How is WAR calculated?

[NeedMoreLike84]

Please don't debate the value. But for a guy who has no head for math how is it calculated? How do you determine "replacement level?"

I am a huge fan of new thinking and saber philosophies but my limited knowledge of math makes it tough for me...

[zimm]

might i suggest this for your christmas list.

the author is purportedly a pretty good guy but his fantasy baseball skills are questionable at best, probably because he spends too much time on this forum.

Beyond Batting Average: Lee Panas: 9780557312245: Amazon.com: Books

[tiger337]

The Baseball-reference WAR calculation can be found here:

Baseball-Reference.com WAR Explained - Baseball-Reference.com

I could not find the new FanGraphs calculation all in one place, so I'll explain it here as best as I can:

(1) Calculate wOBA= (0.91 x 1B + 1.28 x 2B + 1.61 x 3B + 2.18 x HR + 0.71 x BB + 0.74 x HBP)/(PA-IBB)

(2) Calculate wRAA = (wOBA- MLB wOBA)/wobascale*PA. = (wOBA - .315)/1.245*PA for 2012

(3) Batting = wRAA/ ballpark factor (ballpark factors found here: Guts! | FanGraphs Baseball

(4) Calculate SB Runs = .2*SB-.4*CS

(5) Average SB runs is about 0.0011 per PA so wSB = SB Runs - .0011*PA

(6) Get UBR value for other base running.

(7) Base running = wSB+UBR

(8) positional adjustment

C +12.5
1B -12.5
2B +2.5
3B +2.5
SS +7.5
LF -7.5
CF +2.5
RF -7.5
DH -17.5

(9) Fielding = UZR

(10) For catchers, Fielding = rSB + RPP (see Catcher Defense | FanGraphs Sabermetrics Library)

(11) Replacement - Because typical replacement player is considered to be 20 runs below average for a full season, 20 runs per 600 PA are added to a player's total. e.g. player with 600 PA gets 20 added. player with 300 PA gets 10 added.

(12) RAR = Batting + Base running + Positional + Fielding + Replacement

(13) WAR = RAR/9.5

[tiger337]

might i suggest this for your christmas list.

the author is purportedly a pretty good guy but his fantasy baseball skills are questionable at best, probably because he spends too much time on this forum.

Beyond Batting Average: Lee Panas: 9780557312245: Amazon.com: Books

Thanks! It's explained better in my book, but they have tweaked WAR since publication, so I gave the complete new algorithm above.

[redshark63]

Thanks! It's explained better in my book, but they have tweaked WAR since publication, so I gave the complete new algorithm above.

Lee,

By BB, do you mean "unintentional" walks? I assume this, but want to confirm.

[tiger337]

Lee,

By BB, do you mean "unintentional" walks? I assume this, but want to confirm.

yes, just unintentional walks. And IBB should not be in there, so I need to fix my post!

[Antrat]

Don't forget to carry the 1.

[bigtenguy]

My head hurts!

[hueytaxi]

might i suggest this for your christmas list.

the author is purportedly a pretty good guy but his fantasy baseball skills are questionable at best, probably because he spends too much time on this forum.

Beyond Batting Average: Lee Panas: 9780557312245: Amazon.com: Books

I agree. I purchased this a year ago and hopefully can get Lee to sign it on one of his trips. While some of it lost me, he dumbed down some parts to make it easier to follow how, what and what is important.

[Shinzaki]

Based on comments from some sabremetric types I know, I think they just have a footrace and then they call the winner MVP

[tiger337]

Based on comments from some sabremetric types I know, I think they just have a footrace and then they call the winner MVP

Base running is the least influential component of WAR. If that is all a player can do, then he will have a very low WAR. Hitting is the most important thing, then defense.

[Yoda]

Based on comments from some sabremetric types I know, I think they just have a footrace and then they call the winner MVP

I'm guessing you aren't really listening to them.

[EchO]

Base running is the least influential component of WAR. If that is all a player can do, then he will have a very low WAR. Hitting is the most important thing, then defense.

One thing I don't particularly understand is how, or the extent of which, ballpark factors are used (not even necessarily WAR but stats like OPS+ also).

