PDA

View Full Version : What statistic(s) do you like -- and why?



RedsManRick
07-17-2008, 11:10 AM
Given a number of stat v anti-stat threads that have popped up lately, I thought it might be constructive to actually look at some of the stats people like to use.

As us stat-heads have been told over and over, the game is played on the field by real people. While some of us find beauty in the numbers, undeniably the best part of the game is what happens on the field.

So why do we have stats anyways? Certainly we don't need to know what Adam Dunn's slugging percentage is to enjoy a go ahead 3-run homer. We don't need to know the Volquez has an ERA under 3.00 to cheer at another strikeout.

At the simplest level, we use stats because we're trying to understand the game better. We realize that within the course of a given game it's pretty much impossible to figure out who the best players are. At the more complex end, we're trying to build winning franchises or compare Babe Ruth to Josh Hamilton.

In order to do this, we start counting what happened. At some point in the 1800s, people starting combining measures of events in to rate statistics so we could assess how often things happened. Along the way, we've fiddled with how we count things, certain rules we apply to exclude certain events or give extra credit to others.

The key thing to remember in all of this is that stats are just a means to an end. No stat is intrinsically better or worse by nature of it's existence. Rather, we should judge stats based on how well they do what they are supposed to do.

Taking the case of batting average, it was invented to assess basic offensive performance. The goal of offense, by definition, it so score runs. And because most runs are scored as a result of the combined efforts of players, we try to assess each player's individual run production to understand his contribution.

Over time, we've found ways of combining the measurements of offensive events to measure run production that, when added back up at the team level, better match up with how many runs actually end up scoring. Yet many, if not most people, are resistant to ditching batting average even thought its utility has been easily surpassed by other widely available metrics. It's like somebody said they would upgrade your car for free to make it more comfortable, look better, and get more better gas mileage -- and some people aren't interested.

What I'd like to understand is why.

I'd like to have an open conversation about what stats people like to use, what those stats tell them about the players and the game, and why people prefer them to some alternative. This isn't about any player, it's about the way we measure what they do.

I understand that many people don't like to have this conversation for a number of reasons including:
- You just don't care
- You don't want to spend the time
- You hate math
- Stat-heads can be arrogant and condescending
- It just doesn't make sense
- Bill James broke your Optimus Prime toy in 1986

If nobody is interested in doing this, that's cool. But in effect it seems we have an ugly, contentious version of this topic at least twice a week. If you'd rather ask a question about a stat you don't use, feel free. But let's get this conversation going in a constructive manner.

OnBaseMachine
07-17-2008, 11:15 AM
OPS is probably my favorite statistic because it's simple really. It combines the two most important jobs for a hitter - getting on base plus the ability to drive the ball and score the runners. I didn't understand what it was until about six or seven years ago but now it's the main offensive statistic I use when judging a player. OPS against is also a stat I look at for pitchers.

RedsManRick
07-17-2008, 11:36 AM
So OBM, just to play devil's advocate, what's a good OPS? What's a bad one? How many runs does a certain OPS equate to? Are all OPS's created equal? How much "better" is OPS than batting average, OBP, or SLG by themselves?

I don't expect you to answer those questions necessarily, but I'd like to people ask those sorts of questions about stats they don't like. We can collectively tease out our concerns and hopefully move every so slightly towards a better environment when it comes to the use (and debates) of stats.

oneupper
07-17-2008, 11:43 AM
The stat I'd like the most hasn't been invented yet.

What I would really like to see for hitters (and the reverse for pitchers) is something like "Pitcher Quality Dependence Index".

The idea behind this is the following:

We've all heard stuff like "he's a AAAA hitter" or "he isn't clutch" etc and "good pitching always beats good hitting"...about hitters. And observation shows us that there are players who can make the leap from minors to MLB and pretty much put up the same numbers, while there are the Brandon Larsons of the world who tear it up in AAA but can't hit a lick in the bigs.
Similarly we all seem to obsess about hitters who can't seem to get the big hit against the tough pitcher, but completely demolish lesser hurlers.

Its obvious that all hitters hit worse against better pitchers (as an aggregate), but TO WHICH DEGREE. How much are their hitting skills dependent on the quality of the pitching they face?

To get this index, you would run a regression of the hitter's performance (OPS or one of those offensive stats people like) against a pitcher quality stat (OPSA, Dips or slips or whatever). In the case of OPS vs OPSA, you would "in aggregate" get a line that slopes upwards, however graphed by hitter individually...the lines would be much different. Almost all would slope upwards (barring a freak), but with different starting and ending points and different slopes. (Aside..I was thinking it would be simpler to aggregate pitcher into cuartiles or quintiles according to quality...fewer data points)

Why would this be important? Well, there is one constant as one moves into more challenging situations: pitching improves. It improves going up the minor league scales, it improves in high pressure game situations (clutch?)...it improves as you head into the postseason.

If a prospect can't hit the better pitchers in his level...he may not be a candidate for promotion. And a good hitter who performs consistently regardless of pitching qualtity, might be a better pick for a playoff team than a great hitter who is highly pitcher dependent.

Maybe all this is a bit stupid...but I'd like to see that stat.

And well...you asked. :)

bucksfan2
07-17-2008, 11:52 AM
I like the more simple stats. I like stats that are whole numbers or percentages based upon 1.00. I understand OPS is a good stat but at the same time if I am watching the game or driving the game I don't get a real good example of the components that go into the stat. On the offensive side of the game I like BA, OBP, RBI, R, SB, HR, 2B, H. When you look at those stats together you can get a pretty good description of how and what a player is playing like.

You can tell me a guy is OPSing .800 but what does that mean? Is he getting on base at a .400. .500 clip? Is he slugging .400, .500? Its a summation statistic in which the components aren't clear cut.

For pitching again I like the basic. I like era, whip, k/bb, I wish instead of Innings they give you an average of innings pitched per start. For relievers I like WHIP, k, bb, saves.

I know my way of looking at the game is pretty simplistic but at the same time I look at the stats to give myself a general understanding of the game. Its true that a player can be having a lucky season or an unlucky season but thats where my eyes come in. IMO over a 162 game season the unlucky outs tend to even themselves out with the lucky hits.

Spring~Fields
07-17-2008, 11:53 AM
I like the OBP, and I like the SLG stats because I think that they help me to set aside, to filter out my player preferences or biases and to see who really does get on base against the left handed pitcher and the right handed pitcher or both, quickly.

Right or wrong, I can have a sense or consistent measure of knowing that batter x or y is getting on base more often because of my assumption of their ability to either actually hit the ball or to work a walk, avoiding making outs more often than perhaps player z on the whole or aggregate of the season.

Though those two stats may not be of any help tonight to win this single game, (a small sample)

I can see where over time that they would increase the probability of chances to score more often if the batters with those higher stats get the most AB/PA. I think. Anyway I got that from you, Cyclone, WOY and Steel et el. after hours, weeks, months and years, some understanding finally started to sink in. I still like BA as a quick cheap tool to see who is getting hits in x number bats, but nothing more, kind of like using the musing tool for lineups to see what runs they could produce per given lineup, just a small tool, I think it can tell me if a Hairston is hitting the ball better than a Patterson this month or week, just talking about hits only.

I also use the OBP stat to grade the managers choice and decision, to me if the manager made the right choice and decision regarding usage of player personnel, the OBP should support his choice and decisions on offense.

Raisor
07-17-2008, 11:54 AM
I like OPS for the simplicity of it, and I like Runs Created (or even better, RC/TPA).

With RC, I really like knowing, within a very good margin of error, how much of a team's total runs scored, a player is responsible for. If a stat can tell me, within 95% or so) that player A is worth 75 of a team's 600 runs, then I think that's relevent.

BuckeyeRedleg
07-17-2008, 11:54 AM
OPS is probably my favorite statistic because it's simple really. It combines the two most important jobs for a hitter - getting on base plus the ability to drive the ball and score the runners. I didn't understand what it was until about six or seven years ago but now it's the main offensive statistic I use when judging a player. OPS against is also a stat I look at for pitchers.

I agree about OPS. It's simple. It's the first thing I look for in a hitter. Then I like to compare his BA with his OBP and SLG, so BA is useful to determine how plate discipline or power is controlling his OBP or SLG.

RC is just fun to look at. I really enjoy messing around with ESPN.com's sortable statistics. It's a great way to compare players at the same position. There are several filter options so that you can really have a good time with the data.

http://sports.espn.go.com/mlb/stats/batting?league=mlb

For pitchers, the first thing I look for is K/9, then BB/9, and HR/9.

You can tell a lot about a pitcher just from those three, especially K and BB/9. I could care less about ERA and W/L. I think ground ball to fly ball ratio would probably be pretty useful too, but I admit I sometimes skip right past that.


I also check out Baseball Prospectus daily. This is a great page right here.

http://www.baseballprospectus.com/statistics/sortable/

I love to check the adjusted standings. Also, even though I don't quite understand how it is determined, I like to check out VORP occasionally. The Run Expectancy Matrix is pretty nice to look at as well.

SunDeck
07-17-2008, 11:59 AM
W/L - I like that one a lot.

I like OPS because after all, there are only so many bases you can fill. After that, they become runs.

I like WHIP and I like you're basic strikes to balls ratio. Guys who through a lot of balls are not efficient and I think good pitching is about getting guys in and out of the batter's box as quick as possible (although I would also like to have a stat that tells me how many times a pitcher backs guys off, which does not happen nearly enough any more).

Falls City Beer
07-17-2008, 12:00 PM
I like OPS for the simplicity of it, and I like Runs Created (or even better, RC/TPA).

With RC, I really like knowing, within a very good margin of error, how much of a team's total runs scored, a player is responsible for. If a stat can tell me, within 95% or so) that player A is worth 75 of a team's 600 runs, then I think that's relevent.

Yeah, though "these stats may have been passed by" they say a lot to me as well. They jibe nicely with my intuitive sense of what a player is doing relative to the other members of the team and other players in baseball.

For pitchers, I like OPSA--it doesn't do much for "prediction" but it says quite well, in my opinion, who the best starters are in any given season when you arrange them in order of OPSA (with a minimum number of innings pitched of course). The very good/elite starters will be @ .625-.675, the solid to very goods will be .675-.725; the averagey .725-765; the just hanging on to effectiveness .770-800; the useless or nearly useless fodder. .800 +

OnBaseMachine
07-17-2008, 12:00 PM
A good OPS varies for certain positions. There is a big difference between an .800 OPS for a left fielder and an .800 OPS for a shortstop. Without looking it up right now an .800 OPS out of a LFer is around average I suspect whereas an .800 OPS out of a shortstop is outstanding. It all depends on the position.

As for how OPS compares to BA, it doesn't compare IMO. Batting average just measures how often you get a hit and it counts a homerun the same as a bunt single. OPS of course measures how often a hitter gets on base whether it be via hit, walk, or HBP and it's second component, slugging percentage, rewards extra base hits instead of counting them the same as a bunt single such as BA does. To me, On Base Percentage is the single most important statistic for a hitter. Getting on base gets everything started, and that's when slugging comes in. Getting on base and then driving the ball is what scores runs. OPS measures both. A simple stat IMO.

