22-1 - What Is It Good For?

by David S. Cohen on Jul 7, 2009 11:26 AM EDT

The Phillies won a game last night. In the actual standings, that's all it was -- just another win.

But, for those who look to advanced metrics, last night's game was more than that. Sabermetricians have developed ways to evaluate a team's won/loss record based on other metrics to determine whether a team is playing above or below what its actual on-field performance would otherwise dictate.

The simplest, and most well known, of these metrics is Bill James' Pythagorean Record. This measure looks just to runs scored and runs given up to determine what a team's record should be.

A more complicated system is Baseball Prospectus' Third-Order Wins. This measure adjusts runs scored for luck as well as strength of schedule.

More detailed explanations can be found from following the links, but the basic idea is that whereas the actual win total is derived from actual wins (yes, brilliant, I know), these other metrics look to runs scored and runs given up, or some variation thereof.

Which leads to the question - what did last night's 22-1 win do for these measures? The answer - quite a lot:

	Before		After
	Wins	Win %	Wins	Win %
Actual	42.0	0.532	43.0	0.538
Pyth	41.4	0.524	43.9	0.549
BP 3rd Order	38.5	0.487	40.2	0.503

A 22-1 win in game 80 improved the Phillies' actual record by just 1 win. But, the team's Pythagorean record improved by 2.5 wins, and its third-order record improved by 1.7 wins. Not bad for one game's worth of work.

Tweet Comment 17 comments | 0 recs |

Do you like this story?

Comments

It also seems to have done wonders for the statistical output of this blog!

by taco pal on Jul 7, 2009 11:54 AM EDT reply actions

Next time it is just esier to beg…

For Who? My teammates.

For What? To Win.

How Much? Where do I sign?

by jonk on Jul 7, 2009 12:20 PM EDT up reply actions

now imagine if we had pitching

it’s wonderful isn’t it (like the last 4 games)

wtf happened in June?

by sports00fan00 on Jul 7, 2009 1:06 PM EDT reply actions

Read this please.

by taco pal on Jul 7, 2009 1:09 PM EDT up reply actions

What was it good for?

Gave Victorino a crazy line that could help him win the final All Star spot. Too bad Sandoval had to go and hit a grand slam an hour later…

by philiafan14364 on Jul 7, 2009 1:26 PM EDT reply actions

this is the fallacy with the pythags though….
Winning a game 22-1, isn’t really all that different than winning 8-1, but it credits you as such.
If the pythags maxed out at like a 10 run differential, I think it would be more accurate.

by Bilzo on Jul 7, 2009 1:31 PM EDT reply actions

Seems to me that winning 22-1 is very different from winning 8-1. Fourteen runs is a lot.

by taco pal on Jul 7, 2009 1:34 PM EDT up reply actions

Not exactly

After the score was up 16-1, Utley was pulled, Howard was pulled, the Reds put in a position player to pitch. Both teams had effectively given up at that point. Additional runs do not truly represent the performance of either team at that point.

by Bilzo on Jul 7, 2009 1:40 PM EDT up reply actions

The rationale behind the pythags is that it balances out over the season – yes 22 runs will throw it off temporarily – but after the philes go a week of 1-2 run games it comes down precipitously…

Don't frack with me or you'll get a punch in the kidneys...you've been warned

by jemagee on Jul 7, 2009 2:11 PM EDT up reply actions

Not exactly

The Phils would need to lose a game by 16-18 runs to balance it out. Based on run differential, they could lose their next 20 games by one run each, and still have a .500 pythag, but their real record would be 1-20.

Good teams have winning records in 1-2 run games. That’s because that’s what a lot of games are. 1-2 run differentials. The games with the 7+ run differentials skew the results more than they should. While only a good offensive team is likely to have a sniff at a 15+ run game (ie…don’t expect the nats or pirates to do this), when it does happen it skews the stats. The phillies are not 1-2 wins better today than yesterday, just because they had an offensive explosion.

by Bilzo on Jul 7, 2009 3:00 PM EDT up reply actions

But maybe the Phillies were 1-2 wins better than the Pythagorean record would have suggested yesterday, but through mischance had not scored as many runs as their talent would suggest.

There’s nothing wrong with the Pythagorean record. There is a problem with looking at it at the micro level.

by phatj on Jul 7, 2009 3:04 PM EDT up reply actions

This is untrue. Teams don’t have that much control over the 1-2 run games (mostly 1 run games). The best teams tend to have many more blowouts than they are blown out. These is plenty of statisical analysis backing this up.

Bad teams are likely to win close games while good teams win blowouts.

For Who? My teammates.

For What? To Win.

How Much? Where do I sign?

by jonk on Jul 7, 2009 3:48 PM EDT up reply actions

coming in just a second…the phils record by run differential.

by Bilzo on Jul 7, 2009 4:06 PM EDT up reply actions

well.....

looks like I might be in error, but I’m going to play the “small sample size” card!
6 Run games went poorly, but the phils never had a loss in a differential greater than 6 runs in 2008.

Differential W-L
1 27-23
2 14-15
3 16-10
4 14-8
5 9-5
6 2-9
7 4-0
8 1-0
9 2-0
10 1-0
15 1-0
18 1-0

by Bilzo on Jul 7, 2009 4:23 PM EDT reply actions