PDA

View Full Version : Pythagoras smiles



nate
05-12-2009, 09:54 AM
As everyone's favorite announcer, George Grande would say, we have a smiling Pythagoras. The Reds really started turning on their differential since bottoming out on 4/28 of this year. It's almost been a straight climb up to the surface since that time.

Anyhow, here's a little graphic that should warm the cockles of ol' "Py" (and Raisor):

http://www.gluethemoose.com/neen/redszone/RedsRunDif090511.png

Go team!

Homer Bailey
05-12-2009, 10:10 AM
You guys like Pythagoras a lot more than I do, but maybe it's because I just don't understand it that well. Does the fact the Reds piled on a bunch of runs last night when they already had the game in hand make them a better team? Does it mean that they will win more games? I just don't see it.

I understand that the pythag predicts what a teams record would be based on their run differential, but I think it is a very overrated "stat," and isn't a great predictor of how teams play in close ball games, do they score runs when it counts, etc.

I'm sure many will have differing opinions.

BCubb2003
05-12-2009, 10:32 AM
I like to scoff at Pythagoras but its track record is remarkable. Not on a game-by-game basis, but it's a good unsentimental look at how the team is actually playing.

I always imagine that there might be a team that has a low-scoring offense and three excellent starting pitchers, and a couple of very bad ones. They win low-scoring games, and lose by wide margins. But Pythagoras comes through time and again.

One thing you can be sure of is that tomorrow, the number will be different. But over time, it will tell you if the team is playing over their heads or not.

nate
05-12-2009, 10:56 AM
You guys like Pythagoras a lot more than I do, but maybe it's because I just don't understand it that well. Does the fact the Reds piled on a bunch of runs last night when they already had the game in hand make them a better team? Does it mean that they will win more games? I just don't see it.

I guess I'd answer your question with a question. Is a team more likely to win games when it typically scores more runs than the opposition or is it more likely to win games when it typically scores less runs than the opposition?

Lopsided wins offset lopsided losses, that's why there are averages.


I understand that the pythag predicts what a teams record would be based on their run differential, but I think it is a very overrated "stat," What's overrated about it?


and isn't a great predictor of how teams play in close ball games,I don't know if we're talking about it being a predictor. I think we're talking about it being "this is what it is."


do they score runs when it counts, etc. That's why you compare it to the actual record.

I think scoring more runs than the other team is a good thing to sustain.

dougdirt
05-12-2009, 10:57 AM
You guys like Pythagoras a lot more than I do, but maybe it's because I just don't understand it that well. Does the fact the Reds piled on a bunch of runs last night when they already had the game in hand make them a better team? Does it mean that they will win more games? I just don't see it.

I understand that the pythag predicts what a teams record would be based on their run differential, but I think it is a very overrated "stat," and isn't a great predictor of how teams play in close ball games, do they score runs when it counts, etc.

I'm sure many will have differing opinions.
Does it make them a better team? No. Does it mean they have a better chance at a winning record and the playoffs? The numbers say yes. From 1970-2007 91% of teams with a winning record have a positive run differential. 98% of division winners have had a positive run differential. So while piling on some runs after the win was at hand didn't really make a difference, it increased the Reds odds at doing something good on the season.

Homer Bailey
05-12-2009, 11:08 AM
I guess I'd answer your question with a question. Is a team more likely to win games when it typically scores more runs than the opposition or is it more likely to win games when it typically scores less runs than the opposition?

Clearly you know the answer to this question.


Lopsided wins offset lopsided losses, that's why there are averages.

What if a team loses 5 games by 10 runs, but plays well in close games by scoring clutch runs, having a setup man and closer, etc. A lot of people attribute luck to winning or losing 1 run games, but I personally don't think there is as much luck involved as some think.


What's overrated about it?

How is winning that game last night 13-5 vs. winning it 7-5 any different? If anything I think it "hurts" the Reds over the long term to score 13 runs in a game, because it just means that the likely regression to our averages is going to be much more steep (if that made any sense). It's hard to put what I'm saying into words on a computer.


I don't know if we're talking about it being a predictor. I think we're talking about it being "this is what it is."

This goes back to me not really understanding what it is.


That's why you compare it to the actual record.

I guess I don't understand the point of the comparison. If we win 13-5 or 7-5, our record is still the same. In hindsight, those 6 runs meant nothing. However, if I understand it correctly, the Pythag does not factor when then runs are scored into its formula. As it turns out, those runs had no effect on our actual record, but it will on our Pythag. That's why it just doesn't make much sense to me.


I think scoring more runs than the other team is a good thing to sustain.

Clearly.

jojo
05-12-2009, 11:22 AM
Too many things can happen to influence a 1 run difference for a team to have such games be their bread an butter. The BRM of '75 and '76 did have winning records in 1 run games but even their winning percentage was significantly worse in such games than their overall record. Same thing for the '27 and '28 Yanks etc...


A lot of people attribute luck to winning or losing 1 run games, but I personally don't think there is as much luck involved as some think.

What evidence has led you to this conclusion? I'm intrigued.

Falls City Beer
05-12-2009, 11:25 AM
Reds are 4-4 in one-run games this season so far. FTR.

Homer Bailey
05-12-2009, 11:54 AM
What evidence has led you to this conclusion? I'm intrigued.

I don't have any evidence, and never claimed to. I think there are many factors that can lead to success in one run games, most notably a strong bullpen. "Clutch" hitting has been often discussed on here, and some have argued whether or not such a thing exists, but I believe that timely hitting can factor into one run games. I agree that there is a certain level of "luck" or law of averages when it comes to 1 run games, but I think the ability to win those types of games says more about the ability of a team than a blowout win or loss does. And that is just my opinion. I'm sure others feel otherwise.

Homer Bailey
05-12-2009, 11:58 AM
Does it make them a better team? No. Does it mean they have a better chance at a winning record and the playoffs? The numbers say yes. From 1970-2007 91% of teams with a winning record have a positive run differential. 98% of division winners have had a positive run differential. So while piling on some runs after the win was at hand didn't really make a difference, it increased the Reds odds at doing something good on the season.