Chip R
07-17-2008, 12:08 PM
The problem with OPS is that it uses two stats that are difficult to figure off the top of your head. With batting average all you need are the number of ABs and hits and that's pretty much provided for you at every ballpark. It's not a bad stat and it's useful but unless you have OBP and SLG handy, it's difficult to figure out. Then, of course, when you start weighing one of them over the other, it gets even more complex.

Raisor
07-17-2008, 12:16 PM
The problem with OPS is that it uses two stats that are difficult to figure off the top of your head. With batting average all you need are the number of ABs and hits and that's pretty much provided for you at every ballpark. It's not a bad stat and it's useful but unless you have OBP and SLG handy, it's difficult to figure out. Then, of course, when you start weighing one of them over the other, it gets even more complex.

I'm not sure what they do at the GAB, but at the Ted they update OBP/SLG every PA.

gonelong
07-17-2008, 12:17 PM
The problem with OPS is that it uses two stats that are difficult to figure off the top of your head. With batting average all you need are the number of ABs and hits and that's pretty much provided for you at every ballpark. It's not a bad stat and it's useful but unless you have OBP and SLG handy, it's difficult to figure out. Then, of course, when you start weighing one of them over the other, it gets even more complex.

The Dayton Dragons post OPS right in the free program they hand out. Love it.

BA/OBP/SLG/OPS

They also post OBP on the screen when a player is up.

GL

edabbs44
07-17-2008, 12:18 PM
I like the traditional stats. BA/HR/RBI/Runs scored. I then like to supplement with other stats.

For example: If someone has a lot of RBI then that means, to me, that the guy is having a pretty good year in that regard. Unless there is a supplementary stat that can knock that down (i.e. had twice as many runners on base than anyone else in baseball), then I like RBI.

I hate when people look at any stat and refuse to use others to supplement their thoughts. Just because RBI can be discounted in some instances doesn't mean it is worthless in all cases.

Of the "newer" ones, I like OPS. I think it has a lot of value. But just like any other stat, all OPSes aren't created equal.

OnBaseMachine
07-17-2008, 12:18 PM
I'm not sure what they do at the GAB, but at the Ted they update OBP/SLG every PA.

Yeah, they do the same at GAB.

Chip R
07-17-2008, 12:19 PM
I'm not sure what they do at the GAB, but at the Ted they update OBP/SLG every PA.


They do that here too but if you are at a park that doesn't display that info - minor league parks in particular - you're SOL.

Spring~Fields
07-17-2008, 12:22 PM
The stat I'd like the most hasn't been invented yet.

What I would really like to see for hitters (and the reverse for pitchers) is something like "Pitcher Quality Dependence Index".

The idea behind this is the following:

We've all heard stuff like "he's a AAAA hitter" or "he isn't clutch" etc and "good pitching always beats good hitting"...about hitters. And observation shows us that there are players who can make the leap from minors to MLB and pretty much put up the same numbers, while there are the Brandon Larsons of the world who tear it up in AAA but can't hit a lick in the bigs.
Similarly we all seem to obsess about hitters who can't seem to get the big hit against the tough pitcher, but completely demolish lesser hurlers.

Its obvious that all hitters hit worse against better pitchers (as an aggregate), but TO WHICH DEGREE. How much are their hitting skills dependent on the quality of the pitching they face?


Why would this be important? Well, there is one constant as one moves into more challenging situations: pitching improves. It improves going up the minor league scales, it improves in high pressure game situations (clutch?)...it improves as you head into the postseason.

If a prospect can't hit the better pitchers in his level...he may not be a candidate for promotion. And a good hitter who performs consistently regardless of pitching qualtity, might be a better pick for a playoff team than a great hitter who is highly pitcher dependent.


I would like to see something like that also, something that tells us who really is the better hitters, assuming those are the ones that can and do hit the best of pitching more often than the also ranís of hitting lore. Since stats are lumped together over the season and we have that place where all batters pad their stats to a degree against pitchers having a AAA or AAAA night. I would like to see a stat that weeds that out, and gives credit to the batters that can hit good pitching as well as feast off poor pitching from the other lesser batters.

It is nice to get happy when our Reds roll over a pitcher producing a lot of hits, extra base hits, homeruns etc, but it often is because of a pitcher having a bad day vs. the thought that our players being so much better at hitting all of a sudden.

Does Utley and Pujols hit better or good pitching more often than another player? If so then they would be the better hitters because they can hit good and bad pitching, not just feast on a the struggling pitchers. (A lot of Reds batters roll over poor pitching, Ohio States offense rolls over weak defenses, doesnít make Ohio State good on offense, which is exploited when they run into a truly good defense.)

Raisor
07-17-2008, 12:23 PM
They do that here too but if you are at a park that doesn't display that info - minor league parks in particular - you're SOL.

OK, if you're needing an OPS update during a minor league game RIGHT NOW, you probably should be brining your laptop with you.

:thumbup:

gonelong
07-17-2008, 12:23 PM
It'd be interesting to put something up on the scoreboard between innings.

RC/27 attempts to measure blah, blah, blah, but doing this ....
Then list the 10 in RC/27

... wonder what kind of discussions that might stimulate down at the old ball yard?

GL

Raisor
07-17-2008, 12:25 PM
It'd be interesting to put something up on the scoreboard between innings.

RC/27 attempts to measure blah, blah, blah, but doing this ....
Then list the 10 in RC/27

... wonder what kind of discussions that might stimulate down at the old ball yard?

GL


Exactly right. I've always wondered why clubs don't do more to educate their fan base.

RedsManRick
07-17-2008, 12:33 PM
A good OPS varies for certain positions. There is a big difference between an .800 OPS for a left fielder and an .800 OPS for a shortstop. Without looking it up right now an .800 OPS out of a LFer is around average I suspect whereas an .800 OPS out of a shortstop is outstanding. It all depends on the position.

Again, to play devil's advocate, it depends on what you're tyring to measure. From an offensive production perspective, an .800 OPS from a SS is just as productive as an .800 OPS from a LF. Now, if you want to talk about player value, it's much rarer to get that level of production from a SS than from a LF, and thus it's worth more money. But it's not worth more runs.

Not a comment for you in particular, but I think this sort of conflating of concepts, something which we all do, trips up the conversation from time to time. We have to be careful in explicitly describing what we're measuring before we apply value judgments like "better" and "worse".

Chip R
07-17-2008, 12:35 PM
OK, if you're needing an OPS update during a minor league game RIGHT NOW, you probably should be brining your laptop with you.

:thumbup:


Not necessarily just an update but the actual OPS before the game. What if Mrs. Raisor went to a game with you and wanted to know what a certain player's OPS was? You couldn't tell her cause all they have on the scoreboard is batting average and maybe hits and ABs. ;)

durl
07-17-2008, 12:36 PM
I'm old school. My favorite stat is team W/L.

I believe certain stats are interesting like OPS. To me, BA, RBIs, and Runs Scored mean the most for reviewing offense. But my head starts swimming when people begin discussing the value of IBosEOU versus OQerVNC. Or if a player's IPwoXB gives him a higher trade value.

It's odd that I don't get into myriad of modern stats since I deal with numbers as part of my job and I'm been accused (rightfully) of over-analyzing stuff. But baseball is a simple game. I like it that way.

Raisor
07-17-2008, 12:38 PM
Not necessarily just an update but the actual OPS before the game. What if Mrs. Raisor went to a game with you and wanted to know what a certain player's OPS was? You couldn't tell her cause all they have on the scoreboard is batting average and maybe hits and ABs. ;)

Mrs Raisor knows to look at that stuff before we go. Or she BETTER know.

But in seriousness, most of the minor league stadiums I've been to recently have had OBP/SLG listed in the program.

nate
07-17-2008, 12:39 PM
I like OPS+ for comparing a player to his peers.

I like OBP because it measures a player's "not making an out" rate. Just think if the back of baseball cards had 1-OBP for a "FAIL %!" I mean, in 2004, Barry Bonds failed to get on base only 40% of the time. Crz.

RBI is fine but it's a team dependent stat. It doesn't tell me anything about a player's ability to drive in runs. All it tells me is that the batter has a lot of home runs or has gotten hits when there are people on base. But, it doesn't tell me what he'll do if he goes from the team that scores the most runs in the league to the one that scores the fewest.

For pitchers, I like K/9 because it tells me how many outs they can get on their own. Hi K/9 pitchers can still have success on a crappy defensive team (for example, the Reds). I'm trying to figure out which fielding independent pitching stats I like.

For the team, I like to follow the pythag record compared to their actual record.

nate
07-17-2008, 12:40 PM
It'd be interesting to put something up on the scoreboard between innings.

RC/27 attempts to measure blah, blah, blah, but doing this ....
Then list the 10 in RC/27

... wonder what kind of discussions that might stimulate down at the old ball yard?

GL

I like it!

bucksfan2
07-17-2008, 12:42 PM
Again, to play devil's advocate, it depends on what you're tyring to measure. From an offensive production perspective, an .800 OPS from a SS is just as productive as an .800 OPS from a LF. Now, if you want to talk about player value, it's much rarer to get that level of production from a SS than from a LF, and thus it's worth more money. But it's not worth more runs.

Not a comment for you in particular, but I think this sort of conflating of concepts, something which we all do, trips up the conversation from time to time. We have to be careful in explicitly describing what we're measuring before we apply value judgments like "better" and "worse".


I think RMR brings up a good point. Do you really care about the value of each individual player or the overall production of an entire team. Look at the Phillies they have great production coming for the 2b and SS positions. Does that enable you to build your team differently? Do you worry more about order slot hitters say your 2 hitter or 3 hitter? Would a 2B and LF of Utley and Pierre out produce a lineup of two average OPSing 2B and LF?

RedsManRick
07-17-2008, 12:52 PM
I like the traditional stats. BA/HR/RBI/Runs scored. I then like to supplement with other stats.

For example: If someone has a lot of RBI then that means, to me, that the guy is having a pretty good year in that regard. Unless there is a supplementary stat that can knock that down (i.e. had twice as many runners on base than anyone else in baseball), then I like RBI.

Just to tease this out, when you say "in that regard", to what "regard" are you referring? Certainly you don't value RBI in and of itself, but rather as some approximation for run production.

So let's create a better version of RBI, using information we already collect, that better controls for opportunity. One measure in nearly every box score is LOB at both the player and team level - so we already capture how many guys are on base when a guy comes up.

Why not combine these? We could choose to either included the batter or not depending on your preference.

Runners Driven In / Runners On Base.

In his 403 PA, Brandon Phillips has had a total of 282 runners on base and has driven 43 of them. (He has 15 HR). So we can say that Phillips has driven in 15.2% of the runners he could have. If we choose to include him as a potential run, he's driven in 8.5% (43+15/403+282).

By contrast, Adam Dunn has driven in 33 base runners of the 214 that have been on base for him in his 368 plate appearances; a rate of 15.4%. If we include him as a potential run, he's driven in 10.1% (33+26/214+368).