This year Miguel Cabrera had an OPS of 999 with an OPS+ of 165 while Mike Trout had an OPS of 963 and an OPS+ of 171, now was Angel stadium such a disadvantage that it should account for more than 36 points of OPS? I just personally don't think that if Cabrera played in Angel Stadium that he would end the season with an OPS lower than Trout's.

And one other question...shouldn't there be a place for BABIP in WAR? It seems unfair to use BABIP to discredit anomalous seasons (Jackson's 2010 campaign) but not have that reflected in a players WAR. IMO, if a player got exceptionally "lucky" or "unlucky" their WAR should be affected as such. During the entire Cabrera vs. Trout MVP debate it seemed like their was a lack of concern over Trout's .380+ BABIP (especially considering in 40 games in 2011 Trout posted a .247 BABIP).

I would be interested to see what Trout's WAR would be taking into account BABIP and without the added 40+ OPS points given because of "ballpark factors."

[RandyMarsh]

One thing I don't particularly understand is how, or the extent of which, ballpark factors are used (not even necessarily WAR but stats like OPS+ also).

This year Miguel Cabrera had an OPS of 999 with an OPS+ of 165 while Mike Trout had an OPS of 963 and an OPS+ of 171, now was Angel stadium such a disadvantage that it should account for more than 36 points of OPS? I just personally don't think that if Cabrera played in Angel Stadium that he would end the season with an OPS lower than Trout's.

And one other question...shouldn't there be a place for BABIP in WAR? It seems unfair to use BABIP to discredit anomalous seasons (Jackson's 2010 campaign) but not have that reflected in a players WAR. IMO, if a player got exceptionally "lucky" or "unlucky" their WAR should be affected as such. During the entire Cabrera vs. Trout MVP debate it seemed like their was a lack of concern over Trout's .380+ BABIP (especially considering in 40 games in 2011 Trout posted a .247 BABIP).

I would be interested to see what Trout's WAR would be taking into account BABIP and without the added 40+ OPS points given because of "ballpark factors."

I'd like to know this as well, because while I think park factors should play a factor, the fact is that unless you looked at the trajectory and weather conditions of every ball that a player it, it is hard to tell how much a player benefited in that particular park.

Let's use Cano for example since he plays in an extreme hitters park for LH hitters so park factors penalize him big time. But hypothetically what if every one of his HRs would have been HRs in any other park, not just Yankee Stadium? He gets penalized by WAR for playing there, but the fact is in this scenario he would've done the same at any park, in fact if anything the short fence would probably hurt him then because less balls would drop in for hits since the OF would have less space to cover.

So while I do feel that park factors should come into play, unless they look into each players balls in play individually I think that it can be extremely flawed.

[Gehringer_2]

..
So while I do feel that park factors should come into play, unless they look into each players balls in play individually I think that it can be extremely flawed.

I personally think park factor calculation is deeply flawed as well. It should not track the make-up of the home team, but it seems to anyway. There is some dynamic there that is unseen but effective. Further, they certainly should not be recalculated every year unless a park's architecture has actually changed. They should be allowed to average over the long term as long as the park stays the same - otherwise what you end up measuring in reality is mostly the annual departure of the local weather from average - and weather is too mercurial to try to credit or debit players for every year.

[EchO]

Another thing I don't understand about WAR and would love some clarification is why defensive position is incorporated. Isn't the entire purpose of WAR to calculate the wins over a replacement player? A first baseman is never going to be replaced by a CF in a 1 for 1 swap, so why should the first baseman get a -12.5? Or to put it another way, I don't see why a CF or C get bonuses to their position when they will be replaced by another CF or C respectively.

I understand that CF and C are more "important" defensive positions, I just don't see how defensive positions matter when calculating the wins a player has over a replacement player of their same position.

[tiger337]

One thing I don't particularly understand is how, or the extent of which, ballpark factors are used (not even necessarily WAR but stats like OPS+ also).

This year Miguel Cabrera had an OPS of 999 with an OPS+ of 165 while Mike Trout had an OPS of 963 and an OPS+ of 171, now was Angel stadium such a disadvantage that it should account for more than 36 points of OPS? I just personally don't think that if Cabrera played in Angel Stadium that he would end the season with an OPS lower than Trout's.

