Going to WAR...for Classic0

[Goldfan]

Posted this in a different thread, but perfect for here…..

“Heyward was by far the Cards best player in 2015”

Carp
R 101
2b 44
HR 28
RBI 84
.272
.365
.505
.871
135 OPS+

Heyward
R79
2b 33
HR14
RBi 60
.293
.359
.439
.797
117 OPS+

Defense

Carp
CH 370
PO 102
A 254

Heyward
CH303
PO 290
A 10

Carp 4.8WAR
Heyward 7WAR

Now we all know that Carp wasn’t Arenado at 3b, but it is a much more difficult position than RF, with many more chances producing many more outs.
Carps offense was appreciably better and yet theres a HUGE disparity in the WAR #. If you weren’t aware of the giant holes in WAR you’d think Heyward outhit Carp by an incredible margin and Heyward must’ve saved at least as many games with his glove that Carp lost…….

Of course it depends on which WAR model/stat you use. The differences between them, for position players at least, is mostly on the defensive side of things.

The numbers you listed above are baseball-reference's WAR stats, or bWAR. bWAR uses DRS as it's defensive component while fWAR, fangraph's WAR model/stat, uses UZR. DRS typically has a wider spread of defensive metrics than UZR so you will see larger outliers, and Heyward is definitely an outlier defender (at least in 2015 he was).

WAR at its core is batting runs (above or below average) + base running runs + defensive runs + positional adjustment. There's adjustments for league and playing time but those can mostly be ignored when comparing teammates from the same year.

Batting runs 2015:
Carpenter +30.6 (+28 BBR)
Heyward +15 (+14 BBR)

I included the baseball reference number in parenthesis, the two models are in agreement at least in terms of direction and magnitude. Heyward was above average at the plate (117 OPS+, 121 wRC+) while Carpenter was about 20 points above that in both metrics (135 OPS+, 140 wRC+). That's about double the gap to average (100 is average for both) so his batting runs above average is about twice Heyward's. That should make sense.

Base running runs 2015:
Carpenter +1.3 (-1 BBR)
Heyward +7.2 (+6 BBR)

Base running doesn't generate anywhere close to the value of hitting or defense, but Heyward was much better on the bases both by standard metrics (23 SB vs. 4 for Carpenter, 3 CS for each) and advanced (shown above). Advanced metrics show a 6-7 run advantage for Heyward, given he stole 19 more bases and was caught the same amount of times I'd say 6-7 runs seems about right. That reduces the "offense runs" advantage Carpenter had at the place from 14-15 runs down to 8-9 runs overall.

Note: Baseball reference also has a Rdp metric, runs saved by avoiding double plays (Heyward was +5, Carpenter was 0). Fangraphs includes this in their base running runs shown above, so I guess the models actually start to differ a bit here, by 5 runs or so. So BBR it was +11 for Heyward and -1 for Carpenter, a 12 run difference, not a six run difference (that's 0.6 WAR difference right there).

Offense runs 2015:
Carpenter +31.9
Heyward +22.2

This is simply batting runs + base running runs, and the gap, at fangraphs, has been reduced to 9.7 runs. At baseball reference the gap would be 14 - 7- 5 = 2 runs.

Now we get to the point that always drives differences and makes people question things, because it is really hard to quantify defense just watching as a fan.

Fielding runs 2015:
Carpenter -4.8 (-3 BBR)
Heyward +17.4 (+28 BBR)

These metrics are relative to your position. Both UZR and DRS thinks Carpenter was a little below average (5 runs, and 3 runs, respectively), and they both think Heyward was great defensively but DRS thinks he was really great (a whole 10 runs saved better than UZR does, 28 vs. 17 runs saved, respectively). 10 runs is a win, in terms of WAR, so this is a huge part of the gap between the two metrics.

Heyward saved 22 extra runs compared to Carpenter, using UZR, and 31 runs using DRS.

Positional adjustment 2015:
Carpenter +2 (+3 BBR)
Heyward -6 (-5 BBR)

In both models there is a positional adjustment that adds 8 runs (almost a whole win) to Carpenter, relative to Heyward. That cuts those defensive gaps from 22 and 31 to 14 and 23 runs. That's 1.4 to 2.3 wins in WAR.

As we saw above on offense Heyward had a 9.7 run deficit on fangraphs, and a 2 run deficit on baseball reference.