Now, those new stats I just made up (actually stole from Baseball Prospectus) are very straight forward. There's no fancy math going on there. To me, they're petty easy to interpret. Using this year's performances, Brandon Phillips has an 15.2% chance of driving in runners on base and Dunn has a 15.4% of driving them in.

That was the question were really trying to answer, right? Who is the best at driving in runners on base?

Wouldn't this be more meaningful for you than RBI?

The reliance on the "RBI" stat often makes me think that a lot of people's comfort with certain stats is based on their accessibility. If stadium scoreboards showed different stats, if they appeared on the back of the guys baseball card, people wouldn't clamor for RBI. The use RBI comes out of the fact that we've used it for 100 years and we know what the scale of bad-good looks like, not because it's a particularly good way to measure what we're interested in measuring.

To put it another way, if you had both RBI% and RBI, which one would you want to look first if you wanted to know how well a guy has driven in runs this year?

Spring~Fields
07-17-2008, 12:55 PM
My favorite stat is team W/L.



In baseball only, strangely enough I have found myself getting bored with that stat if there is an imbalance or habituation of winning or losing too much. Conversely, in a short series though like in 1990 I was very pleased with the quick win stat.

Raisor
07-17-2008, 12:58 PM
Would a 2B and LF of Utley and Pierre out produce a lineup of two average OPSing 2B and LF?

Yes, Utley and Pierre would out produce league average 2b and LF.

A league average LFer outproduces Pierre, but Utley is so far above average he makes up for it.

RedsManRick
07-17-2008, 01:00 PM
I'm old school. My favorite stat is team W/L.

Why? What does W/L tell you? That's not the end goal, right? It's about winning the World Series. Getting to the World Series doesn't require winning the most games, just winning your division or the wild card.

So isn't your standing in your division relative to your competitors more important? I mean, who cares if we're 72-90, if we win the division, that's all that counts, right? W/L record is just a proxy for this...

But if the question is "Which teams are more likely to win in the future?", W/L isn't really the best predictor of that either. We'd be better off looking at RS & RA.

One could argue that W/L is completely unnecessary. :evil:

RedsManRick
07-17-2008, 01:03 PM
Yes, Utley and Pierre would out produce league average 2b and LF.

A league average LFer outproduces Pierre, but Utley is so far above average he makes up for it.

But to take the generic case, a 650 OPS LF and 950 OPS 2B doesn't lead to any more runs scoring than a 950 OPS LF and 650 OPS 2B -- or arguably, an 800 OPS LF and 800 OPS 2B.

The only time position matters from an offensive perspective is when you have to pay the guys -- or try to replace them.

Raisor
07-17-2008, 01:14 PM
But to take the generic case, a 650 OPS LF and 950 OPS 2B doesn't lead to any more runs scoring than a 950 OPS LF and 650 OPS 2B -- or arguably, an 800 OPS LF and 800 OPS 2B.

The only time position matters from an offensive perspective is when you have to pay the guys -- or try to replace them.

That's right. Team wise, it doesn't matter how you get to that 800 OPS. It matters when you're comparing individual players.

oneupper
07-17-2008, 01:14 PM
Why? What does W/L tell you? That's not the end goal, right? It's about winning the World Series. Getting to the World Series doesn't require winning the most games, just winning your division or the wild card.

So isn't your standing in your division relative to your competitors more important? I mean, who cares if we're 72-90, if we win the division, that's all that counts, right? W/L record is just a proxy for this...

But if the question is "Which teams are more likely to win in the future?", W/L isn't really the best predictor of that either. We'd be better off looking at RS & RA.

One could argue that W/L is completely unnecessary. :evil:

You could take that concept a bit further...the goal of a fan isn't necessarily winning the World Series...it's enjoyment. Perhaps a WS would provide the maximum enjoyment, but one could argue against. Imagine a superteam incapable of losing...wins every game by a blowout. How much fun would that be?

We should construct a "joy index", which would have as variables:

Win-Loss record (wins ARE usually more fun)
Season Contention (are we IN it?)
Close games/walkoffs
Brawls won/lost
Stadium giveaway quality
Announcer enthusiasm
Milestones/records attained or thwarted (in case of negative records).
and so on...

If the team doesn't win it ALL (as only one team can)...at least one can say if the season was enjoyable or not...

Isn't that the goal of all entertainment? To entertain?

Wheelhouse
07-17-2008, 01:16 PM
Team Wins -- reason: it's the only statistic that counts.

Raisor
07-17-2008, 01:19 PM
Team Wins -- reason: it's the only statistic that counts.

Personally I like to know WHY a team is winning, or losing.

RedsManRick
07-17-2008, 01:21 PM
Oneupper, that's precisely my point. We need to figure out what our goal is and then figure out how best to measure it. It's when we disagree about what we're trying to measure, but don't realize it, that we get in to this tail chasing debates about what stats are "best".

As for your specific case, that's a great point. We could take this from either the fan or the team perspective. What, ultimately, do you care about as a fan? Winning? Fun? As organization or owner -- Winning? Fun? Money?

Once you answer that top line question (Winning is priority #1), then you can you ask the next questions (How do you win?) and the next questions (How do you score and prevent runs?). Eventually, to answer all of the questions, you need metrics. By thinking about metrics in this way, you pretty quickly can start to see that stats aren't all created equal. They only exist to serve some purpose, to help us achieve something we actually care about. And we should evaluate them in that light.

RedsManRick
07-17-2008, 01:22 PM
Personally I like to know WHY a team is winning, or losing.

Why? Because it's interesting? Because you need to know it in order to win more? So you know which things to cheer and which to boo?

westofyou
07-17-2008, 01:23 PM
Personally I like to know WHY a team is winning, or losing.

Imagine if Copernicus had actually just said...

"Oh.. the Sun came up again... that's good, who cares why... it's just good that it happened."

Raisor
07-17-2008, 01:24 PM
Why? Because it's interesting? Because you need to know it in order to win more? So you know which things to cheer and which to boo?

Because it's interesting.

I probably should have been an engineer or something. I like know how/why things work.

Cooper
07-17-2008, 01:29 PM
I like them all. The more the better. If I don't like the stat -i quickly look away so it does not effect my overall enjoyment of the game.

dougdirt
07-17-2008, 01:33 PM
A lot of you guys are talking offensive stats, I will go the other way and get into pitching stats.

I like K%, BB% and GB%.

I like K% because unlike K/9 or even straight up strikeouts, it tells me how often the pitcher strikes someone out. In K/9 a pitcher could strikeout 1 guy, walk the next 2, strikeout another guy, then allow 2 hits and strike out the final guy of the inning. That tells you that the guy is striking out 27 guys per 9 innings, but he is only striking out 57%. If another pitcher strikes out the side, he is doing a better job. Obviously thats an example on a small scale, but over a season a guy could have a relatively higher K/9 but strike out a significantly lower % of batters because of the amount of walks he gives up.

As for walk %, it is a lot closer to BB/9, but I feel its better to say over 200 innings that a guy is walking 7% of the batters he faces rather than he is walking 2.3 guys per 9 innings. Its not easily accessible really, but I feel it gives you a better understanding.

Groundball % is something I like with pitchers because groundballs don't go for home runs, and they also induce a lower slugging percentage (possibly because of the lack of HR's). That leads to lower scoring in the long run, and that is one thing a pitcher is supposed to do, keep runs to a minimum.

Those are my favorite things to look at when it comes to a pitcher.

Caveat Emperor
07-17-2008, 01:39 PM
Team Wins -- reason: it's the only statistic that counts.

I see your Team Wins and raise you World Series Titles.

There's a reason the Buffalo Bills of the 90s were the butt of so many jokes, and it had very little to do with their Team Wins.

dougdirt
07-17-2008, 01:42 PM
I see your Team Wins and raise you World Series Titles.

There's a reason the Buffalo Bills of the 90s were the butt of so many jokes, and it had very little to do with their Team Wins.

Pft, of course they couldn't win the World Series, they play fooseball :thumbup:

remdog
07-17-2008, 01:53 PM
Why? What does W/L tell you? That's not the end goal, right? It's about winning the World Series. Getting to the World Series doesn't require winning the most games, just winning your division or the wild card.

So isn't your standing in your division relative to your competitors more important? I mean, who cares if we're 72-90, if we win the division, that's all that counts, right? W/L record is just a proxy for this...

But if the question is "Which teams are more likely to win in the future?", W/L isn't really the best predictor of that either. We'd be better off looking at RS & RA.

One could argue that W/L is completely unnecessary. :evil:

I realize that you're kidding here but the general sendup that comes from this very closely parallels my opinion of the 'statheads', i.e., the numbers (stats) are more important than the game, winning/losing, enjoying the play, etc., etc., etc.

Furthermore, the emphasis on stats has taken such a stranglehold on this board that the board has, to me at least, become almost unreadable. It's boring, repetitve and is dominated by a small group of people who seem to have a need to reduce everything to a finite number. Here's a hint: if you're posting more than 5 or 6 times a day you need to (IMO) get a life! Not only that, as in most things in life, you learn more by keeping your ears open and your mouth shut than the other way around.

And, it's great if you like stats but really, at what point is enough enough? How thinly must a hair be split? Are you so busy crunching numbers that you missed the shear joy of suddenly seeing a great play or moment? Honestly, reading this board I have to believe that many here missed those plays and moments more than they saw them.

I suppose I'll get ripped for this post but, guess what, I really don't give a rat's behind. (Famous shrug) It may be a reflection of the Reds performance this year (as well as several before it) but, honestly, I come by here more for any breaking news not for baseball discussion because, well, frankly, no one discusses much anymore. They just throw out numbers, argue minutiae and beat on Dunn, Jockety, Dusty, etc.

I think I'll just sit back tonight and enjoy Johnny Cueto perform his role as that of a rookie pitcher with loads of talent trying to find his way to the top. I'll look at the stats in the morning. Was it the Bard that said, 'The play is the thing?'

Rem

*BaseClogger*
07-17-2008, 02:01 PM
Let's see...

Honestly, I'm not a fan of OPS. I don't like not knowing the parts that go into it and its wierd scale, so I much prefer seeing a hitter's BA/OBP/SLG. I also like OPS+ when I compare hitters from different environments because it normalizes and has a cool scale. Looking at a hitter's BABIP, LD%, and HR/FB% helps explain a batters success or lack there of. LD% explains BABIP and HR/FB% explains why a hitter might have less home runs but more doubles.

For pitchers, I first look at their K/9, BB/9, and HR/9. Then, GB% because it includes outs and all other outcomes. Finally, I see how they all add up for the pitcher's xFIP, which I like because it is on a scale I am confortable with (ERA). To check their luck (or randomness :)) I look at BABIP, LD%, and HR/FB%.

*BaseClogger*
07-17-2008, 02:05 PM
It may be a reflection of the Reds performance this year (as well as several before it) but, honestly, I come by here more for any breaking news not for baseball discussion because, well, frankly, no one discusses much anymore.

The only thing I want to know is... how is this not discussion? What made the "glory days" of RZ more of a disscussion than today's RZ?

Raisor
07-17-2008, 02:06 PM
I realize that you're kidding here but the general sendup that comes from this very closely parallels my opinion of the 'statheads', i.e., the numbers (stats) are more important than the game, winning/losing, enjoying the play, etc., etc., etc.