And one other question...shouldn't there be a place for BABIP in WAR? It seems unfair to use BABIP to discredit anomalous seasons (Jackson's 2010 campaign) but not have that reflected in a players WAR. IMO, if a player got exceptionally "lucky" or "unlucky" their WAR should be affected as such. During the entire Cabrera vs. Trout MVP debate it seemed like their was a lack of concern over Trout's .380+ BABIP (especially considering in 40 games in 2011 Trout posted a .247 BABIP).

I would be interested to see what Trout's WAR would be taking into account BABIP and without the added 40+ OPS points given because of "ballpark factors."

Ballpark factors try to tell you what the average player would have done playing in the same parks as the given player. It's true that some players react differently a ballpark than the average player would but it's not practical to have a different factor for every player when calculating WAR for every player in baseball. You could eliminate ballpark factors altogether, but I don't think it's fair to compare a Rockie or Padre that way. When you are just looking at a few players to determine awards, it's reasonable to try doing it a few different ways. In the case of Trout versus Cabrera, I think it's reasonable to consider removing the ballpark factor if you think that Comerica Park does not help Cabrera. Just looking at one final number is not a good way to do analysis.

I would not want BABIP to be included in discussion of awards. BABIP is a stat used to predict the future. I use WAR more for looking at the past. Once the season is over, it doesn't matter whether a player was luck or unlucky. It's the final results that count.

[tiger337]

Another thing I don't understand about WAR and would love some clarification is why defensive position is incorporated. Isn't the entire purpose of WAR to calculate the wins over a replacement player? A first baseman is never going to be replaced by a CF in a 1 for 1 swap, so why should the first baseman get a -12.5? Or to put it another way, I don't see why a CF or C get bonuses to their position when they will be replaced by another CF or C respectively.

I understand that CF and C are more "important" defensive positions, I just don't see how defensive positions matter when calculating the wins a player has over a replacement player of their same position.

I'm thinking of changing the name of WAR to something else when I use it because people focus too much on the word "replacement" and it really isn't a huge part of the whole equation (Maybe I'll call it Wins Added). The replacement part of WAR is really just giving players credit for playing time. In the case of Cabrera versus Trout, you are just giving Cabrera some extra points because he had more plate appearances.

The extra positional points are not part of the replacement. They are part of a player's defensive value. A shortstop adds more value than a first baseman simply because he's a shortstop and not as many players can play shortstop as first base. If Player A is an average defensive shortstop and player B is an average defensive first baseman, you want to give Player A more defensive points than Player B.

[tiger337]

I personally think park factor calculation is deeply flawed as well. It should not track the make-up of the home team, but it seems to anyway. There is some dynamic there that is unseen but effective. Further, they certainly should not be recalculated every year unless a park's architecture has actually changed. They should be allowed to average over the long term as long as the park stays the same - otherwise what you end up measuring in reality is mostly the annual departure of the local weather from average - and weather is too mercurial to try to credit or debit players for every year.

I like ballpark factors which are calculated over a lot of years (like five) and also regressed to the mean. That eliminates a lot of the noise of year to year changes. I actually think BP factors can legitimately change over time though even without a change in architecture. If one park changes its configuration, then that affects the BP factors for all the other parks in comparison.

[Gehringer_2]

If one park changes its configuration, then that affects the BP factors for all the other parks in comparison.

Yes, certainly you are correct there would have to be a renormalization if any one park changes significantly. But even given that one park may move through the ranking from one spot to another changing the absolute factors, the relative rankings should pretty much stay put from year to year other than the park that actually did change.

[EchO]

I like ballpark factors which are calculated over a lot of years (like five) and also regressed to the mean. That eliminates a lot of the noise of year to year changes. I actually think BP factors can legitimately change over time though even without a change in architecture. If one park changes its configuration, then that affects the BP factors for all the other parks in comparison.

I don't know if my issue with ballpark factors is necessarily that it is used, but maybe more so the extent to which it is used. I know it is a single case but I just honestly don't think that if Miguel Cabrera and Mike Trout were on the same team (and played 81 of their games in Angel Stadium) that Cabrera would have a lower OPS than Trout as Trout's greater OPS+ would suggest.

