My Thoughts on WAR and RBI

[willshad]

People seem to think that just because someone talks about the value of RBI, that they think its all that should be used. Likewise, because someone talks about how 'off' WAR could be , they think that they dismiss it entirely. The ironic part, is that THEY seem to think that WAR is actually the only stat that should be used, even while admitting it is far from perfect.
One guy has more WAR than another guy...fine, that's a point in his favor. If that's the deciding factor between two otherwise close players, then so be it. Same with RBI..if two players have similar stats, and yet one drives in 130 and the other drives in 80, then guess who I would say did better? Sure, we can make 'adjustments' mentally based on the player's lineup, era, stadium, etc, but we should also make 'adjustments' mentally for WAR as well. I have found that we can pretty much add 1 or 2 WAR to any great season by a catcher. We can pretty much add 1 or 2 WAR to a poor defensive player, and subtract 1 or 2 from a good one. We can add another WAR point for a guy who doesn't walk a lot, and subtract 1 from a guy who walks a lot.
This can make a huge difference in some players, and logic backs this up. A guy who is a poor defensive catcher who doesn't walk, like Piazza, or perhaps Berra would end up with 75 or 80 WAR for their careers, instead of 56, like they have now.
The thing is, that we already KNOW that these players, Berra and Piazza, had careers closer to Bagwell, Griffey, and Dimaggio(the 75- 80 range) than they did to Bobby Bonds, Willie Davis and Sal Bando (the 55-57 range), so not only does it prove my 'adjustments' to be correct, but it kind of defeats the purpose of WAR to begin with.
If you want to accept as gospel a stat that shows that players who are universally regarded as top 5 (sometimes number 1) at their position had the same value in their career as other players who are not even top 20 at their position, so be it. Me, I am just going to use it as one factor.

[filihok]

People seem to think that just because someone talks about the value of RBI, that they think its all that should be used. Likewise, because someone talks about how 'off' WAR could be , they think that they dismiss it entirely. The ironic part, is that THEY seem to think that WAR is actually the only stat that should be used, even while admitting it is far from perfect.

Who are these people of whom you speak?

One guy has more WAR than another guy...fine, that's a point in his favor. If that's the deciding factor between two otherwise close players, then so be it. Same with RBI..if two players have similar stats, and yet one drives in 130 and the other drives in 80, then guess who I would say did better? Sure, we can make 'adjustments' mentally based on the player's lineup, era, stadium, etc, but we should also make 'adjustments' mentally for WAR as well. I have found that we can pretty much add 1 or 2 WAR to any great season by a catcher. We can pretty much add 1 or 2 WAR to a poor defensive player, and subtract 1 or 2 from a good one. We can add another WAR point for a guy who doesn't walk a lot, and subtract 1 from a guy who walks a lot.
This can make a huge difference in some players, and logic backs this up. A guy who is a poor defensive catcher who doesn't walk, like Piazza, or perhaps Berra would end up with 75 or 80 WAR for their careers, instead of 56, like they have now.

Does this make WAR more or less accurate than it is now?

http://www.hardballtimes.com/main/article/what-is-war-good-for/

[ATTACH=CONFIG]118206[/ATTACH]

I'll await your results

I hope my wait isn't as in vain as this wait was/is

Like I said, ANY positive offensive stat will correlate almost perfectly with the team's run scoring, and winning.

[Jackaroo Dave]

So, Willshad, something like this would be a good idea?

[B]Rank ▴ 	First Base	RBI 	      Catcher	RBI 	Diff	Pos RBI[/B]
1 	Lou Gehrig 	1995 	Yogi Berra 	1430 	565	1870
2 	Jimmie Foxx 	1922 	Ted Simmons 	1389 	533	1829
3 	Eddie Murray 	1917 	Johnny Bench 	1376 	541	1816
4 	Cap Anson 	1880 	Mike Piazza 	1335 	545	1775
5 	R. Palmeiro 	1835 	Ivan Rodriguez 	1332 	503	1772
6 	Jake Beckley 	1578 	Carlton Fisk 	1330 	248	1770
7 	Willie McCovey 	1555 	Gary Carter 	1225 	330	1665
8 	Fred McGriff 	1550 	Bill Dickey 	1209 	341	1649
9 	Jeff Bagwell 	1529 	Gabby Hartnett 	1179 	350	1619
10 	Carlos Delgado 	1512 	Lance Parrish 	1070 	442	1510
				                    [B]AVG 439.8	[/B]

Since RBI is biased against catchers, we should add about 440 RBI to the catchers' total when comparing them to first basemen?

[brett]

We know that WAR messes up on catchers. Part of it is because a franchise might go through 3 catchers in the same amount of time that they go through 2 players at other positions. As I've pointed out, I give catchers 150% of the replacement level runs per game caught, or about 1.1 extra was per 162 games. We also should probably find a way to credit catchers for some of the credit that a pitching staff gets. A replacement pitcher with a good catcher might outperform replacement level for example.

As for RBI, the big problem is that if increased RBI come at the expense of creating more outs (putting say 40 more balls in play but netting 25 outs, and 25 fewer walks), then the RBI are superficially linear to value. 120 RBI versus 105 with 25 more outs and 25 fewer walks is not 120/105 as valuable. Since a team scores about .15 runs per out, perhaps we could look at RBI minus .15 per out (or RBI above average per out). We would also have to give the other guy credit for getting on base except in situations where the value of a walk is minimized.

If a guy puts up his hit and walk and extra base rates randomly then metrics work. If a guy puts it in play more in the 1/3 of situations where a ball in play is worth more then he should gain some value.

[Francoeurstein]

About two years ago, I started to just completely ignore RBI. It's just not a valid stat for player comparison to me.