Furthermore, the emphasis on stats has taken such a stranglehold on this board that the board has, to me at least, become almost unreadable. It's boring, repetitve and is dominated by a small group of people who seem to have a need to reduce everything to a finite number. Here's a hint: if you're posting more than 5 or 6 times a day you need to (IMO) get a life! Not only that, as in most things in life, you learn more by keeping your ears open and your mouth shut than the other way around.

And, it's great if you like stats but really, at what point is enough enough? How thinly must a hair be split? Are you so busy crunching numbers that you missed the shear joy of suddenly seeing a great play or moment? Honestly, reading this board I have to believe that many here missed those plays and moments more than they saw them.

I suppose I'll get ripped for this post but, guess what, I really don't give a rat's behind. (Famous shrug) It may be a reflection of the Reds performance this year (as well as several before it) but, honestly, I come by here more for any breaking news not for baseball discussion because, well, frankly, no one discusses much anymore. They just throw out numbers, argue minutiae and beat on Dunn, Jockety, Dusty, etc.

I think I'll just sit back tonight and enjoy Johnny Cueto perform his role as that of a rookie pitcher with loads of talent trying to find his way to the top. I'll look at the stats in the morning. Was it the Bard that said, 'The play is the thing?'

Rem


Rem,

Honest question.

If you dislike what you think RZ is so much, why come back? Or why not just read subjects you aren't going to think are "unreadable"?

RedsManRick
07-17-2008, 02:15 PM
Rem, in a very odd way, I think we agree. Too much analysis, particularly bad analysis, can detract from the game itself. While I certainly get more enjoyment out the analysis process, and you clearly don't, the real goal of a post like this, for me, is to not waste my time on stats that don't matter and maybe help other fans do the same. When I hear an announcer drone on about batting average w/ RISP it significantly detracts from my enjoyment of the game. When I watch a game, I want to enjoy the aesthetic and dramatic beauty. I derive no pleasure during the game from Adam Dunn's OBP, nor from talking about it.

My entry in to the sabermetric world was based on an annoyance when announcers and commentators moved away from describing the game and started trying to analyze it. The game was less interesting to watch/listen to, and the insights they provided were marginal at best, misleading at worst. It's hard to enjoy the beauty of watching Dunn work a pitch count when I'm being told how overpaid he is. So I decided to do my homework about stats when I wasn't watching the game.

In the moment, who cares about how many RBI a guy has? It doesn't help me enjoy his double in the corner and the subsequent play at the plate. And likewise, I'm as frustrated as the next guy when Dunn strikes out with a guy on 3B and 1 out, missing an opportunity to drive in the run. But when the announcer tells me that Dunn sucks because of his strikeouts, he's opened that analysis door. Suddenly we're not talking about the game, but of the value of the player. And if we're going to take the time to time split hairs, to turn our attention away from the aesthetic and dramatic enjoyment the game provides, I want to spend it splitting the right ones.

Perhaps this is a case where the treatment is worse than the disease for most folks. Certainly I'm an offender of your 6 post rule and I'm sure my job performance would be better if I posted less. That said, the discussion itself is a source of great enjoyment for me. But that doesn't mean I watch and enjoy the game played on the field any less than you do and I take offense at the notion. It's 1:06 and the Reds aren't playing baseball right now. I don't see how it detracts from my enjoyment of the game tonight that I like to analyze the game when I'm not watching it.

Back to my initial point. I think my enjoyment of the game, as I'm watching it, would benefit significantly from announcers who spent more time perfecting their craft and less time trying to do statistical analysis. And when there's no game to watch and I want analysis, I want to get it from people who are experts in analysis. We'd all benefit from more Vin Scullys and Nate Silvers and less "EE is NOT clutch" calls and "Adam Dunn strikes out too much" posts. I can't do much about the announcers, but I'm trying to do my part in eliminating the repetitive, voluminous debates about what stats matter most.

Rojo
07-17-2008, 03:05 PM
The explosion of stats has coincided with the explosion of fantasy leagues. Evaluating individual performances is crucial of course and sabremetrics does this well. The unknown country is analyzing how teams perform as a whole and why.

SMcGavin
07-17-2008, 03:11 PM
I like OPS+, just because it's simple to look at and a pretty good barometer of overall offensive prowess. RC/27 is another good one, but it's not as easy to access or calculate.

But like others have said, not every OPS is not created the same. So I actually look at batting average every now and then. For a batter, if I can see a line of AVG/OBP/SLG/OPS, I feel like I have a good idea of that hitter's ability.

Always Red
07-17-2008, 03:11 PM
I realize that you're kidding here but the general sendup that comes from this very closely parallels my opinion of the 'statheads', i.e., the numbers (stats) are more important than the game, winning/losing, enjoying the play, etc., etc., etc.

Furthermore, the emphasis on stats has taken such a stranglehold on this board that the board has, to me at least, become almost unreadable. It's boring, repetitve and is dominated by a small group of people who seem to have a need to reduce everything to a finite number. Here's a hint: if you're posting more than 5 or 6 times a day you need to (IMO) get a life! Not only that, as in most things in life, you learn more by keeping your ears open and your mouth shut than the other way around.

And, it's great if you like stats but really, at what point is enough enough? How thinly must a hair be split? Are you so busy crunching numbers that you missed the shear joy of suddenly seeing a great play or moment? Honestly, reading this board I have to believe that many here missed those plays and moments more than they saw them.

I suppose I'll get ripped for this post but, guess what, I really don't give a rat's behind. (Famous shrug) It may be a reflection of the Reds performance this year (as well as several before it) but, honestly, I come by here more for any breaking news not for baseball discussion because, well, frankly, no one discusses much anymore. They just throw out numbers, argue minutiae and beat on Dunn, Jockety, Dusty, etc.

I think I'll just sit back tonight and enjoy Johnny Cueto perform his role as that of a rookie pitcher with loads of talent trying to find his way to the top. I'll look at the stats in the morning. Was it the Bard that said, 'The play is the thing?'

Rem

I think this post is very timely, and I agree with nearly all of it. Rem's post has also given me the courage to speak up a bit, as I have felt this same way for a while here now.

Everyone who comes here loves baseball, and loves the Reds; I think that's a given. I certainly respect that about everyone who stumbles through these threads.

Stats have certainly improved my enjoyment of the game; or maybe I should say my understanding of the game. Frankly, when I'm sitting in section 526, I'm not paying all that much attention to FIP, BABIP, Win Shares or RC. Instead I'm wondering if BP is ever going to be able to hit a crooked pitch from a RH pitcher, or if Keppinger's range is decreasing by the day like I think it is, or why all of my friends hate Adam Dunn so much, when it is so clear to me that he's one of the best players on this team. Yes, i can go back and use stats to give me some ideas and explanations for all of those things. In some cases, they can even tell me that what i see with my eyes is not really the truth. And I respect that, I really do.

Stats are pure, cold truth. They have never told a lie. They describe, mathematically, what has taken place on the field of play. There is no emotion in them, and they have no agenda.

However, we cannot say the same about those who analyze stats. This is only my 3rd baseball season here on ORG, but it seems like things have changed a lot during that time. As rem mentioned, many of the threads are unreadable, at least without a subscription to Baseball Prospectus opened in another window. Posters arrogantly clobbering each other over the head with minutia doesn't make for pleasant reading. I try to avoid those when they occur, and of course the best thing is not to get involved when numbers, decimals and pitching abuse points start getting sprayed around.

And yet I still come here. Why? Well, I love baseball, I love the Reds, and I like most of the posters here. And every once in a while, there is a magical thread here, one that I just can't get enough of. For me, of late it has been 2 threads that Cyclone (what a great poster Cyclone is!) started, one last week on baseball's golden age, and one that he started in 2006 on historical baseball photographs. That thread is sheer magic to me, I must have downloaded over 30 old pictures, and then went to the referenced web site and downloaded some more.

I think of the ORG as a big tent. There is a lot here, and room for a lot of folks with a lot of different opinions. I hope that there will always be room for those of us who just like to talk baseball, without being talked down to because we might not agree that, for example, over 100 pitches a game means that a guy's arm is going to fall off, or that Belisle really might suck, and he's not just unlucky because his BABIP against is kind of high, or that his FIP is lower so it's just his defenders behind him who suck. I understand that stats can quantify things, but not all of them are exactly scientific.

Some days I am not so sure that a guy like me belongs here anymore, and some days I am. I actually feel sometimes like I have nothing to add because I don't post a lot of stat-related stuff. Then I realize that I belong here as much as anyone else does, that we all love the game, and that we can enjoy it differently from each other as well. For me, one of my best, most sacred memories of the game was being at Crosley Field as a young boy, sitting next to that little postage stamp of green in a bad neighborhood, between my dad and my grandpa (both of whom are deceased), watching the Reds and having my dad point out Willie Mays, telling me he's the best of the best. We all have different memories, we all love the game just a little bit differently. We need to all remember to respect that in each other.

_Sir_Charles_
07-17-2008, 03:32 PM
I'll probably get blasted for it, but I still look at batting average first and foremost. For the power numbers I look straight at HR's and RBI's. For pitchers, I look at ERA, W/L, K's & BB's. I know, I know. But some habits are hard to break. I know that OBP is a huge deal, but I'd personally love to sacrifice half of Dunn's walks for some more hits and him having a respectable batting average. It would just make it easier for me to stomach having a player on my team making the kind of money he is when I see that batting average.

Jpup
07-17-2008, 03:33 PM
The ones that make me right.:cool:

Rojo
07-17-2008, 03:43 PM
The ones that make me right.:cool:

Support not illumination, that's what its all about. ;)

_Sir_Charles_
07-17-2008, 03:44 PM
I realize that you're kidding here but the general sendup that comes from this very closely parallels my opinion of the 'statheads', i.e., the numbers (stats) are more important than the game, winning/losing, enjoying the play, etc., etc., etc.

Furthermore, the emphasis on stats has taken such a stranglehold on this board that the board has, to me at least, become almost unreadable. It's boring, repetitve and is dominated by a small group of people who seem to have a need to reduce everything to a finite number. Here's a hint: if you're posting more than 5 or 6 times a day you need to (IMO) get a life! Not only that, as in most things in life, you learn more by keeping your ears open and your mouth shut than the other way around.

And, it's great if you like stats but really, at what point is enough enough? How thinly must a hair be split? Are you so busy crunching numbers that you missed the shear joy of suddenly seeing a great play or moment? Honestly, reading this board I have to believe that many here missed those plays and moments more than they saw them.

I suppose I'll get ripped for this post but, guess what, I really don't give a rat's behind. (Famous shrug) It may be a reflection of the Reds performance this year (as well as several before it) but, honestly, I come by here more for any breaking news not for baseball discussion because, well, frankly, no one discusses much anymore. They just throw out numbers, argue minutiae and beat on Dunn, Jockety, Dusty, etc.

I think I'll just sit back tonight and enjoy Johnny Cueto perform his role as that of a rookie pitcher with loads of talent trying to find his way to the top. I'll look at the stats in the morning. Was it the Bard that said, 'The play is the thing?'