[EchO]

I'm thinking of changing the name of WAR to something else when I use it because people focus too much on the word "replacement" and it really isn't a huge part of the whole equation (Maybe I'll call it Wins Added). The replacement part of WAR is really just giving players credit for playing time. In the case of Cabrera versus Trout, you are just giving Cabrera some extra points because he had more plate appearances.

The extra positional points are not part of the replacement. They are part of a player's defensive value. A shortstop adds more value than a first baseman simply because he's a shortstop and not as many players can play shortstop as first base. If Player A is an average defensive shortstop and player B is an average defensive first baseman, you want to give Player A more defensive points than Player B.

Please don't think this is me being rude, but what then is the point of this statistic? In what scenarios are you going to be needing to compare a shortstop with a first baseman? The MVP award? If you are looking to replace a player on your team there are very few scenarios I can think of when you would be replacing a 1B with a shortstop or vice versa. I agree that an average defensive shortstop should be given more "defensive points" than an average defensive first baseman...but why are you comparing shortstops to first baseman?

Maybe I just need to buy your book to get a better understand of WAR and what it is used for, I'm a huge fan of more advanced statistics (OPS/UZR/etc.) but there are some like WAR that I just don't seem to get. I do appreciate your replies though, they are very insightful and I am admittingly pretty ignorant on the subject of WAR and what it's used for.

[TigerNation]

Please don't think this is me being rude, but what then is the point of this statistic? In what scenarios are you going to be needing to compare a shortstop with a first baseman? The MVP award? If you are looking to replace a player on your team there are very few scenarios I can think of when you would be replacing a 1B with a shortstop or vice versa. I agree that an average defensive shortstop should be given more "defensive points" than an average defensive first baseman...but why are you comparing shortstops to first baseman?

Maybe I just need to buy your book to get a better understand of WAR and what it is used for, I'm a huge fan of more advanced statistics (OPS/UZR/etc.) but there are some like WAR that I just don't seem to get. I do appreciate your replies though, they are very insightful and I am admittingly pretty ignorant on the subject of WAR and what it's used for.

The point of the statistic is to factor in all variables, fielding, hitting, baserunning, position, and come up with a number that quantifies how much value each player provides, to show who the best players are.

[Corky]

Please don't debate the value. But for a guy who has no head for math how is it calculated? How do you determine "replacement level?"

I am a huge fan of new thinking and saber philosophies but my limited knowledge of math makes it tough for me...

Bigger question is What is it good for?

[tiger337]

Please don't think this is me being rude, but what then is the point of this statistic? In what scenarios are you going to be needing to compare a shortstop with a first baseman? The MVP award? If you are looking to replace a player on your team there are very few scenarios I can think of when you would be replacing a 1B with a shortstop or vice versa. I agree that an average defensive shortstop should be given more "defensive points" than an average defensive first baseman...but why are you comparing shortstops to first baseman?

Maybe I just need to buy your book to get a better understand of WAR and what it is used for, I'm a huge fan of more advanced statistics (OPS/UZR/etc.) but there are some like WAR that I just don't seem to get. I do appreciate your replies though, they are very insightful and I am admittingly pretty ignorant on the subject of WAR and what it's used for.

I can think of a few scenarios where I want to compare players at different positions - Awards, Hall of Fame, potential trades. In all of those cases, I want to know whether one player did more to help his team win games. WAR is one way of looking at that. Where I frequently look at WAR is in making historical comparisons. Let's say I want to put together a list of the top fifty Tigers ever. WAR would be a good place to start because it tries to combine all of a player's skills into one value which allows for ranking. I wouldn't stop there of course. I'd look at other things and make adjustments to the list, but WAR helps me get a general idea of who should be on the list.

[tiger337]

I don't know if my issue with ballpark factors is necessarily that it is used, but maybe more so the extent to which it is used. I know it is a single case but I just honestly don't think that if Miguel Cabrera and Mike Trout were on the same team (and played 81 of their games in Angel Stadium) that Cabrera would have a lower OPS than Trout as Trout's greater OPS+ would suggest.

The standard WAR is intended to look at what an average player would do in a particular park. When I do my own WAR calculations, I generally use more conservative estimates of ballpark factors, so that only the extreme parks have a really big effect. When I did my own quick WAR comparing Cabrera and Trout, the difference in BP factors was very small.

[Charles Liston]

Bigger question is What is it good for?