Clearly a team that has a large run differential is going to win more games. My "argument" is that I don't see much need to celebrate having a positive run differential because we tacked on a bunch of runs in a blowout game. If we ran off a bunch of 3-4 run wins in a row, and our run differential went up by 12-15 runs, then I could see that being something to be excited about.

acredsfan
05-12-2009, 12:01 PM
As with any mathematical statistic, nothing is 100% accurate, but when you are dealing with accuracy well over 90% then that is not an "overrated" or "useless" stat. Sometimes it's hard to accept that a simple equation can paint such a clear picture of what is going on, but when you have a season with 162 games, that is more than enough of a sample size to paint a clear picture. Teams generally won't only win games by a few runs and get blown out in every one of their losses or vice versa.

nate
05-12-2009, 12:03 PM
Clearly a team that has a large run differential is going to win more games. My "argument" is that I don't see much need to celebrate having a positive run differential because we tacked on a bunch of runs in a blowout game. If we ran off a bunch of 3-4 run wins in a row, and our run differential went up by 12-15 runs, then I could see that being something to be excited about.

If you look at the chart, that's basically what we've been doing since the 4/28.

nate
05-12-2009, 12:05 PM
I'm not arguing that this is a magic 8-ball. I'm saying the Reds are now, on average, scoring more runs than their opponents. Doing that, as Doug pointed out, leads to a winning record over 90% of the time. I don't treat this as predictive, I treat it as "right now" and "right now" I see a team that's winning by scoring more runs than their opponents rather than lucking their way into a bunch of wins.

Homer Bailey
05-12-2009, 12:12 PM
If you look at the chart, that's basically what we've been doing since the 4/28.

I realize that, and that is obviously a good thing. We can do this run around all day long. I agree that Pythagoras is very accurate in predicting a team's record, and teams with positive run differentials are generally going to be much more successful. The difference is that I don't celebrate big margins of victory ("Pythag, here we come!") nor do I fret about it if we are gonna be blown out ("Kiss Pythag goodbye"). The quotes in parentheses are things I read on here (not actual quotes, just summarized), and they are things I just don't worry about.

edabbs44
05-12-2009, 12:14 PM
At what point in the season does this theory become a solid predictor of the rest of the season?

Homer Bailey
05-12-2009, 12:18 PM
I'm not arguing that this is a magic 8-ball. I'm saying the Reds are now, on average, scoring more runs than their opponents. Doing that, as Doug pointed out, leads to a winning record over 90% of the time. I don't treat this as predictive, I treat it as "right now" and "right now" I see a team that's winning by scoring more runs than their opponents rather than lucking their way into a bunch of wins.

And I agree 100% with you there. I didn't mean to accuse of saying that it is a magic 8-ball. I understand the desire to have a high run differential. My point is that tacking on a bunch of runs in a blowout, or having a pitcher pitch the 9th inning and give up 5 runs, isn't necessarily indicitive of the overall success of a team. I also understand that over 162 this will even out.

nate
05-12-2009, 12:30 PM
At what point in the season does this theory become a solid predictor of the rest of the season?

October?

I don't see it as a predictor.

nate
05-12-2009, 12:33 PM
I realize that, and that is obviously a good thing. We can do this run around all day long. I agree that Pythagoras is very accurate in predicting a team's record,

I'm not using it to predict. I'm using it to say, based on run scoring and prevention, we're were we should be.


and teams with positive run differentials are generally going to be much more successful. The difference is that I don't celebrate big margins of victory ("Pythag, here we come!") nor do I fret about it if we are gonna be blown out ("Kiss Pythag goodbye"). The quotes in parentheses are things I read on here (not actual quotes, just summarized), and they are things I just don't worry about.Well, those games count too.

BTW, anyone notice the "streakiness" in that pictures?

jojo
05-12-2009, 12:44 PM
I realize that, and that is obviously a good thing. We can do this run around all day long. I agree that Pythagoras is very accurate in predicting a team's record, and teams with positive run differentials are generally going to be much more successful. The difference is that I don't celebrate big margins of victory ("Pythag, here we come!") nor do I fret about it if we are gonna be blown out ("Kiss Pythag goodbye"). The quotes in parentheses are things I read on here (not actual quotes, just summarized), and they are things I just don't worry about.

The whole celebrating "pythag" around here is mostly tongue in cheek. Heck I bet less than half of Raisor's 53 nobel prizes were actually related to pythag....

BearcatShane
05-12-2009, 01:15 PM
With a +2 run differential right now, what should the Reds record be?

jojo
05-12-2009, 01:15 PM
With a +2 run differential right now, what should the Reds record be?

16-16

Ravenlord
05-12-2009, 01:20 PM
With a +2 run differential right now, what should the Reds record be?

16-16 by the Therom. be interesting to see where it is at the end of May.

VR
05-12-2009, 01:37 PM
All I see is a smiley face when I look at that chart. I guess that a good thing. :)

MississippiRed
05-12-2009, 02:19 PM
Clearly a team that has a large run differential is going to win more games. My "argument" is that I don't see much need to celebrate having a positive run differential because we tacked on a bunch of runs in a blowout game. If we ran off a bunch of 3-4 run wins in a row, and our run differential went up by 12-15 runs, then I could see that being something to be excited about.

We have a saying at our house, and that is "Two out base hits will get you into heaven." Not sure where we got it, but those sure are nice in the one-run games.

Sorry, I meant to quote the other post by Homer Bailey about clutch hits.

osuceltic
05-12-2009, 02:24 PM
2008: 12-20 record, 132 runs scored, 164 runs against
2009: 18-14 record, 145 runs scored, 143 runs against.

Also: Reds record after a loss: 11-3

westofyou
05-12-2009, 02:25 PM
Every season this greek's name comes up and every season someone says it's a load, or specious at best.

Nevertheless the reaction to the theory is getting less and less angry, five years ago it was pretty humorous.

Homer Bailey
05-12-2009, 02:31 PM
Every season this greek's name comes up and every season someone says it's a load, or specious at best.

Nevertheless the reaction to the theory is getting less and less angry, five years ago it was pretty humorous.

I'm not saying it's a load. I just don't get the excitement over winning one game by a bunch of runs.

That being said, the Reds recent surge of winning (thus higher run differential over the last few games) is worth noting.

BRM
05-12-2009, 02:33 PM
Every season this greek's name comes up and every season someone says it's a load, or specious at best.

Nevertheless the reaction to the theory is getting less and less angry, five years ago it was pretty humorous.

I remember one thread in particular that was very humorous. Raisor became an even bigger legend in that one.

nate
05-12-2009, 03:50 PM
I'm not saying it's a load. I just don't get the excitement over winning one game by a bunch of runs.