Rem

Just so you know, you're FAR from alone in this feeling. I like looking at box scores in the paper as much as the next guy. But stats are a VERY small part of the game for me...especially all the new stats we've seen creep up in the last 20 years. Just give me a TV or a radio, 2 hours (or so) of peace and quiet and I'm a happy camper.

Puffy
07-17-2008, 03:48 PM
I'm old fashioned - my favorite stat is Bra/Cup Size.

bucksfan2
07-17-2008, 03:50 PM
Stats are pure, cold truth. They have never told a lie. They describe, mathematically, what has taken place on the field of play. There is no emotion in them, and they have no agenda.


Explain.

Lets look at OPS it is OPB + SLG. That is not a lie. But what exactly is the intrepretation of OPS? If I say player X has an OPS of .800, what does that mean? The key with any statistical analysis is the intrepretation of that. True if you can not dispute the components of OPS but why those two components?

The RC stat. Granted it is a calculation that returns a number. But there is a human manipulation to that stat. Why are certain numbers used? Why are certain stats used and others not? Somewhere in another thread it was said that the RC stat was 95% accurate but when I calculated what the 5% difference was it accounted for 35 runs. Is that a cold hard fact or is that an intrepretation?

Why is Slugging based ot of 4 instead of 1? A single in the forumla counts as 1 but in reality it is only a quarter of the way to a whole. In reality that single doesn't count as much unless it plates a run or the batter ends up scoring.

In all reality I could come up with this stat, BA+OBP+SLG and call it UPS. It would be pretty useless but would it lie? In reality a statistic is any combination of numbers. Some are good, some are bad, but when you take it as the absolute truth it must be 100% instead of that 95%. That 5% you are off can mean the differnece from a World Series Championship or sitting home during October, think St. Louis circa 2006.

RichRed
07-17-2008, 03:51 PM
I'm old fashioned - my favorite stat is Bra/Cup Size.

Big John Kruk and Dmitri Young fan, are ya?

BuckeyeRedleg
07-17-2008, 04:06 PM
How could a message board dedicated to a baseball team not discuss statistics?

*BaseClogger*
07-17-2008, 04:10 PM
How could a message board dedicated to a baseball team not discuss statistics?

Nah... let's talk about the aura of the game instead... :rolleyes:

Seriously though, sometimes these bickering stat v. scout fights get out of control. But, generally, I don't know what some of these posts are reffering to. I'd say at least 75% of the threads on here are limited to simple baseball discussion which do not require a degree in statistics. Meanwhile, it's the 20 page threads that do. Why? Becauses they are the ones that turn into stat v. scout struggles...

Ltlabner
07-17-2008, 04:15 PM
Understanding the game and enjoying the game are two different things. Sometimes they overlap (i.e. analyisis that leads to a deeper appricatation), sometimes they don't (no spreadsheet can equal the shear beauty of a well turned double play).

As far as statistics go I'm still in novice land. OPS, VORP, RAP and RAR are about the extent of my statistical depth and I'm sure my analysis often leads the more stat-savy scratching their heads and wondering what the hell I am thinking.

Something like below makes sense to me to explain a players performance but much beyond that (into winshares and WARP's and all the ISO/LD% and whatnot) and I've just not gotten that far yet. Not becuase I think it's useless, but because I'm satisfied with what I know and understand that while it's not the most accurate picture, its clear enough for me to function.


# YEAR NAME PA EqA OBP SLG VORP RAR RAP
1. 2008 Brandon Phillips 403 .280 .323 .476 20.5 22.1 8.6
2. 2008 Ken Griffey Jr. 374 .267 .348 .400 3.4 14 -2.3
3. 2008 Adam Dunn 368 .313 .380 .538 24.0 33.5 18.8
4. 2008 Joey Votto 343 .280 .350 .464 13.1 18.5 1.3
5. 2008 Edwin Encarnacion 337 .284 .341 .481 13.5 19.7 5.5

Yes, I have a subscription to BP and yes I know how to post code. But I don't think anyone here would mistake me for being statitically minded.

While I do think RZ has become a tedious bickerfest, I'm glad to see folks are trying to post new and different content. Especially the efforts to have more "soft material" ala "what's your favorite baseball memory" or the historical baseball threads.

Sure beats the heck out of arguing over the same crap.

RichRed
07-17-2008, 04:27 PM
I like OPS but I like to see the whole picture: BA, HR, RBI, OBP, SLG, OPS. And then occasionally delve into some of the more "fine-tuning" stats, like BABIP and P/PA. My eyes still glaze over at some of the more advanced ones though.

As for pitchers, ERA and WHIP are so deeply ingrained in my head from my younger days that even though I get that K/9, BB/9, etc., are better indicators, it's hard to keep my eye from gravitating back to ERA.

I guess the stats I like are the ones that give me the best overall picture of a player's, and team's, performance without being overly complex. But I also like when I learn about a new stat, usually on this board, that tells me something I didn't know before.

And for the record, I also enjoy the sight of a nicely turned double play, a knee-buckling curveball, Ken Griffey Jr.'s smooth swing and all the other moments that make baseball such a unique game.

cincinnati chili
07-17-2008, 04:29 PM
A lot of you guys are talking offensive stats, I will go the other way and get into pitching stats.

I like K%, BB% and GB%.

I like K% because unlike K/9 or even straight up strikeouts, it tells me how often the pitcher strikes someone out. In K/9 a pitcher could strikeout 1 guy, walk the next 2, strikeout another guy, then allow 2 hits and strike out the final guy of the inning. That tells you that the guy is striking out 27 guys per 9 innings, but he is only striking out 57%. If another pitcher strikes out the side, he is doing a better job. Obviously thats an example on a small scale, but over a season a guy could have a relatively higher K/9 but strike out a significantly lower % of batters because of the amount of walks he gives up.

As for walk %, it is a lot closer to BB/9, but I feel its better to say over 200 innings that a guy is walking 7% of the batters he faces rather than he is walking 2.3 guys per 9 innings. Its not easily accessible really, but I feel it gives you a better understanding.

Groundball % is something I like with pitchers because groundballs don't go for home runs, and they also induce a lower slugging percentage (possibly because of the lack of HR's). That leads to lower scoring in the long run, and that is one thing a pitcher is supposed to do, keep runs to a minimum.

Those are my favorite things to look at when it comes to a pitcher.

I agree 100%. If you care about being precise (I realize some of us don't need to or don't have time to), then K/IP is a lazy stat. K/PA is much more precise, and it weeds out the chaff like Glendon Rusch, who for many years had a pretty good K/IP rate, but actually struck out guys less frequently than other pitchers who didn't get chance after chance at the major league level.

For offense, OPS is also a lazy stat. Not as lazy as batting average, but lazy nonetheless. Below is something I wrote on the topic many years ago. For shorthand stats, RC/27 is EASIER to visualize than batting average or OPS anyway. I can picture that if Joey Votto's RC/27 is 5.5, then a team of Joey Vottos would - almost certainly - score an average of 5.5 runs per game.

If you want to build a good team, then you have to look at scarcity. You start by calculating "average" levels and replacement levels. You want as many above-average guys as possible. When that is not economically feasible, you want guys who are as above replacement level as possible.

You then look at a guy's salary and figure out how much you're paying for every run above average and every run above replacement level. You then spend your money accordingly.

I am not good at math, but I don't think it's all that complicated.



How Orlando Palmeiro Got Hosed: Exactly How Full of S is OPS?
5/26/03

Michael Lewis’ new book Moneyball revisits the issue of the relative importance of acquiring players who get on base versus acquiring players who hit for power. Specifically, the book shows Oakland A’s Assistant General, Paul DePodesta, disparaging the stat of “On Base Plus Slugging Percentage (OPS),” because of its implicit prejudice against high on-base percentage (OBP) and low slugging percentage (SLG) players.

Close followers of sabermetrics should not be surprised to hear DePodesta state that the “O” in OPS is more important than the “S.” Rob Neyer, for example, has expressed in columns such as this one that, when precision is necessary, OPS should be weighted to approximately:

(1.4 x on-base%) + slugging%

Furthermore, when one evaluates the math in more complex (and accurate) run estimators [such as Clay Davenport’s Equivalent Average (EQA) and Bill James’ Runs Created Per 27 outs (RC/27)], one finds that high-OBP/low-SLG guys perform much better than low-OBP/high-SLG guys.

What makes DePodesta’s statement stand out as special is the degree to which he believes that the “O” is more important than the “S.” In DePodesta’s model for run production, “an extra point of on-base percentage was worth three times an extra point of slugging percentage” [p. 128, Moneyball, emphasis mine].

If DePodesta’s model for run production is accurate, then certain players are not only routinely getting undervalued by OPS (in arenas such as free agency, salary arbitration, and the court of public opinion), they are in fact routinely and overtly getting cheated by it.

How Much More Important is On-Base Percentage Than Slugging Percentage?

Since I know that many readers are “executive summary” type of people, I will first disclose what I found out in the bullet points below. If the words “regression analysis” don’t float your boat, then feel free to just read the bullet points and then skim until you see the next bold subtitle.

• Team on base percentage is, not surprisingly to statheads, more important than team slugging percentage
• The Neyer magic number of 1.4 for adjusted OPS,
[(1.4 x OBP) + SLG],
while adequate for most of the 1980s, is now too low
• The DePodesta magic number of 3 for adjusted OPS,
[(3 x OBP) + SLG],
while perhaps accurately reflecting the conditions of a few seasons ago, is now too high
• The relative value of team on base percentage vs. team slugging percentage has not remained constant over the past 25 years. Therefore, OPS should periodically be tweaked to reflect the fluctuating scarcity of OBP and SLG
• After evaluating the data carefully, the proper OPS adjustment for the conditions from 2001 to 2002 season should be approximately:
[(1.64 x OBP) + SLG]

Now, for those of you who are interested in the data...

One difficulty I had in this project was trying to reconcile the need for a significant sample size with the need to capture values of a small time frame. My first inclination was to run a regression equation for the 2003 season thus far. However, I was concerned that some statisticians would argue that an “n-30” (30 major league teams) would be too small and would be prone to biases. Sample size could also be a problem due to the fact that those 30 teams have played fewer than 50 games apiece as of May 23rd. On the other extreme, I considered looking at extensive history – such as the relationship between OBP, SLG and runs over the past 25 seasons. But I was concerned that such data would be unreliable due to the completely different offensive conditions that existed back when Oscar Gamble had a 12” ‘fro and out-machines like Omar Moreno were routinely written into the leadoff spots on major league lineup cards.