Good God y'all.

[Mr. Bigglesworth]

A first baseman is never going to be replaced by a CF in a 1 for 1 swap, so why should the first baseman get a -12.5? Or to put it another way, I don't see why a CF or C get bonuses to their position when they will be replaced by another CF or C respectively.

I understand that CF and C are more "important" defensive positions, I just don't see how defensive positions matter when calculating the wins a player has over a replacement player of their same position.

It matters because a replacement level 1B hits better than a replacement level C or CF.

I think a better question is why wouldn't a guy who plays a key defensive position receive some form of credit for manning the position at a league acceptable level versus another who fields a less demanding position at a league acceptable level? There are a far fewer guys who can do the former than the latter, making it more scarce (and more valuable).

[EchO]

It matters because a replacement level 1B hits better than a replacement level C or CF.

I think a better question is why wouldn't a guy who plays a key defensive position receive some form of credit for manning the position at a league acceptable level versus another who fields a less demanding position at a league acceptable level? There are a far fewer guys who can do the former than the latter, making it more scarce (and more valuable).

I said that CF is obviously a more important defensive position...I was just more hung up on the "replacement" aspect. You will NEVER replace a 1B with a CF so to give them additional (or less) "points" seems useless. I just didn't understand the use for WAR because in a trade scenario if you need to replace a 2B you won't do so by a getting a CF...you may upgrade at CF to compensate for a loss but in doing so you are replacing both CF and 2B which would essentially cancel out the need for additional or less points due to defensive position. But Lee clarified that WAR is less about replacement than it is about overall ranking for awards and such.

And btw, thanks for being patient in your replies Lee, I'm not trying to instigate and I know WAR/advanced statistics can be a touchy subject at times.

[Sven Draconian]

I said that CF is obviously a more important defensive position...I was just more hung up on the "replacement" aspect. You will NEVER replace a 1B with a CF so to give them additional (or less) "points" seems useless. I just didn't understand the use for WAR because in a trade scenario if you need to replace a 2B you won't do so by a getting a CF...you may upgrade at CF to compensate for a loss but in doing so you are replacing both CF and 2B which would essentially cancel out the need for additional or less points due to defensive position. But Lee clarified that WAR is less about replacement than it is about overall ranking for awards and such.

And btw, thanks for being patient in your replies Lee, I'm not trying to instigate and I know WAR/advanced statistics can be a touchy subject at times.

A replacement level first baseman is still going to be a good hitter. There are a lot of guys sitting in AAA and on the free agent market who could walk onto a big league team and be an average(ish) hitter. Because the defensive requirements for the position are minimal, there are a lot of big, strong dudes who can hit a baseball a long way and mange the position. If a team's first baseman goes down, they do not replace him with a light hitting middle infielder; there is an ample amount of league average batters ready to step in. That is why to be a "good" first baseman you need to be an exception hitter.

A replacement level CF is probably a guy who can field the position well, but can't stick in the major leagues as a hitter. (un)luckly for us as Tigers fans, we got to experience the exact definition of a replacement level player in Quintein Berry. He can just barely hold the job defensively and can't hit a lick.

What WAR measures is the relative difference between Quintin Berry and Austin Jackson vs the relative difference between Alex Avila and Bryan Holaday, Prince Fielder and Brad Eldred; Victor Martinez and Delmon Young, Torii Hunter vs Brennan Boesch ect. ect).

So when you are comparing players for a trade; lets say we wanted to deal Rick Porcello (who is basically a 1.5 - 2.0 WAR pitcher) if we get back a player who has a WAR of 2.5 - 3.0, we can consider that an upgrade. If we get back a player with a WAR under 1.0, it's a loss.

[EchO]

A replacement level first baseman is still going to be a good hitter. There are a lot of guys sitting in AAA and on the free agent market who could walk onto a big league team and be an average(ish) hitter. Because the defensive requirements for the position are minimal, there are a lot of big, strong dudes who can hit a baseball a long way and mange the position. If a team's first baseman goes down, they do not replace him with a light hitting middle infielder; there is an ample amount of league average batters ready to step in. That is why to be a "good" first baseman you need to be an exception hitter.