That's not what the "excitement" is over. Nor is it about a single game.


That being said, the Reds recent surge of winning (thus higher run differential over the last few games) is worth noting.

That's what the original post was about.

nate
05-12-2009, 03:51 PM
2008: 12-20 record, 132 runs scored, 164 runs against
2009: 18-14 record, 145 runs scored, 143 runs against.

Also: Reds record after a loss: 11-3

That's an interesting stat there too. It used to seem like the Reds would often lose 2-3 in a row, win 1, lose 2-3. This year, it seems like win 2-3 in a row, lose 1, a nice change.

Homer Bailey
05-12-2009, 03:59 PM
That's not what the "excitement" is over. Nor is it about a single game.



That's what the original post was about.

I guess my posts were more directed to people that fret about the Pythag when we are getting killed in a particular game, and "cheer" when we are making our way back by killing a team (last night).

VR
05-12-2009, 04:11 PM
That's an interesting stat there too. It used to seem like the Reds would often lose 2-3 in a row, win 1, lose 2-3. This year, it seems like win 2-3 in a row, lose 1, a nice change.

It's nice to see the nightly pitching matchups and see that the Reds typically have the advantage 3/5 games....and when they don't have the edge, it's not a runaway. It seems the bullpen doesn't spend the 1st 5 innings in the bullpen filling up gas cans and creating molotov cocktails in preparation for their nightly bonfire.

Even tonights game....with our #5 vs. their #2, will be an entertaining game.

I go back to the reminder that 5 years ago, Micah Owings would be the ace of that staff....and it wouldn't really be close.

Johnny Footstool
05-12-2009, 04:46 PM
If you look at historical data, teams with positive overall W-L records tend to have higher winning percentages in blowouts. Teams with negative overall W-L records tend to have very low winning percentages in games decided by a lot of runs. As margin of victory decreases, W-L records tend to approach .500. There are outliers, to be sure, but that's basically how it goes.

Basically, good teams win by a lot, and when they do lose, it's usually a close game. Bad teams tend to get blown out, and when they do win, they squeak by.

So it makes sense to get excited at big wins and get discouraged by big losses, not simply for the sake of Pythagoras, but in recognition of what those kinds of wins might be saying about the talent level of your team.

nate
05-12-2009, 04:52 PM
I guess my posts were more directed to people that fret about the Pythag when we are getting killed in a particular game, and "cheer" when we are making our way back by killing a team (last year).

I think it can be concerning when the Pythag record radically disagrees with the real record. This is where we get into "luck," un- and difficult to measure things.

Ltlabner
05-12-2009, 05:02 PM
I think it can be concerning when the Pythag record radically disagrees with the real record. This is where we get into "luck," un- and difficult to measure things.

That's where I think it's a very useful tool. It's one way to gauge if a poorly performing team could potentially have a bounce back. It's one way to reality check yourself if you're team's on a hot streak up to the ASB but it's possible they are going to crash back to earth.

It's a macro tool, not a micro.

Homer Bailey
05-12-2009, 05:12 PM
I think it can be concerning when the Pythag record radically disagrees with the real record. This is where we get into "luck," un- and difficult to measure things.

Just an FYI, I meant last night, not last year. I edited it. Not that it matters.

medford
05-12-2009, 05:48 PM
I think it can be concerning when the Pythag record radically disagrees with the real record. This is where we get into "luck," un- and difficult to measure things.

I think that gets into the heart of the debate. When your team is doing well, nobody wants to look at a stat that on the surface seems to mean very very little and hear how their team is "lucky". Afterall, Luck is where hard work meets preperation :p:

the 90% success rate of this stat is what's most interesting, but when thought about it makes sense. I doubt you'd find a similar correlation in football, as well as basketball (to a smaller extent), but over the course of 162 games things even out and all teams tend to play to their ability to score runs, vs their ability to defend runs.

The other issue, is that its mearly a tracking stat of the past, but can't really incorporate trades or injuries that may greatly drastic a team's makeup over the coarse of several months than it was early in the year. Get to the end of the season and you tend to see the whole pictures. Its pretty simple, good teams score runs and prevent them, bad teams give up runs and fail to score them, and that is all the Pythag theory measures and the reason its so accurate over a 162 game span.

Ravenlord
05-12-2009, 06:02 PM
if memory serves, aren't the D-Backs of last year (or maybe two years ago) and the '81 Brewers the only teams to make the playoffs while being outscored by at least 100 runs?

TheNext44
05-12-2009, 06:55 PM
That's where I think it's a very useful tool. It's one way to gauge if a poorly performing team could potentially have a bounce back. It's one way to reality check yourself if you're team's on a hot streak up to the ASB but it's possible they are going to crash back to earth.

It's a macro tool, not a micro.

That's the most important aspect to remember about the Pythag. Even though it is accurate 90-95% of the time, that means that it wrong on average by 4-8 wins each 162 game season for each team. And since it is dead on many times, that means that there are times that it is over 10 games off, like last years Angels and the D-back's the year before.

That's accurate enough to make general observations like what Nate was referring to, but not accurate enough to be predictive about things like who will make the playoffs or the final standings.

If people are using it to make such predictions, then they are overrating it. I just haven't seen anyone on this board do that.

nate
05-12-2009, 07:27 PM
if memory serves, aren't the D-Backs of last year (or maybe two years ago) and the '81 Brewers the only teams to make the playoffs while being outscored by at least 100 runs?

The 2007 D-Backs sent ol' Py a -20.

traderumor
05-12-2009, 07:28 PM
if memory serves, aren't the D-Backs of last year (or maybe two years ago) and the '81 Brewers the only teams to make the playoffs while being outscored by at least 100 runs?And if that is so with the D-Backs, it goes to show that, while you may have been able to beat the odds one year, you better not think you have a winner going forward with the same guys.

jojo
05-12-2009, 08:59 PM
Bavasi says pythag smythag.....

Raisor
05-13-2009, 12:18 PM
I remember one thread in particular that was very humorous. Raisor became an even bigger legend in that one.

http://www.redszone.com/forums/showthread.php?t=37424

BRM
05-13-2009, 12:22 PM
http://www.redszone.com/forums/showthread.php?t=37424

That's the one. A true classic! :thumbup:

nate
05-13-2009, 12:40 PM
http://www.gap-system.org/%7Ehistory/BigPictures/Pythagoras_7.jpeg

SMcGavin
05-13-2009, 01:17 PM
http://www.redszone.com/forums/showthread.php?t=37424

Awesome thread... I especially like the first post in response to BF in that thread:


This will be a funny thread when it is brought back up in about a month.