In my indecisiveness, I ran numerous regression equations to see if I was simply worrying too much. These were very basic linear regressions done in SPSS, under the following conditions:

Dependent variable: team runs scored per game (I realize that runs scored per inning would be slightly more accurate, but I didn’t have that data)

Independent variables (2): team OBP; team SLG

Here are the results:
Year OBP to SLG Ratio R - squared OBP multiplier SLG multiplier Constant F statistic
2003 (thru May 23) 2.07 0.909 20.96 10.10 -6.50 134.9
2002 0.82 0.878 12.17 14.80 -5.57 201.5
2001 3.51 0.908 24.81 7.08 -6.48 276.4
2000 4.81 0.913 28.06 5.84 -7.09 295.3
1999 1.75 0.934 19.09 10.88 -6.21 395.2
1998 4.11 0.928 25.39 6.18 -6.30 360.8
1997 2.95 0.927 21.77 7.39 -5.66 331.9
1996 1.07 0.921 12.97 12.11 -4.50 302.5
1995 2.12 0.924 19.01 8.98 -5.27 315.5
1982-1992 1.42 0.866 14.76 10.39 -4.54 1831.7
1993-2000 1.78 0.904 17.89 10.07 -5.41 2158.5
1970-2000 1.53 0.926 16.38 10.73 -5.20 10098.0
1998-2000 2.79 0.925 22.52 8.08 -6.16 1085.3
1996-2000 2.24 0.91 20.03 8.95 -5.66 1460.4

I was surprised by the fluctuation in the relative value of on base percentage and slugging percentage that was estimated in these equations. If these equations could be believed (and that’s a big if), SLG was more important than OBP in 2002, but OBP was 4.8 times more important than SLG in 2000. Given that all of the multi-year regressions seemed to indicate that OBP was more important than SLG, I figured that the 2002 regression was “wrong.” I theorized that one-year regressions simply lacked a big enough sample size. Thirty teams were not enough to tell a story.

According to Stat Soft Inc., “many researchers follow a rule of thumb that if your sample size is 50 or more then serious biases are unlikely.” If you are one who believes that an n=50 is too small, then I refer you to the previous table. Note that between 1996 and 2000 (an n=146), on base percentage was 2.24 times more “important” than slugging percentage. This suggests that Mr. DePodesta was on to something. 2.24 is not three. But it’s closer to three than one. Should we, therefore, re-write OPS to [(2.24 x OBP) + SLG)]? Not so fast.

Now, if you are one who believes that a sample size of 50 is adequate for this type of regression model, then you might prefer to see a regression equation that applies more closely to today. Who cares how important OBP was in 1996. Those findings don’t apply to the current game if the conditions of scarcity have changed.

For this reason, I ran regressions for every two-year period going back to 1977 (when baseball expanded to 26 teams). This way, every period would yield a sample size of 52 or greater. Without further ado, here are those results:


Year OBP to SLG Ratio R - squared OBP multiplier SLG multiplier Constant F statistic
2001-2002 1.64 0.882 17.85 10.87 -5.81 212.6
2000-2001 3.63 0.92 25.05 6.91 -6.50 329.8
1999-2000 2.37 0.917 21.76 9.18 -6.39 313.7
1998-1999 2.43 0.929 21.02 8.66 -5.90 374.7
1997-1998 3.26 0.923 23.02 7.07 -5.92 331.7
1996-1997 1.61 0.895 16.67 10.34 -5.09 226.2
1995-1996 1.49 0.92 15.95 10.67 -4.93 304.9
1994-1995 1.65 0.897 17.21 10.45 -5.29 231.3
1993-1994 1.02 0.893 13.55 13.23 -5.26 221.1
1992-1993 1.34 0.899 15.47 11.54 -5.25 227.0
1991-1992 1.59 0.852 16.76 10.55 -5.25 140.6
1990-1991 0.83 0.816 10.92 13.10 -6.14 109.0
1989-1990 0.51 0.841 6.78 13.27 -5.49 129.9
1988-1989 1.28 0.874 13.22 10.32 -4.00 169.4
1987-1988 2.22 0.903 19.28 8.68 -5.32 227.8
1986-1987 2.49 0.852 21.35 8.57 -5.97 147.7
1985-1986 2.17 0.856 19.07 8.79 -5.32 145.1
1984-1985 2.31 0.878 19.48 8.43 -5.31 176.6
1983-1984 2.06 0.892 17.29 8.41 -4.61 201.7
1982-1983 1.07 0.868 11.78 11.01 -3.83 161.2
1981-1982 0.97 0.836 11.57 11.92 -4.12 124.6
1980-1981 1.43 0.877 15.57 10.88 -5.04 173.9
1979-1980 1.86 0.927 18.02 9.71 -5.39 312.6
1978-1979 1.46 0.919 15.65 10.69 -5.02 277.4
1977-1978 1.76 0.932 17.81 10.10 -5.51 337.1
Mean 1.78
Median 1.64

The importance of on-base percentage shot up in the ’98-’99 time frame, just as McGwire and Sosa were eclipsing Roger Maris’ single season homer record, and just as 40 and 50 homer seasons were becoming more routine than ever before. Some of this can be explained by the close bunching of slugging percentage among the top scoring teams. For example, note below that the four best runs-per-game American League offenses in 1998 generated nearly identical slugging percentages, but the Yankees’ ability to put runners on base allowed them to score the most runs.

Team On Base% Slugging% Runs Per Game
Yankees ‘98 .362 .460 5.96
Rangers ‘98 .356 .462 5.80
Red Sox ‘98 .343 .463 5.41
Mariners ‘98 .341 .468 5.34

However, the relative importance of on base percentage regressed to “only” 1.64 during the 2001-2002 time frame. Is it possible that DePodesta’s run production model was based largely on the conditions of the late 1990’s, but needs to be readjusted for today? Perhaps. Note that the 2001-2002 ratio of 1.64 is right at the median of all the time frames surveyed since 1977. This suggests that today’s conditions are much more “normal” than those in the late 1990’s.

Nevertheless, there are some aspects of this data that are difficult explain. Despite my deeply held intuitive belief that on-base percentage is always more important than slugging percentage, the two-year time period chart shows three eras where SLG appears to be more important than OBP. Those eras are 1981-1982, 1989-1990, and 1990-1991. Oddly enough, these periods happened to produce the lowest “r-squared” totals. For those who haven’t taken a stats class, or who had one but didn’t pay attention, “r-squared” tells you how much is explained by the regression equation. So for example, in the 2001-2002 time period ‘runs scored’ were approximately 88.2% (r-squared of .882) determined by OBP and SLG. The other 11.8% can be explained by other stuff. This “other stuff” might include baserunning, clutch hitting, and number of times reaching base on an error. Why does a “low r-squared” period correspond with a period where SLG is important? My opinion is that these eras, 1989 in particular, are characterized by a great deal of offensive parity. For example, in the National League in 1989, every team besides Atlanta had an on base percentage in the range of .305 to .321. In the American League in 1989, all but two teams scored between 4.13 and 4.78 runs per game. But these years are anomalies, and hence, I would not recommend that Major League GM’s attempt to build an offense based on numerous low-OBP/high-SLG Dave Kingman types.

Finally, since some of you will undoubtedly want to test the validity of my prescribed 1.64 ratio of OBP to SLG, here is what happens when you apply the regression equation to 2003 (in games played through May 23rd). While 1.64 denotes the relative importance of OBP v. SLG, the exact regression formula is:

runs scored per game = (17.853 x OBP) + (10.874 x SLG) – 5.805

TEAM Actual Runs Per Game Estimated Runs Per Game Error, Runs/Game
Florida 4.29 4.82 0.53
Chicago Sox 3.98 4.34 0.36
Arizona 4.13 4.48 0.35
San Diego 3.83 4.15 0.31
San Francisco 4.81 5.07 0.26
Los Angeles 3.67 3.90 0.23
Houston 4.79 4.86 0.07
Minnesota 4.79 4.86 0.07
Texas 5.55 5.61 0.06
Milwaukee 4.46 4.51 0.05
NY Mets 4.17 4.20 0.03
Seattle 5.38 5.38 0.00
Cleveland 3.98 3.95 -0.03
Chicago Cubs 5.00 4.97 -0.03
Montreal 4.71 4.67 -0.04
St. Louis 5.41 5.37 -0.05
Anaheim 4.93 4.88 -0.06
Kansas City 4.85 4.77 -0.08
Pittsburgh 3.94 3.86 -0.08
Atlanta 5.85 5.74 -0.12
NY Yankees 5.71 5.58 -0.12
Colorado 5.42 5.27 -0.15
Tampa Bay 4.55 4.35 -0.20
Oakland 4.74 4.54 -0.20
Toronto 6.02 5.81 -0.21
Cincinnati 5.13 4.89 -0.23
Philadelphia 4.81 4.56 -0.25
Detroit 3.00 2.69 -0.31
Baltimore 4.72 4.35 -0.37
Boston 6.19 5.77 -0.42


Perhaps due to some very poor clutch hitting (or so my Marlin fan friend tells me), the Marlins are vastly underplaying their regression-projected runs per game total. However, most teams’ offensive fortunes are closely projected by the regression equation. Half of the teams’ projections are accurate within 0.15 runs per game, the equivalent of about 24 runs or fewer over a full season. Since this is really only about 30% of a season’s worth of data, this equation should probably be tried later in the season. It’s possible that outliers such as the Marlins will regress to the mean.

Market Inefficiencies Created When Actors in the Market Don’t Value “O”

The A’s belief in OBP over SLG, among other reasons, helps explain the club’s pursuit of Scott Hatteberg after the 2001 season. Moneyball suggests that the A’s were the only club, besides possibly the Rockies, who valued Hatteberg’s abilities enough to sign him. Hatteberg’s OPS totals in prior recent seasons were not impressive, and he could no longer throw well due to injury. However, the “O” in Hatteberg’s career OPS was attractive to the A’s. Hatteberg was only two years removed from a .367 OBP, clearly above the league mean of .345 that year. The A’s were more than happy to get a guy who could simply get on base and agree to a salary under $1 million.

Other than Scott Hatteberg, who are some other players in recent years that have been underappreciated due to the gross lack of emphasis of OBP in the OPS equation? Likewise, which guys have been the most overrated by OPS?

Just concentrating on 2002 statistics, I looked at every player who had at least 200 at bats during the season. I then calculated the standard deviations from the mean for each player’s OPS and subtracted that amount from the standard deviations from the mean for each of the 3 modified versions of OPS:

“DePodesta*” OPS = [(3 x OBP) + SLG]
“Neyer” OPS = [(1.4 x OBP)+ SLG]
“Chili’s ‘01-‘02 OPS” = [(1.64 x OBP) + SLG]

The results of those calculations follow the upcoming disclaimer paragraph.

It is important to point out that Moneyball does not reveal DePodesta’s complete model. The book only discloses the fact that OBP is worth approximately three times as much as SLG in DePodesta’s model. Certainly, this mysterious model contains more independent variables than OBP and SLG, and a hitter’s poor performance in these other categories might somewhat mitigate the fact that he has a very good OBP. There’s also a remote possibility, not one that I subscribe to personally, that some of the statements in Moneyball from DePodesta and General Manager Billy Beane are deliberate bits of misinformation, carefully selected to throw off sucker GM’s who wish to copy the Oakland model. Perhaps, DePodesta’s model actually does not value OBP. After all, the A’s did sign Chris Singleton this year!