A replacement level CF is probably a guy who can field the position well, but can't stick in the major leagues as a hitter. (un)luckly for us as Tigers fans, we got to experience the exact definition of a replacement level player in Quintein Berry. He can just barely hold the job defensively and can't hit a lick.

What WAR measures is the relative difference between Quintin Berry and Austin Jackson vs the relative difference between Alex Avila and Bryan Holaday, Prince Fielder and Brad Eldred; Victor Martinez and Delmon Young, Torii Hunter vs Brennan Boesch ect. ect).

So when you are comparing players for a trade; lets say we wanted to deal Rick Porcello (who is basically a 1.5 - 2.0 WAR pitcher) if we get back a player who has a WAR of 2.5 - 3.0, we can consider that an upgrade. If we get back a player with a WAR under 1.0, it's a loss.

You are misunderstanding my argument.

Okay, lets say we use your example of Rick Porcello...say we get a C back in return. The catcher has a defensive point total of (+12.5) and he would replace the previous catcher (Avila in the Tigers' scenario) who also had a defensive point total of (+12.5) so the two would cancel out. My point is, even if you trade a pitcher for a fielder that acquired fielder will than replace the incumbent player at said position and nullify the defensive position "points."

Either way, I hope I made myself more clear.

[waynr]

Bigger question is What is it good for?

Absolutely nothing

[RandyMarsh]

Absolutely nothing

What's batting average good for? It doesn't tell you how many walks a guy has, how many HRs a guy has, how many doubles a guy has, how many triples he has, how many outs he makes, how good he runs the bases, or how good his defense is. So to me it tells very little about a player yet people continue to use it.

[Sven Draconian]

You are misunderstanding my argument.

Okay, lets say we use your example of Rick Porcello...say we get a C back in return. The catcher has a defensive point total of (+12.5) and he would replace the previous catcher (Avila in the Tigers' scenario) who also had a defensive point total of (+12.5) so the two would cancel out. My point is, even if you trade a pitcher for a fielder that acquired fielder will than replace the incumbent player at said position and nullify the defensive position "points."

Either way, I hope I made myself more clear.

We are having a mis-communication in one of two ways here.

1) You are not grasping he concept of how replacement level is determined for each position. They are not assigned points because they play a position. They are not assigned points for anything (they earn runs above or below a threshold). This is determined using a formula to determine how many runs a player created for his team.

Lets use Miguel Cabrera playing 3B.
So a replacement level 3B would produce a slash line in the area of .220/.280/.350 and OPS in the range of .630 (this is guess, but I think it's pretty close). That is basically saying what a solid AAA 3rd baseman could produce at the major league level. If this replacement level player played a full season he would create around 70 runs.

So a replacement player at 3B is worth 70 runs. Miguel Cabrera created 139 runs last year, 69 runs more than the replacement player. So his Value of Replacement Player (VORP) is 69. For the sake of simplicity VORP is divided by 10 (roughly) to create WAR (Wins above Replacement. 10 runs = 1 win). So his WAR is 6.9.

WAR is just a way to measure a players performance above a certain threshold. That threshhold being, essentially, a AAA player.


2) You are making the argument that you must account for the the value above the player you are replacing. IE, while Victor Martinez is a 3 WAR player as a DH, he represents a net gain of nearly 4 WAR because he is replacing Delmon Young (roughly -1 WAR).

[tiger337]

Replacement level doesn't necessarily have to do with a player getting replaced. It is just a baseline for comparison. When we ask, "How many wins is a player worth?", we need to have a comparison. We could use average as the baseline: "How many wins is a player above an average player?" If that was the baseline, Wins would be equal to zero implying the player is worthless. That's misleading though because an average player is a good player.

You could choose zero as the baseline. How many wins is a player worth above .000/.000/.000? In that case, a player who batted .120/.150/.175 would have a positive number for wins. That doesn't make sense, because a player batting .120/.150/.175 would hurt his team. Thus they settled on a value in between average and zero.

[TigersSlappy]

Replacement level doesn't necessarily have to do with a player getting replaced. It is just a baseline for comparison. When we ask, "How many wins is a player worth?", we need to have a comparison. We could use average as the baseline: "How many wins is a player above an average player?" If that was the baseline, Wins would be equal to zero implying the player is worthless. That's misleading though because an average player is a good player.