Maybe the most accurate prediction in the history of Redszone?

jojo
05-13-2009, 02:01 PM
Awesome thread... I especially like the first post in response to BF in that thread:



Maybe the most accurate prediction in the history of Redszone?

The real irony is that the premise of the thread was hopelessly flawed dooming the thread before the first response. Even if the Nats would've had a similar second half that year, it wouldn't be proof that pythag is invalid.

Ltlabner
05-13-2009, 03:59 PM
Who knew that a key component of playing baseball was to play angry and direct said anger at winning?

No wonder BPhil is struggling...too many smiles.

Edit: If you play angry and run really fast to 1B when walked, does that make you a shoe-in for the HOF ?

kpresidente
05-13-2009, 04:30 PM
Guys, don't get too smug. If you're right about Pythagoras, that means your reward is the Reds really aren't that good.:(

westofyou
05-13-2009, 05:55 PM
Guys, don't get too smug. If you're right about Pythagoras, that means your reward is the Reds really aren't that good.:(

You mean smug like some of the responses in this thread?

http://www.redszone.com/forums/showthread.php?t=11888

VR
05-13-2009, 05:56 PM
Guys, don't get too smug. If you're right about Pythagoras, that means your reward is the Reds really aren't that good.:(

Right....but they are a LOT better in Pythag's eyes than they were a two weeks ago.

jojo
05-13-2009, 06:02 PM
Right....but they are a LOT better in Pythag's eyes than they were a two weeks ago.

Actually, not really.

nate
05-13-2009, 06:12 PM
You mean smug like some of the responses in this thread?

http://www.redszone.com/forums/showthread.php?t=11888

Ooof, that's a humdinger.

WebScorpion
05-14-2009, 02:24 AM
You people do know that Pythagoras died about 2500 years ago and he was probably one of the first psycho cult leaders. Punishment for breaking an imposed silence was said to be death. :eek::eek:

Mario-Rijo
05-14-2009, 05:00 AM
You mean smug like some of the responses in this thread?

http://www.redszone.com/forums/showthread.php?t=11888

And all this time I thought M2 had some sort of grander meaning. ;)

RichRed
05-14-2009, 11:09 AM
Pythagoras: The first Stat G(r)eek

May he continue to smile on our Redlegs.

Falls City Beer
05-14-2009, 11:40 AM
Actually, not really.

Maybe not a lot, but better. A slow climb can be accomplished in a 6 month season, provided no injuries.

nate
05-14-2009, 11:52 AM
Actually, not really.

I had it at around a 2 game over performance as of today. But due to the relatively small number of games, that's a good number of percentage points.

membengal
05-14-2009, 11:58 AM
Nate: mlb.com's standings have it at that as well. X W/L of 18-16. So they are within range.

D-Man
05-14-2009, 08:25 PM
What's missing here is that the Reds pitching staff has had shutouts in 18 percent of their starts (6 of 34 games). In addition, they've given up one run in five other games. Overall, they've given up one or zero runs in a third of their contests. . . That's huge, when it comes to wins and losses.

If that sort of pitching performance continues throughout the season, the "normal" rules of run assignment become irrelevant.

bucksfan2
05-15-2009, 08:38 AM
Two questions.

1. As the post-steroid era begins to come along do you think that the pythag win number will decrease from 10? It just seems as power numbers decline less runs will be scored in a game. Therefor I would assume the number would come down from 10runs=1win to some lower run total. With that lower run total wouldn't the variation become a little greater?

2. Does the pythag do anything with outliers or do they just hope that the outliers will of set themselves? Every team has "those" types of games. The game in which the starter gives up 8 runs in the first couple of innings, is run, and the opposing team scores 15+ runs in a game. It is just the nature of playing a 162 games season, every team is going to have an off day occasionally. My assumption would be that pythag would just assume (if figures can assume) that one bad lose will be off set with on big win. However over the course of a season lets say a team sufferend 4 10+ run losses but only scored 1 10+ win. That would have the potential of skewing the pythag. Also wouldn't it stand to make sense that a maximum run differential be credit to any win? If you put the cap at 8 runs/victory wouldn't that stand to make the pythag a little more accurate?

nate
05-20-2009, 11:29 AM
Pythag is sated.

Reds record: 20-18
Pythag record: 20-18

Ron Madden
05-21-2009, 05:01 AM
Pythag is sated.

Reds record: 20-18
Pythag record: 20-18


You stat geeks just twist and turn numbers until they support your agenda.

;)

nate
05-21-2009, 09:11 AM
You stat geeks just twist and turn numbers until they support your agenda.

;)

Oh yeah!!??

:cool:

nate
09-30-2009, 09:41 AM
http://www.gluethemoose.com/neen/redszone/pythag.png

nate
10-05-2009, 10:50 AM
http://www.gluethemoose.com/neen/redszone/pythag_0910050947.png

Pythagoras rests satisfyingly.

BCubb2003
10-05-2009, 11:55 AM
I like Pythagoras and what he stands for, but I wonder what use it has. It ought to be able to tell us when a team is playing above itself, and don't expect it to last. The wildly fluctuating early part of the chart is mostly from the small sample size of early games. The gap in the middle part of the season is interesting. Was the team playing out of their heads? By the end of the season, the pythag rose to meet this level. What can pythag tells us besides a truism at the end of season?

Falls City Beer
10-05-2009, 11:57 AM
What can pythag tells us besides a truism at the end of season?

It's not really predictive. Any more than taking someone's temperature is predictive.

edabbs44
10-05-2009, 12:16 PM
I like Pythagoras and what he stands for, but I wonder what use it has. It ought to be able to tell us when a team is playing above itself, and don't expect it to last. The wildly fluctuating early part of the chart is mostly from the small sample size of early games. The gap in the middle part of the season is interesting. Was the team playing out of their heads? By the end of the season, the pythag rose to meet this level. What can pythag tells us besides a truism at the end of season?