Now back to the issue of who gets slighted or hyped by OPS:

DePodesta's most slighted Top 20 list OPS rank (out of 329) DePod rank (out of 329) Difference StDevMean DePodesta's most hyped Bottom 20 list OPS rank (out of 329) DePod rank (out of 329) Difference StDevMean
Palmeiro,Orlan. 193 110 0.552 Garcia,Karim 38 83 -0.655
Mclemore,Mark 124 72 0.469 Soriano,Alf. 49 89 -0.518
Castillo,Luis 194 128 0.460 Crede,Joe 76 124 -0.469
Alicea,Luis 309 282 0.414 Grissom,Marq. 73 107 -0.394
O'Leary,Troy 174 108 0.413 Anderson,Gar. 46 66 -0.387
Counsell,Craig 227 172 0.401 Hunter,Torii 57 88 -0.375
Suzuki,Ichiro 84 52 0.385 Gibbons,Jay 104 158 -0.361
Justice,David 111 74 0.373 Rodriguez,Alex 7 14 -0.356
Roberts,Dave 210 151 0.361 Fullmer,Brad 50 68 -0.347
Loretta,Mark 112 77 0.361 Valentin,Jose 105 155 -0.343
Bonds,Barry 1 1 0.350 Phelps,Josh 25 38 -0.339
Franco,Julio 176 122 0.349 Mabry,John 79 111 -0.338
Mientkiewicz,D. 159 104 0.341 Batista,Tony 152 206 -0.336
Dunn,Adam 63 36 0.331 Butler,Brent 237 281 -0.322
Jones,Chipper 14 8 0.330 Ordonez,Magg. 15 21 -0.320
Kendall,Jason 231 188 0.329 Taylor,Reggie 204 253 -0.314
Redmond,Mike 171 119 0.325 Ibanez,Raul 41 55 -0.312
Vazquez,Ramon 221 175 0.324 Paquette,Craig 328 329 -0.309
Williams,Bernie 31 19 0.322 Wells,Vernon 141 195 -0.304
Sanchez,Alex 230 187 0.318 Helms,Wes 246 283 -0.286

Neyer's most slighted Top 20 list OPS rank (out of 329) Neyer rank (out of 329) Difference StDevMean Neyer's most hyped Bottom 20 list OPS rank (out of 329) Neyer rank (out of 329) Difference StDevMean
Palmeiro,Orlan. 193 174 0.163 Garcia,Karim 38 47 -0.191
Mclemore,Mark 124 109 0.140 Soriano,Alf. 49 58 -0.150
Castillo,Luis 194 179 0.136 Crede,Joe 76 86 -0.137
Bonds,Barry 1 1 0.126 Grissom,Marq. 73 80 -0.115
O'Leary,Troy 174 152 0.122 Anderson,Gar. 46 51 -0.111
Alicea,Luis 309 304 0.118 Hunter,Torii 57 66 -0.108
Counsell,Craig 227 211 0.117 Gibbons,Jay 104 115 -0.106
Suzuki,Ichiro 84 77 0.117 Valentin,Jose 105 116 -0.101
Justice,David 111 100 0.112 Batista,Tony 152 170 -0.100
Loretta,Mark 112 103 0.108 Paquette,Craig 328 329 -0.100
Roberts,Dave 210 192 0.106 Fullmer,Brad 50 53 -0.099
Jones,Chipper 14 11 0.106 Mabry,John 79 87 -0.098
Franco,Julio 176 164 0.103 Butler,Brent 237 250 -0.098
Dunn,Adam 63 52 0.102 Rodriguez,Alex 7 9 -0.097
Mientkiewicz,D. 159 135 0.101 Phelps,Josh 25 32 -0.095
Williams,Bernie 31 27 0.101 Taylor,Reggie 204 223 -0.095
Redmond,Mike 171 153 0.096 Wells,Vernon 141 161 -0.090
Kendall,Jason 231 221 0.096 Truby,Chris 326 327 -0.089
Vazquez,Ramon 221 207 0.095 Ibanez,Raul 41 43 -0.089
Sanchez,Alex 230 217 0.093 Ordonez,Magg. 15 15 -0.088

Chili's most slighted Top 20 list OPS rank (out of 329) Dom rank (out of 329) Difference StDevMean Chili's most hyped Bottom 20 list OPS rank (out of 329) Dom rank (out of 329) Difference StDevMean
Palmeiro,Orlan. 193 161 0.244 Garcia,Karim 38 51 -0.286
Mclemore,Mark 124 100 0.210 Soriano,Alf. 49 62 -0.226
Castillo,Luis 194 171 0.203 Crede,Joe 76 92 -0.206
O'Leary,Troy 174 142 0.183 Grissom,Marq. 73 87 -0.172
Bonds,Barry 1 1 0.182 Anderson,Gar. 46 54 -0.167
Alicea,Luis 309 301 0.178 Hunter,Torii 57 69 -0.163
Counsell,Craig 227 205 0.176 Gibbons,Jay 104 120 -0.159
Suzuki,Ichiro 84 74 0.174 Valentin,Jose 105 121 -0.151
Justice,David 111 93 0.167 Fullmer,Brad 50 56 -0.150
Loretta,Mark 112 96 0.162 Batista,Tony 152 175 -0.150
Roberts,Dave 210 187 0.159 Rodriguez,Alex 7 9 -0.148
Jones,Chipper 14 10 0.156 Mabry,John 79 89 -0.148
Franco,Julio 176 152 0.154 Paquette,Craig 328 329 -0.147
Mientkiewicz,D. 159 127 0.151 Butler,Brent 237 254 -0.146
Dunn,Adam 63 50 0.151 Phelps,Josh 25 34 -0.144
Williams,Bernie 31 23 0.149 Taylor,Reggie 204 227 -0.141
Redmond,Mike 171 146 0.144 Wells,Vernon 141 166 -0.135
Kendall,Jason 231 214 0.144 Ibanez,Raul 41 46 -0.134
Vazquez,Ramon 221 201 0.142 Ordonez,Magg. 15 15 -0.134
Sanchez,Alex 230 211 0.139 Truby,Chris 326 327 -0.130


Note that the names on these lists are quite similar, but that the difference lies mainly in the degree to which Model DePodesta, Model Neyer, and Model Chili show that 2002 hitters were helped or hurt by the OPS stat. Some readers might not understand why Barry Bonds came up as underrated on all three lists, even though he had baseball’s highest OPS in 2002. What these lists demonstrate are that, while OPS may designate Bonds as best hitter in 2002, it does not adequately express just how much he blew away the competition.

The guy at the top of all three “most slighted” lists is Orlando Palmeiro. OPS unfairly brands Palmeiro as a below average hitter, ranking him 193rd out of 329 players. Conversely, DePodesta’s extrapolated model ranks Palmeiro at 110th out of 329 – significantly above average – while Neyer’s model and mine both rank him near the median of players. What’s more, 2002 wasn’t the first time that Palmeiro posted a very good on base percentage. He has done so in four of the past five years. Nevertheless, Palmeiro was not tendered a contract by the Angels following the 2002 season, and ultimately accepted a 1-year contract with the Cardinals worth $700,000 - a $300,000 paycut. While it has been well documented that Major League owners tightened up the purse strings this past off-season, it seems odd that Palmeiro could not find someone who valued his talents a little more. After all, even in this tight market, GMs managed to fork over $1.875 million to Marquis Grissom and $3.25 million to David Bell, just for 2003 alone.

Why Bother?

Serious statheads might wonder why I’ve wasted so many words and hours of research time on OPS, when there are clearly more sophisticated run estimators already out in the public and private domain. I’ve already referred to some of these, such as EQA and RC/27, the former of which accounts for park effects. Moneyball alludes to a complex system now at work in Oakland that breaks offensive production down into fragments, such as fly balls and ground balls, rather than singles, doubles, etc. Harvard statistics professor Carl Morris has apparently devised a formula that is “an exact calculation, not an estimate” of offensive production. Why should believers in these complex methods care about nuances in the shortcomings of OPS?

It’s because a stathead who ignores OPS is like a parent or schoolteacher who ignores Eminem and 50 Cent. Informed individuals sometimes have a vested interest in the deciphering the vernacular of even the most severely uninformed.

When average baseball fans hear the words “Carl Morris, Harvard statistics professor,” they run away in terror. Further, if Moneyball is to be believed, most Major League GMs do the same. The average baseball fan, and perhaps the average baseball executive, does not speak Carl Morris’ language. They speak OPS – or in some cases a language that is even more primitive. Consequently, the members of the statistical aristocracy who can best understand languages of the statistical proletariat will be the ones who best exploit the proletariat to their own advantage. Shrewd baseball executives can identify a team that believes in OPS and trade Karim Garcia for more than he’s worth, and acquire Orlando Palmeiro for less than he’s worth.

To draw an analogy, the Morris system is like an electron microscope. It’s arcane – few understand how to use it – but instruments of the sort are absolutely necessary when precision is essential. And with all the money at stake in pro baseball, precision should be essential. OPS is like your regular old microscope that you used back in high school chemistry class. It is quickly replacing batting average (the cracked magnifying glass) as the measuring instrument of the common man. But before OPS becomes any more pervasive than it already is, I would ask that everyone refocus their microscopes by simply multiplying the “O” by somewhere between 1.5 and 2. We should continue to do so until the next dramatic increase or decrease in the scarcity of power. Now, I realize the purpose of shorthand stats, like OPS, is to keep the math simple. But if we lack the mathematical ability to multiply by two, then we’re too stupid to participate in the class, and it’s time for the teacher to take our microscopes away.

BuckeyeRedleg
07-17-2008, 04:43 PM
Nah... let's talk about the aura of the game instead... :rolleyes:

Seriously though, sometimes these bickering stat v. scout fights get out of control. But, generally, I don't know what some of these posts are reffering to. I'd say at least 75% of the threads on here are limited to simple baseball discussion which do not require a degree in statistics. Meanwhile, it's the 20 page threads that do. Why? Becauses they are the ones that turn into stat v. scout struggles...


Agree. And there are people that enjoy such discussion and debate (hence the 20 page thread) and you'd think it would be easy to stay out of a thread like that if it brings you down. If someone just wants free information and updates on the team they can go here:

http://sports.espn.go.com/mlb/clubhouse?team=cin

It's like someone wants the board to do their work for them to get updated information, but then they don't want the board to discuss certain topics. Why can't you just take the good with the bad, as with all things in life? If something big happens, it will be stickied at the top anyway. There are threads I have zero interest in, but that doesn't mean I criticize the topic and personally, I think the topic of this thread was pretty clear from the beginning and it was not meant to stir up any debate or controversy.

I'm not sure I understand the hand-wringing. I also find it rude to say that someone needs to "get a life" if they post more than six times a day, especially when that person criticizing has thousands of posts themselves. It's kind of hard for me to sympathize with someone that gets personal and slams a large group of posters with their rant.

If you don't like it, don't read it and find a thread that you do like and participate in that. If it causes you that much stress in your life, then move on.