[filihok]

We know that WAR messes up on catchers.

How do we know that WAR messes up on catchers?

WAR is exceptionally accurate in measuring how much individual players contribute to team wins.

WARvs.Wins_.jpg

How much of the error is due to not properly valuing catchers?

Part of it is because a franchise might go through 3 catchers in the same amount of time that they go through 2 players at other positions. As I've pointed out, I give catchers 150% of the replacement level runs per game caught, or about 1.1 extra was per 162 games.

How did you come to make these adjustments?

We also should probably find a way to credit catchers for some of the credit that a pitching staff gets. A replacement pitcher with a good catcher might outperform replacement level for example.

Absolutely, there is work being done in that area.

http://www.hardballtimes.com/main/article/another-one-bites-the-dust/

[Francoeurstein]

How do we know that WAR messes up on catchers?

WAR is exceptionally accurate in measuring how much individual players contribute to team wins.

[ATTACH=CONFIG]118211[/ATTACH]

How much of the error is due to not properly valuing catchers?



How did you come to make these adjustments?




Absolutely, there is work being done in that area.

http://www.hardballtimes.com/main/ar...ites-the-dust/

You just need to make that picture your sig, big guy.

[Bothrops Atrox]

Absolutely, there is work being done in that area.

http://www.hardballtimes.com/main/ar...ites-the-dust/

Lots of work being done re: pitch framing too.

[Sultan_1895-1948]

We know that WAR messes up on catchers. Part of it is because a franchise might go through 3 catchers in the same amount of time that they go through 2 players at other positions. As I've pointed out, I give catchers 150% of the replacement level runs per game caught, or about 1.1 extra was per 162 games. We also should probably find a way to credit catchers for some of the credit that a pitching staff gets. A replacement pitcher with a good catcher might outperform replacement level for example.

As for RBI, the big problem is that if increased RBI come at the expense of creating more outs (putting say 40 more balls in play but netting 25 outs, and 25 fewer walks), then the RBI are superficially linear to value. 120 RBI versus 105 with 25 more outs and 25 fewer walks is not 120/105 as valuable. Since a team scores about .15 runs per out, perhaps we could look at RBI minus .15 per out (or RBI above average per out). We would also have to give the other guy credit for getting on base except in situations where the value of a walk is minimized.

If a guy puts up his hit and walk and extra base rates randomly then metrics work. If a guy puts it in play more in the 1/3 of situations where a ball in play is worth more then he should gain some value.

Holy smokes, someone who is engulfed into the stat world actually gets it. This is indeed a breakthrough Ray Schalk is smiling

[filihok]

Holy smokes, someone who is engulfed into the stat world actually gets it. This is indeed a breakthrough Ray Schalk is smiling

Lots of people get it. It's just very hard to quantify

[Bothrops Atrox]

Holy smokes, someone who is engulfed into the stat world actually gets it. This is indeed a breakthrough Ray Schalk is smiling

This has been a hot-topic issue for research and discussion for the stat-world for years.

[Sultan_1895-1948]

Absolutely, there is work being done in that area.

http://www.hardballtimes.com/main/ar...ites-the-dust/

Does that link mostly have to do with pass balls vs wild pitches? Someone spent a lot of time going into detail about what can be said in a simple way.

The simplest being, pitches that are bounced in front of the plate have a good chance to get past the catcher. Slightly more complex would be, those of the in-between hop variety, are likely to go past the catcher or get away far enough for a runner to easily advance. Even more complex, would be, those that fall to the catchers backhand are most likely to go past or far enough for a runner to easily advance, and even going a step further, those pitches that include all the other factors, and happen to be a curveball are all but certain to do the same. It's not rocket science.

[brett]

How do we know that WAR messes up on catchers?

WAR is exceptionally accurate in measuring how much individual players contribute to team wins.

[ATTACH=CONFIG]118211[/ATTACH]

How much of the error is due to not properly valuing catchers?

I am suggesting 2 things.
1 is that the replacement level should be dropped slightly because teams must utilize a catcher and a backup catcher. Changing the replacement level would simply change the constant (B) in the regression equation and the correlation would be the same. It would just shift axes values but the graph would look identical.

2, you are showing team war versus wins. If we simply give the catcher a fraction of the pitcher's value for calling a game, the team war also stays the same.

I made 2 assumptions: 1 that if a starting catcher went out for a season a team would probably get about 2/3 of its games from a replacement catcher and 1/3 from a sub replacment level catcher, or about 108 and 54. Second is that the sub replacment would be fractionally as good as the replacement as the replacement was as good as the average player. This drops the replacment level by 1.1 war. In other words a team of replacments would not win 52 games, but would win 50.9 because they would also use an even lower level catcher about 1/3 of the time.

Edit: upon working out the math, if I assume the sub replacment catcher to be fractionally as valuable relative to the replacement as the replacement is to average, the sub-replacement catcher would be a .32 x .32 or about a 10% winning percentage player about 35 runs below average, and if he filled in 1/3 time then a catcher should actually get about 1.33 extra war per 162 games caught. That would put Johnny Bench up to 86.6 war, or move him from 45th to 30th all time (which with LQ and post season adjustments might put him in the 22-26 range. Berra would rise from 56.1 to 70.0 or from 117 to 49.

I also think that the catchers should get a small share of pitching war too.

[brett]

Consider this as well. An average pitching staff gets about 8 pitching war just for pitching at an average level for the season, but the catcher gets none for catching all those pitches at an average level. If the catcher deserves just a share of those 8 war then it not only would make more sense to me, but it would also decrease the huge war from the 300+ inning pitchers just for being above replacement.

[Second Base Coach]

I love how WAR gives credit for the position typed out on the front of the baseball card even before we turn it over to inspect the statistics on the back.