My thoughts exactly. Besides being a "neat" tool, I'm not sure how it benefits anyone.

nate
10-05-2009, 12:38 PM
I like Pythagoras and what he stands for, but I wonder what use it has. It ought to be able to tell us when a team is playing above itself, and don't expect it to last. The wildly fluctuating early part of the chart is mostly from the small sample size of early games. The gap in the middle part of the season is interesting. Was the team playing out of their heads? By the end of the season, the pythag rose to meet this level. What can pythag tells us besides a truism at the end of season?

A few notes:

After game 5, the Reds never underplayed their Pythag.

After the first few weeks (eyeballed at around game 36), the largest difference between Pythag and actual record was .058.

The least amount of difference between the two after game 36 was .002.

On average, the difference between Pythag and actual record was .025.

The best Pythagorean "record" the Reds had all year was .534. The best "real" record (after game 36): .571.

Final winning %: .481
Pythag: .464

BCubb2003
10-05-2009, 12:48 PM
It's not really predictive. Any more than taking someone's temperature is predictive.

This is where I sound naive and confused, but don't we already have the winning percentage?

Falls City Beer
10-05-2009, 01:04 PM
This is where I sound naive and confused, but don't we already have the winning percentage?

Yeah, I don't know. Maybe there's something to be found among teams that have bucked the Pythagorean. But maybe not. I don't disagree; as a tool, it's basically useless, or at best, simply an affirmation of the old saw: get me a bigger sample size.

nate
10-05-2009, 01:04 PM
This is where I sound naive and confused, but don't we already have the winning percentage?

I think it can give you an idea of your team's "luck" during the season. If you're really overplaying your pythag, you might not be as good as your record would indicate. If you're underplaying it, you might be better than your record.

Some might even say it measures "clutch."

Not going there!

:cool:

BCubb2003
10-05-2009, 01:23 PM
I think it can give you an idea of your team's "luck" during the season. If you're really overplaying your pythag, you might not be as good as your record would indicate. If you're underplaying it, you might be better than your record.

Some might even say it measures "clutch."

Not going there!

:cool:

That's what I'd like it to mean, but I still have my doubts. Is there a point in that season where you can derive something like that? Maybe all it says is that we had some blowouts, then we didn't. And isn't it inevitable that there's lots of variation early, which closes in the longer you go? Just like winning percentage? (If you're 0-1 and you win a game, you jump .500 points. Later in the season, not so much.)

And clutch is just the absence of choke.

nate
10-05-2009, 01:36 PM
That's what I'd like it to mean, but I still have my doubts. Is there a point in that season where you can derive something like that?


Maybe all it says is that we had some blowouts, then we didn't.

I think it correlates very highly with winning percentage over the years. We "were blown out by" and "blew out" several games this year and yet the pythag and actual winning percentage are within a couple of games or .017 percentage points.


And isn't it inevitable that there's lots of variation early, which closes in the longer you go? Just like winning percentage? (If you're 0-1 and you win a game, you jump .500 points. Later in the season, not so much.)

Yes, but the same goes for any "rate" statistic.

IslandRed
10-05-2009, 01:44 PM
As an analytical tool, I don't think it's very useful by itself. But I think the concept of a relatively stable relationship between RS-RA and wins was a prerequisite for a lot of the statistical analysis that followed. There isn't much point of doing all the run-based metrics for individual players without that tieback to team performance.

edabbs44
10-05-2009, 01:45 PM
Final winning %: .481
Pythag: .464

Is that good?

jojo
10-05-2009, 02:15 PM
Is that good?

Kansas City and Pittsburgh are jealous....

edabbs44
10-05-2009, 02:17 PM
Kansas City and Pittsburgh are jealous....

I meant the 3 game differential.

Johnny Footstool
10-05-2009, 02:28 PM
It's a tool that can temper your expectations. If the team is beating Pythag by 5 or more games, you can pretty much rest assured that there is a losing streak on the horizon. Conversely, if the team is 5 or more games under Pythag, things should turn around for the better.

The temperature analogy is pretty appropriate. It's basically a symptom, an indicator that something is not gelling. When you see a skewed Pythag differential, you know it's time to start looking for the real causes.

jojo
10-05-2009, 02:31 PM
I meant the 3 game differential.

I don't think three games is meaningful one way or the other due to noise.

Pythag predicts a teams record based upon the relative total RS and RA of a team.

However, the order/context in which those runs were scored/prevented (usually considered a random effect) can effect the actual winning percentage.

BCubb2003
10-05-2009, 02:34 PM
It's a tool that can temper your expectations. If the team is beating Pythag by 5 or more games, you can pretty much rest assured that there is a losing streak on the horizon. Conversely, if the team is 5 or more games under Pythag, things should turn around for the better.

The temperature analogy is pretty appropriate. It's basically a symptom, an indicator that something is not gelling. When you see a skewed Pythag differential, you know it's time to start looking for the real causes.

This is what I'm looking for, but can you find it in the Reds chart above?

dougdirt
10-05-2009, 02:36 PM
It's a tool that can temper your expectations. If the team is beating Pythag by 5 or more games, you can pretty much rest assured that there is a losing streak on the horizon. Conversely, if the team is 5 or more games under Pythag, things should turn around for the better.

The temperature analogy is pretty appropriate. It's basically a symptom, an indicator that something is not gelling. When you see a skewed Pythag differential, you know it's time to start looking for the real causes.
Except thats not really true. All it says is that they were probably lucky or unlucky to that point in the season. It doesn't tell you what is likely to happen because from that point forward they may be + or - in their RS/RA.

Brutus
10-05-2009, 02:40 PM
Except thats not really true. All it says is that they were probably lucky or unlucky to that point in the season. It doesn't tell you what is likely to happen because from that point forward they may be + or - in their RS/RA.

This is my take on it as well. In real time, it tells you whether a team is over or under-performing their production based on record. However, it does not account for whether a team will produce more or less later. If a team maintains that same differential, you could get a good idea of what the team's final record would be based on that differential. But it doesn't predict the future, it simply predicts what the future would be based on the current results.

I think it's a handy tool, but I like the analogy to a thermometer. It's really just a current temperature reading.

Edit: I think the biggest use for it is after trying to figure out where to deploy resources in the offseason and mixing and matching various transactions. With estimated WAR and total run figures, you can use these and the pythag to get an idea of what a team's record should be with or without certain players.

jojo
10-05-2009, 02:57 PM
Actual W-L record is pretty much like any other aspect.