You could choose zero as the baseline. How many wins is a player worth above .000/.000/.000? In that case, a player who batted .120/.150/.175 would have a positive number for wins. That doesn't make sense, because a player batting .120/.150/.175 would hurt his team. Thus they settled on a value in between average and zero.

Though I see the reasoning for "replacement" value, I do think it might be more understandable if calculated from zero/zero/zero.

So on a 92 win team you could see more clearly how the formula distributes the credit for the wins among the team.

[tiger337]

Though I see the reasoning for "replacement" value, I do think it might be more understandable if calculated from zero/zero/zero.

So on a 92 win team you could see more clearly how the formula distributes the credit for the wins among the team.

I think John Dewan at Baseball Info Solutions does that. At least, he was doing that a couple of years ago in the Fielding Bible. What ends up happening in the calculation is the lower you choose your baseline, the more credit you give to playing time. If zero is used as the baseline, then players get more extra credit for playing more games. A 600 PA player would benefit more from WA0 than a 300 PA player because you are assuming that the 300 missing PAs for the second player are being taken by a complete zero. That's not necessarily wrong, but it will give you different rankings of players. Most (not all) sabers agree that the best way to reward playing time is to assume that the missing at bats would be taken by a player somewhere between zero and average.

[TigersSlappy]

I think John Dewan at Baseball Info Solutions does that. At least, he was doing that a couple of years ago in the Fielding Bible. What ends up happening in the calculation is the lower you choose your baseline, the more credit you give to playing time. If zero is used as the baseline, then players get more extra credit for playing more games. A 600 PA player would benefit more from WA0 than a 300 PA player because you are assuming that the 300 missing PAs for the second player are being taken by a complete zero. That's not necessarily wrong, but it will give you different rankings of players. Most (not all) sabers agree that the best way to reward playing time is to assume that the missing at bats would be taken by a player somewhere between zero and average.

Again in some cases this would state value more clearly if the missed PA's were the result of injuries. Usually a player that can stay on the field for 162 games is more valuable than one that consistently misses 40 games a year for various reasons.

[Gehringer_2]

I think John Dewan at Baseball Info Solutions does that. At least, he was doing that a couple of years ago in the Fielding Bible. What ends up happening in the calculation is the lower you choose your baseline, the more credit you give to playing time. If zero is used as the baseline, then players get more extra credit for playing more games. A 600 PA player would benefit more from WA0 than a 300 PA player because you are assuming that the 300 missing PAs for the second player are being taken by a complete zero. That's not necessarily wrong, but it will give you different rankings of players. Most (not all) sabers agree that the best way to reward playing time is to assume that the missing at bats would be taken by a player somewhere between zero and average.

After following along on this my 2 cents would be that the most understandable baseline would not be the average call up player but the average existing MLB player, and thus a team full of 'replacement' players would have a win expectancy of 81 games (as compared to who knows how many games a team of AAA players is expected to win?) Then WAR becomes more plainly expressible as the predictor of team record above 500, which would both be very understandable, and also provide a very nice overall fan accessible track on the quality of the predictors from year to year.

If you want to do a playing time credit in this scenario, you just credit a players IP excess over the average for the position in the MLB rather than a zero or other baseline - the assumption basis being the guy is replaced by the average major leaguer at his position rather than the average call up guy.

It seems that the current scheme is mixing two concepts that don't need to be mixed: which are) how a team can actually replace a player during the season; and) how good a player is. In the latter case you are most interested in a player's comparison to his peers. The two bases don't need to be mixed at all. Why build the relatively hard to access increment from average AAA to average MLB player into the system for evaluation of MLB players at all?

[tiger337]

After following along on this my 2 cents would be that the most understandable baseline would not be the average call up player but the average existing MLB player, and thus a team full of 'replacement' players would have a win expectancy of 81 games (as compared to who knows how many games a team of AAA players is expected to win?) Then WAR becomes more plainly expressible as team record above 500, which would both be very understandable, and also provide a very nice overall fan accessible track on the quality of the predictors from year to year.

I think WAA makes sense for the average fan who is not attempting to play General Manager. I believe a lot of fans think: "Do we have enough Above Average players to make the playoffs?" rather than thinking about about replacement level. They answer two different questions and there's no reason why we can't have both WAA and WAR.