Baseball has always been about numbers. If you just like the game of baseball and don't like reading about numbers, then a message board dedicated to that team might not the the right place to be, because that's what people are going to discuss. That's one of the things that makes baseball so special. Thought-provoking debate with years of statistical data. On the other hand, it doesn't mean one can't appreciate baseball if they don't care about the numbers. The game is a beautiful work of art, numbers or not. It's just that some see the masterpiece and enjoy it for what it is and some enjoy it, but feel compelled to quantify it. That's all. It's just the difference in the way certain people think. Some are type-A and some are not, but I don't get complaining about people that are attempting to understand something deeper.

It's like going to a Mexican restaurant and complaining that the food is too spicy and that the restaurant should change their menu because you don't like spicy foods and it's not fair to you.

In my opinion you could do one of three things:

1. Learn to appreciate the spicy foods.
2. Order something less spicy on the menu.
3. Go to a different restaurant, that doesn't have spicy foods.

Baseball cannot be back soon enough.

Chip R
07-17-2008, 04:44 PM
I like the statistics that I can understand and I don't like the ones I don't understand. ;)

westofyou
07-17-2008, 04:49 PM
Baseball has always been about numbers. If you just like the game of baseball and don't like reading about numbers

Since the beginning they've been trying to measure it.

http://www.redszone.com/forums/showthread.php?t=52377




.

RedsManRick
07-17-2008, 04:57 PM
Explain.

Lets look at OPS it is OPB + SLG. That is not a lie. But what exactly is the intrepretation of OPS? If I say player X has an OPS of .800, what does that mean? The key with any statistical analysis is the intrepretation of that. True if you can not dispute the components of OPS but why those two components?

The RC stat. Granted it is a calculation that returns a number. But there is a human manipulation to that stat. Why are certain numbers used? Why are certain stats used and others not? Somewhere in another thread it was said that the RC stat was 95% accurate but when I calculated what the 5% difference was it accounted for 35 runs. Is that a cold hard fact or is that an intrepretation?

Why is Slugging based of of 4 instead of 1? A single in the forumla counts as 1 but in reality it is only a quarter of the way to a whole. In reality that single doesn't count as much unless it plates a run or the batter ends up scoring.

In all reality I could come up with this stat, BA+OBP+SLG and call it UPS. It would be pretty useless but would it lie? In reality a statistic is any combination of numbers. Some are good, some are bad, but when you take it as the absolute truth it must be 100% instead of that 95%. That 5% you are off can mean the differnece from a World Series Championship or sitting home during October, think St. Louis circa 2006.

Bucks, this is a great post and exactly the kind of questions I was hoping this thread would raise. I would suggest that the answer to your OPS question is that OPS is a measure of offensive production -- the higher the better. That's it. A guy with a higher OPS generally contributes more to your team scoring runs than a guy with a lower OPS.

As you've pointed out, a major problem with it in terms of not being widely used is that it doesn't have a widely known, obvious, convenient interpretation. It's fundamentally different than something like batting average or RBI, which is merely a tally of how often specific events have occurred. I call those things "stats". I would call OPS a "metric". It's a combination of stats meant to approximate something (overall run production in this case) that we can't easily measure with a single "stat". The same goes for RC, in principle.

We can't directly measure the effect of a single in runs, because it has value in a number of ways including maybe driving somebody in, maybe setting up the batter to score a run, creating an at bat for the guy behind him, etc. There's no one thing to measure. So if we want to understand the value of the single in terms of runs, we have to manipulate it. We do that by running regressions and the like, looking at how things tend to add up over time. Two singles can be of very different values. One single might drive in two runs and lead to a run scored. Another might be completely wasted. But over time, we can assess the average run value for singles in general.

The basic argument in favor of OPS over batting average is that for the purposes of player valuation knowing how good a guy is at producing runs (OPS) is more informative than just knowing how often he gets a hit when he doesn't walk.

Of course, if it's the bottom of the 9th, 2outs, down 1, with a man on 3B and we have the pitcher is up next, I might be more interested in a guy's batting average than his OPS. Again, this goes back to the issue of what question you're trying to answer. OPS doesn't attempt to answer the question of what a guy is likely to do this at bat.

Perhaps it be helpful to publish a handy little table that says that an 800 OPS over 600 PA is worth about 60 runs -- or whatever the math ends up being. Because at the end of the day, that's the thing we care about, scoring runs. Citing OPS makes it seem like OPS itself is the end. It's certainly not. It's just a pretty easy and pretty accurate way to assess run production. That way you could say, for example, that an 850 OPS is worth 15 more runs for your over the season than an 800 OPS.

What makes OPS better than batting average, to illustrate the point very generally is this (I'm making up the numbers for the purposes of explanation, but there are real numbers that we could put here):

Take any guy with an 800 OPS and he will have created somewhere between 50 and 70 runs. There's a 20 run variation there because OPS doesn't account for important things like the timing of events, base running, etc. So while it's a pretty good measure, it only gets us so far.

But take any guy with a .280 batting average, and he will have created somewhere between 20 and 90 runs.

OPS is "better" because it does a better job at approximating the thing which we'd like to know (how many runs does this guy create) but can't measure directly. You can't have an 800 OPS and be a quite unproductive or quite productive. With a 280 average, you can be either. And to be fair, OPS isn't really all that great of a metric either. As long as we're going to move from stat to metric, we're better off using something like Eqa, which would be even more precise yet.

durl
07-17-2008, 05:00 PM
Anybody remember the old Looney Tunes cartoon of when Foghorn Leghorn was trying to teach Egghead Jr. about baseball? Egghead started writing all kinds of formulas to explain how he accomplished stuff...

To each his own. I won't knock anyone for studying numbers because I've been known to do it myself. Basically, I just enjoy watching the game.

Spring~Fields
07-17-2008, 05:03 PM
Furthermore, the emphasis on stats has taken such a stranglehold on this board that the board has, to me at least, become almost unreadable. It's boring, repetitve and is dominated by a small group of people who seem to have a need to reduce everything to a finite number. Here's a hint: if you're posting more than 5 or 6 times a day you need to (IMO) get a life! Not only that, as in most things in life, you learn more by keeping your ears open and your mouth shut than the other way around.

Always Red
07-17-2008, 08:52 PM
Explain.

Lets look at OPS it is OPB + SLG. That is not a lie. But what exactly is the intrepretation of OPS? If I say player X has an OPS of .800, what does that mean? The key with any statistical analysis is the intrepretation of that. True if you can not dispute the components of OPS but why those two components?


Simply that OPS= OPB+SLG. It is what it is; a cold, hard fact. How you interpret it is all part of the analysis, which is all subject to debate. Which is what I was saying in my post.

You're right, you can make up new statistics and then analyze them until the cows come home. IMO, the truth is that there will never be any final answers; there will be no miracle stat that will wrap them all up in a box with a nice bow on top of it. Have we reached the point where there are already too many stats- too many soft stats? RC and WS come to mind, for me. I realize there are many here who will disagree with me on that. So be it.

SteelSD
07-17-2008, 11:15 PM
Runs Created and RC/27 Outs. Fantastic statistical representation of player performance. Correlates very highly to actual performance (higher correlations than OPS). Easy to create (even the most complex versions) with a simple Excel spreadsheet. Build it and you can figure out the relative values of an event- even GIDP, Strikeouts, and...yes...Base Hits versus Walks.

Runs Above Position. Thank you, Clay Davenport. Viewed at BP's "Equivalent Average" page, this is an excellent way to stack up players versus league average by position.

For those who like Batting Average, Equivalent Average breaks down performance to a BA-like metric.

Isolated Power (IsoP) is a good measure of how much power a player actually has and projects. But Isolate Discipline (IsoD) is probably the most overlooked offensive projection tool of all. If nothing else, these two make Batting Average useful for something.

For pitchers, DIPS, FIP, and xFIP are all far more useful than ERA to determine and project pitcher performance. BABIP and ball-in-play rates (GB%, FB%, LD%, HR/FB%, etc.) are great drill-downs that allow us to better understand how the game's inherent randomness may affect performance. K/9, BB/9, and HR/9 are the "trifecta" of pitcher projection statistics and we should note that standardizing by batters faced rather than IP (as stated earlier in the thread) can also be useful.

macro
07-17-2008, 11:19 PM
You could take that concept a bit further...the goal of a fan isn't necessarily winning the World Series...it's enjoyment. Perhaps a WS would provide the maximum enjoyment, but one could argue against. Imagine a superteam incapable of losing...wins every game by a blowout. How much fun would that be?

We should construct a "joy index", which would have as variables:

Win-Loss record (wins ARE usually more fun)
Season Contention (are we IN it?)
Close games/walkoffs
Brawls won/lost
Stadium giveaway quality
Announcer enthusiasm
Milestones/records attained or thwarted (in case of negative records).
and so on...

If the team doesn't win it ALL (as only one team can)...at least one can say if the season was enjoyable or not...

Isn't that the goal of all entertainment? To entertain?

Hey, this is my favorite post of the thread. Case in point, the 1985 season may go down as my favorite Reds season ever. Coming off three dismal seasons, low expectations, Pete's first full season back in a Reds uniform, Tom Browning's 20-win rookie season, Dave Parker's MVP-like season, 4192, and hanging in the pennant race down to the last few days of the season.

It was the light at the end of the dark tunnel of 1982-84, and it was a fun season.

On the other hand, the Reds won the division by 15 games and made the LCS in 1995, but I don't recall much enjoyment from that season, including even the sweep of the Dodgers in the LDS.

AmarilloRed
07-18-2008, 01:27 AM
Let's divide it into all 3 categories: Hitting, Fielding, Pitching. Each has its own statistics, and I prefer the older statistics.

Hitting: BA, HR, RBI. BA is still a good indicator of how efficient a hitter a player is. HR and RBI will shows us how well that player can produce, and a player with a good BA has a very good chance of getting the hits to drive those runners in.

Fielding: Errors and Fielding Percentage. Both show how well a particular player is at catching the ball, although some positions will naturally be positions that will have more errors. I think the newer fielding statistics have merit, but I am not fully convinced on their usefulness just yet.

Pitching: I have not made a good determination yet on what would be considered good pitching statistics just yet, although strikeouts,hits ,and walks, can all be very good statistics.

Patrick Bateman
07-18-2008, 01:45 AM
Fielding: Errors and Fielding Percentage. Both show how well a particular player is at catching the ball, although some positions will naturally be positions that will have more errors. I think the newer fielding statistics have merit, but I am not fully convinced on their usefulness just yet.


Those 2 metrics are the same thing. They also fail to show the most important aspect of a player's fielding ability, range.

mth123
07-18-2008, 07:08 AM
I like to see a little more of the elements of performance than the all encompassing stats like OPS, OPS+, FIP and XFIP.

For Hitters:

BA/OBP/SLG and for fun I still like HR, RBI and Runs Scored.

For Pitchers:

ERA coupled with K/9, BB/9, HR/9, BABIP. ERA by itself tells me little.

Defense:

Don't like any of them. My eyes are more dependable at this point.

Baserunning:

All are too simplistic for my taste. CS% ignore times picked-off and there isn't much about out there on how many outs guys make on the bases. Making outs on the bases is my number one peeve when watching a ball game. Some one needs to invent a Freel meter. It would be akin to the "Mendoza Line" as an indication of awful. Mendoza line for Hitting, Freel Meter for Baserunning.