For instance if a player was OPSing .890 over a 20 game stretch but was really a true talent .750 OPS guy, one should expect him to post an OPS of .750 over the next X number of games (i.e. his OPS will regress to his true talent).

Same with W-L record. If a team is a true talent .500 team despite playing .520 ball up to a point, one should expect them to play .500 ball going forward.

Given a large enough sample, pythag is a poor man's estimate of what a team's true talent winning percentage should be based upon the RS/RA it's accrued. But just as the actual record is subject to noise, so to is pythag (i.e. pythag doesn't care if a player is having an unusually good or unusually bad year, it just tallies the runs). But over the course of a season, it's probably true that those kind of issues cancel themselves out on a roster...

IslandRed
10-05-2009, 03:29 PM
It's a tool that can temper your expectations. If the team is beating Pythag by 5 or more games, you can pretty much rest assured that there is a losing streak on the horizon. Conversely, if the team is 5 or more games under Pythag, things should turn around for the better.

The temperature analogy is pretty appropriate. It's basically a symptom, an indicator that something is not gelling. When you see a skewed Pythag differential, you know it's time to start looking for the real causes.

Maybe. Maybe not. To use another analogy, coin flips: maybe we end up getting heads 60 times out of the first 100 flips. Probability and regression to the mean tell us it will eventually come back to 50% and we shouldn't expect 60% going forward. But it isn't necessarily going to come back to 50% by Flip 162 just because we decided to draw a line there.

Johnny Footstool
10-05-2009, 04:14 PM
Of course, sometimes it's luck. Actually, MOST of the time, it's luck.

But that shouldn't stop a diligent GM from checking.

*BaseClogger*
10-05-2009, 04:24 PM
I like using Pythag as a baseline for the team's true talent level. For example, the 2007 Seattle Mariners were 88-74 and accordingly viewed themselves as immediate contenders in 2008. However, if they would have used their Pythag record rather than actual won-loss to judge their position they would have seen they were much farther from contention (Pythag record of 79-83)...

PuffyPig
10-05-2009, 07:15 PM
You guys like Pythagoras a lot more than I do, but maybe it's because I just don't understand it that well. Does the fact the Reds piled on a bunch of runs last night when they already had the game in hand make them a better team? Does it mean that they will win more games?

In a word, yes.

Stats show in in judging who is more likely to win a playoff series, the team with a better run differential has a better chance to win the series than the team with the better record.

Which suggests that run differential is better way of evaluating a team than W-L record.

One other point.

My personal opinion is that team with a strong back end of the bullpen (but weak 5-7 bullpen guys) should be expected to beat their pythag. When games are close, you will see the top bullpen guys, when games are not close, you will see the worse guys. The run differential will get inflated in more meaningless games. You would expect that team to win their fair share of close games.

PuffyPig
10-05-2009, 07:19 PM
Maybe. Maybe not. To use another analogy, coin flips: maybe we end up getting heads 60 times out of the first 100 flips. Probability and regression to the mean tell us it will eventually come back to 50% and we shouldn't expect 60% going forward. But it isn't necessarily going to come back to 50% by Flip 162 just because we decided to draw a line there.

Actually, once you flip 60 heads in 100 throws, you would expect that going forward, you would from then on get 50% heads. So, the expectation is that you will always have more heads than tails if you kept a runniung total, though the more flips you make, the less chance the initial 60/40 flips will make much of a difference.

jojo
10-06-2009, 07:21 AM
Dave Cameron at USSM weighs in on pythag (http://ussmariner.com/) and talks about some of the issues being discussed in this thread (basically talks about how randomness, i.e. luck, effects pythag):



WAR and the 2009 Mariners
Dave October 5, 2009 at 8:49 pm Filed Under Mariners

A few facts you’ve probably heard tossed around the last few days:

1. The Mariners won 85 games
2. The M’s were outscored by 52 runs
3. Their pythag record, based on the above run differential, was just 76-86.

People have extrapolated all kinds of things from those facts. One of the more popular explanations for those three facts is that the team isn’t as good as their final record indicates, and that the extra wins are either due to unrepeatable good luck or the bountiful harvest of unquantifiable team chemistry. I’m here to say that both of those explanations are bunk.

Run differential is a decent tool when its used correctly and its limitations are understood. Good teams have large positive run differentials, bad teams have large negative run differentials, and teams in the middle are generally mediocre. It’s pretty hard to succeed without regularly outscoring your opponents, as should be pretty obvious, and no team is good enough to consistently win enough one or two run games to overcome a lack of talent. So, you can generally do a pretty decent job of projecting a team’s Win-Loss record based on RS-RA. No team was more than +/- 9 wins away from their pythag record this year, for instance.

That said, run differential is certainly not a perfect estimator of a team’s real abilities. Run totals can be skewed heavily by performance with men on base, either by hitters coming through in the clutch or pitchers stranding runners. In general, teams do not have a real ability to be significantly better or worse than you would expect in these situations based on their overall production, but over the sample of one season, teams performance can vary enough to affect their run totals. And that affects their pythag record, even though it’s not a real indicator of talent.

This can have a real impact on how people see a team overall. Not surprisingly to anyone who watched regularly, the Mariners were the worst hitting team in baseball with runners on base this year, hitting .254/.317/.396 with a man on. They were also worst in baseball at hitting with runners in scoring position, coming in at just .234/.312/.358. As a team, the M’s hit better with the bases empty than in pretty much any scenario where they had a chance to drive in a run (which is not normal, if you’re wondering), and while that’s frustrating to watch, it’s not an indicator that the M’s were really the worst offense in baseball – they were just a bad offense that failed in the clutch more often than they should have. And that skews their runs scored total down, which pushes their pythag record down, and, well, you get the idea.

Anyway, that’s a lot of words used to say that pythag is not luck independent, and shouldn’t be used as the be-all, end-all determinant of how “good” a team really was. In fact, we have a better way of measuring team talent level, and it’s one you should probably be familiar with by now – Wins Above Replacement.

You’ve seen us talk about WAR a lot. It’s the best measure of player value that we have, and while it’s not perfect, it’s pretty darn good. It sums up a player’s offensive and defensive value, as well as accounting for time on the field, and puts it on a scale over the production that could be expected from a pretty good Triple-A player would offer for the league minimum. At the team level, it’s the total of all the wins added by players on the roster throughout the year. And, because all of the inputs used in the formula are context free, it doesn’t know anything about the timing of specific events and isn’t affected by “clutch” performances in the way that run totals are.

WAR is a better indicator of talent level than pythag. And you know what WAR thought the M’s record “should be” this year? 83-79. The M’s got 21 wins from their position players (mostly thanks to their league best defense) and 16 wins from their pitching staff. Based on the calculations from FanGraphs, a replacement level team this year would have won ~46 games, so add those 46 wins to the 37 extra that the M’s got, and you have 83 expected wins.

In other words, the M’s weren’t a 76 win team that got really lucky or willed themselves to 10 extra wins through their harmony and hugging. They were a team that played well enough to finish two games over .500 and actually finished four games over .500. There’s nothing to explain. The M’s finished right about where we’d have expected them to, given how well they hit, caught, and pitched.

Don’t let the Pythag Police try to convince you that there’s an inevitable massive regression coming because the M’s outperformed their pythag. Their pythag underperformed their actual offensive level, and once you adjust for all of the facts that could be considered “luck” (not just some of them, as pythag does), you have to conclude that the M’s were basically the team that their final record indicates. There is certainly still a lot of work to do to get this team to be a playoff contender, but in order to know how much work needs to be done, you have to start from the right foundation. And that foundation is not pythag record.





The difference between actual record and pythag record is almost always due to a deviation from normal distribution of run scoring.

For instance, a team that wins a bunch of close games and loses a bunch of blowouts will outperform their pythag record. That’s been the M’s recipe in 2007 and 2009. Their record in one run games was terrific, but they never blew anyone out, so their wins didn’t inflate their RS in the same way that their losses inflated their RA.

The opposite is also true. This year’s Indians are a great example. Won a bunch of games by big margins, but lost a bunch of close games. As such, they vastly underperformed their pythag.

In general, teams don’t really have control over the distribution of their run scoring. They can’t stop scoring in one game to save more runs for the next. Over a big enough sample, this would even out, but 162 games is not a big enough sample to get rid of all the noise.


Also here's the final standings with second and third order pythag (third order is considered a better indicator than just the straight actual unadjusted RS/RA-based pythag):

http://www.baseballprospectus.com/statistics/standings.php

Third order pythag has the Ms at 83-79 and the Reds at 71-91 (WAR suggests the Reds were a 73 win team).

Third order pythag is basically derived from "expected RS/RA". In other words, a run estimator is used (linear weights or base runs) to estimate the expected RA/RS based upon player performance. These expected runs are used to calculate pythag and are then adjusted for strength of schedule.

Brutus
10-06-2009, 03:31 PM
I respect Cameron's work a lot. And most certainly I believe WAR is the best estimate we have (I think numerous studies have shown this to correlate the best with actual W-L record). I don't buy, however, that "clutch" hitting and situational hitting doesn't matter and does not explain anything. That's something I'll never believe. It might not be easily explained, quantified or measured, but in every facet of life, some people handle stress, pressure and tricky situations differently than how they handle normal situations. Some of the best people in their profession don't handle high-stress situations as well. Meanwhile, there are some people that are just average in performance, but have a knack of coming through with natural leadership qualities when the going gets tough.

I think it's silly to think that the goal of a baseball player is to take a bat with him to the plate, swing at a pitch, drive a runner in, and then when a team has a set of players doing that extraordinarily well or extremely poorly, dismiss it as luck and has little to do with the team's ability to win or lose games. Sure, there will always be some regression, and talent and ability affect these numbers more than anything, but when we figure out how to best quantify it, I think we'd see these situations can, do and should affect wins and losses.

I still agree with Cameron about WAR. I do think it's the most comprehensive stat we have thus far to gauge performance. But I don't think these other factors should be so easily dismissed as having an impact.

Eric_the_Red
10-06-2009, 03:33 PM
I respect Cameron's work a lot. And most certainly I believe WAR is the best estimate we have (I think numerous studies have shown this to correlate the best with actual W-L record). I don't buy, however, that "clutch" hitting and situational hitting doesn't matter and does not explain anything. That's something I'll never believe. It might not be easily explained, quantified or measured, but in every facet of life, some people handle stress, pressure and tricky situations differently than how they handle normal situations. Some of the best people in their profession don't handle high-stress situations as well. Meanwhile, there are some people that are just average in performance, but have a knack of coming through with natural leadership qualities when the going gets tough.

I think it's silly to think that the goal of a baseball player is to take a bat with him to the plate, swing at a pitch, drive a runner in, and then when a team has a set of players doing that extraordinarily well or extremely poorly, dismiss it as luck and has little to do with the team's ability to win or lose games. Sure, there will always be some regression, and talent and ability affect these numbers more than anything, but when we figure out how to best quantify it, I think we'd see these situations can, do and should affect wins and losses.

I still agree with Cameron about WAR. I do think it's the most comprehensive stat we have thus far to gauge performance. But I don't think these other factors should be so easily dismissed as having an impact.


WAR? Huh? What is it good for?

Brutus
10-06-2009, 03:37 PM
WAR? Huh? What is it good for?

Absolutely nothing... or a lot.

Heh.

RedsManRick
10-06-2009, 03:43 PM
I respect Cameron's work a lot. And most certainly I believe WAR is the best estimate we have (I think numerous studies have shown this to correlate the best with actual W-L record). I don't buy, however, that "clutch" hitting and situational hitting doesn't matter and does not explain anything. That's something I'll never believe. It might not be easily explained, quantified or measured, but in every facet of life, some people handle stress, pressure and tricky situations differently than how they handle normal situations. Some of the best people in their profession don't handle high-stress situations as well. Meanwhile, there are some people that are just average in performance, but have a knack of coming through with natural leadership qualities when the going gets tough.


It's not that clutch doesn't exist. I don't know of anybody in sabermetrics that believes that and the most recent studies suggest it does, though the effect is somewhat smaller than most would suspect.

No, the larger issue is what you said, it's very difficult to measure. It is virtually impossible to tell the difference between "luck" and "clutch". And so, when you're look to explain changes or variance, there's no reason to talk about them in two different buckets. Does it exist? Yes. Can we do anything with the knowledge that it exists? No. Particularly in making projections or decisions about who should bat when, our best guess remains how well the player performs generally -- using some other estimate based on past "clutch" performance is likely to be less accurate.

jojo
10-06-2009, 03:43 PM
I respect Cameron's work a lot. And most certainly I believe WAR is the best estimate we have (I think numerous studies have shown this to correlate the best with actual W-L record). I don't buy, however, that "clutch" hitting and situational hitting doesn't matter and does not explain anything. That's something I'll never believe. It might not be easily explained, quantified or measured, but in every facet of life, some people handle stress, pressure and tricky situations differently than how they handle normal situations. Some of the best people in their profession don't handle high-stress situations as well. Meanwhile, there are some people that are just average in performance, but have a knack of coming through with natural leadership qualities when the going gets tough.

I think it's silly to think that the goal of a baseball player is to take a bat with him to the plate, swing at a pitch, drive a runner in, and then when a team has a set of players doing that extraordinarily well or extremely poorly, dismiss it as luck and has little to do with the team's ability to win or lose games. Sure, there will always be some regression, and talent and ability affect these numbers more than anything, but when we figure out how to best quantify it, I think we'd see these situations can, do and should affect wins and losses.

I still agree with Cameron about WAR. I do think it's the most comprehensive stat we have thus far to gauge performance. But I don't think these other factors should be so easily dismissed as having an impact.

I think Dave's point in the quote above was that teams generally perform similarly in situational hitting as they do overall. The Ms however had a pretty significant disparity between the two this year but the difference wouldn't likely be as dramatic going forward. Thus that shouldn't be a peg to hang ones hat upon when evaluating their offensive woes.

Johnny Footstool
10-06-2009, 03:50 PM
It's not that clutch doesn't exist. I don't know of anybody in sabermetrics that believes that and the most recent studies suggest it does, though the effect is somewhat smaller than most would suspect.

No, the larger issue is what you said, it's very difficult to measure. It is virtually impossible to tell the difference between "luck" and "clutch". And so, when you're look to explain changes or variance, there's no reason to talk about them in two different buckets. Does it exist? Yes. Can we do anything with the knowledge that it exists? No. Particularly in making projections or decisions about who should bat when, our best guess remains how well the player performs generally -- using some other estimate based on past "clutch" performance is likely to be less accurate.

Clutch situations definitely exist.

Players who reliably perform better in those situations than in normal situations are much harder to pin down (Mark Grace being the exception).

RedsManRick
10-06-2009, 03:55 PM
Clutch situations definitely exist.

Players who reliably perform better in those situations than in normal situations are much harder to pin down (Mark Grace being the exception).

It's a question of scale and the definition of "reliable" in my mind. When the effect has been found it's on the order of 10 to 20 points of OPS.

Even if we know that a guy is "reliably" clutch, incorporating our knowledge of his clutchiness tells us virtually no more about what is likely happen in a given PA than do his overall numbers.

And in any event, the actual results of what happen will vastly affect our perceptions that we won't want to believe what the numbers say anyways. In the end, most people will only accept those explanations that confirm their personal experiences.

Brutus
10-06-2009, 03:59 PM
I think Dave's point in the quote above was that teams generally perform similarly in situational hitting as they do overall. The Ms however had a pretty significant disparity between the two this year but the difference wouldn't likely be as dramatic going forward. Thus that shouldn't be a peg to hang ones hat upon when evaluating their offensive woes.

I think that's the philosophical difference I'm referring to, though. There will always be fluctuations in any measurement going forward. It's those fluctuations that are due to varying level of performances in certain situations that we cannot measure. Instead of assuming they'll regress to the same level as overall performance, I would argue it's those fluctuations that do separate teams from each other even above and beyond their actual talent level.

My outlook is that the Mariners did perform that poorly in those situations. That's why they play the games. I absolutely agree with you that it does not mean they will always perform that bad in those circumstances, but I hate to automatically dismiss it as something that shouldn't be taken into account with the difference between actual record and Pythagorean record. Situational hitting should absolutely matter, and accordingly, the spread between overall production and situational production can and will vary by team (in other words, the same differential between overall talent level and situational production will not exist from team to team). It's for that reason I don't think it's prudent to dismiss this factor in analyzing the run differential.

I think we agree on what it is he's trying to convey, but we disagree with the logic on why he's saying it.

jojo
10-06-2009, 04:08 PM
I think that's the philosophical difference I'm referring to, though. There will always be fluctuations in any measurement going forward. It's those fluctuations that are due to varying level of performances in certain situations that we cannot measure. Instead of assuming they'll regress to the same level as overall performance, I would argue it's those fluctuations that do separate teams from each other even above and beyond their actual talent level.

My outlook is that the Mariners did perform that poorly in those situations. That's why they play the games. I absolutely agree with you that it does not mean they will always perform that bad in those circumstances, but I hate to automatically dismiss it as something that shouldn't be taken into account with the difference between actual record and Pythagorean record. Situational hitting should absolutely matter, and accordingly, the spread between overall production and situational production can and will vary by team (in other words, the same differential between overall talent level and situational production will not exist from team to team). It's for that reason I don't think it's prudent to dismiss this factor in analyzing the run differential.

I think we agree on what it is he's trying to convey, but we disagree with the logic on why he's saying it.

I think he's basically saying, the offense sucked, no doubt about it. Just don't try to fix it by focusing upon how they hit situationally (or to put it another way, don't try to fix their offense by looking specifically to improve their situational hitting).

Brutus
10-06-2009, 04:25 PM
I think he's basically saying, the offense sucked, no doubt about it. Just don't try to fix it by focusing upon how they hit situationally (or to put it another way, don't try to fix their offense by looking specifically to improve their situational hitting).

If that's what he's saying, then I can buy into that. I didn't read it that way, but I kind of felt he didn't clarify his position so perhaps I just interpreted it wrong.

jojo
10-06-2009, 04:30 PM
If that's what he's saying, then I can buy into that. I didn't read it that way, but I kind of felt he didn't clarify his position so perhaps I just interpreted it wrong.

He's not a proponent of clutch. At least he would never argue that a GM should pay extra for the perceived attribute (actually he'd likely argue the opposite) so I doubt you guys would agree on this issue.

Brutus
10-06-2009, 04:40 PM
He's not a proponent of clutch. At least he would never argue that a GM should pay extra for the perceived attribute (actually he'd likely argue the opposite) so I doubt you guys would agree on this issue.

Well, I am obviously a believer in clutch existing (in some form or another). But until there's a good, consistent way of measuring it, I certainly would not spend resources on it unless I'm certain of how it's quantified.