Offense + defense (including position adj) runs 2015:
Carpenter +31.9 - 2.8 = 29.1 RAA (runs above average)
Heyward +22.2 + 11.4 = 33.6 RAA

Using baseball reference you get:
Carpenter +27 + 0 (+3 - 3) = 27 RAA
Heyward +25 + 23 = 48 RAA

This is runs above average. They had similar playing time so to drop the baseline from average to replacement level 18-20 runs (Carpenter got one extra run due to more playing time, he hit leadoff). Then you divide by 10 to convert runs to wins.

So fangraphs thinks Heyward was 3.5 runs or so better than Carpenter, or 0.3 fWAR (5.6 fWAR for Heyward, 5.3 fWAR for Carpenter).

Baseball reference thinks the gap is bigger because of how it views defense (11 extra runs saved, 2 extra for Carpenter) and base running (5 extra runs for avoiding GIDP by Heyward), leading to a gap about 1.5 WAR larger than fangraphs (it had an extra run in the batting gap, and another in regular base running as well) so the gap is about 1.7 WAR larger, plus the 0.3, which is really 3.5 runs difference, on fangraphs and you get an expected gap of about 2 WAR between Heyward and Carpenter in 2015, which is what baseball reference shows . . . 7.0 vs. 4.8 bWAR.

RBI, excellent breakdown of the equations.
Only thing is, baseball isn’t open to interpretation. An out is an out, a strike is a strike, a hit is a hit, a run is a run
This great disparity between the competing WAR houses completely invalidates WAR
Stats have Been recorded in the books since inception and those stats are what the player physically accomplished. Not what they might have or could have or a divider, multiplier, equator skewed to Offense or Defense.

[NYCardsFan]

In 12,232 words, Melville did not answer my question (no surprise).

Oh, but I did.
You did not see it.
I will not judge as to whether that is a surprise.
By the way, your word count stat is incorrect.

He’s not wrong. You did some tribal dance around the question, but in the end, never came around to answering it.

I most certainly answered with specificity and clarity - and also illustrated by an analogy.

Ultimately, your answer was:

Give me the roster and the standings - and I know all that is knowable.

Which is in no way was a specific or clear answer to his question.

Meanwhile, Melville compares individual players (across multiple positions--and teams) on here all the time and routinely deems one superior to the other(s). Further, he regularly opines on individual contract values and the relative values involved in various trade proposals. How can he do all of that if the individual contributions and values of players are "unknowable" and therefore incomparable? Almost everything he posts on this forum directly undercuts his nihilistic "symphony analogy."

Pure sophistry and bad faith all the way to the bottom.

[Ozziesfan41]

In 12,232 words, Melville did not answer my question (no surprise).

Oh, but I did.
You did not see it.
I will not judge as to whether that is a surprise.
By the way, your word count stat is incorrect.

No one sees an answer. What specific number or stat do you do to compare value in trades or drafts? That is what WAR attempts to provide. You may say it isn't a stat (true) but it uses stats to create a singular number for comparison and is as far as I can see the best thing out there.

The answer is that everybody uses WAR either directly by using the metrics available or indirectly by subjectively creating their own version of it. And the real kicker is, everybody has been doing this for decades.

It is exactly how we answer questions like the ones you asked. "who should we draft?" "Should I trade my ace for a MOTO bat who has 40 HR, 100 RBI but is a slug on the bases and a 1B, or should I go after the 20 HR, 40 SB GG SS instead?" Before WAR came around and objectively measured all components of baseball (with a repeatable but albeit imperfect method) we would have to weight power, RBIs, base running (usually steals), and defense (position and usually awards won) in our heads to determine if a trade was "fair value" or not. That is exactly what WAR does, just in a structured framework to allow you to reduce all those inputs to a single number, to make a decision. It's just measuring skills that lead to run differential and stacking them together. Then you put the stacks on a scale to see if a trade is fair or in your favor. As I mentioned prior to WAR the "weights" on the scale were GG defense, .300 hitter, 100 RBI bat, 30 SBs, 40 HRs, workhorse SP, shutdown reliever, etc.

WAR is fine unless you use it as the definitive stat. WAR is why people were saying Heyward who was a 13 homerun 60 RBI 79 Run outfielder was worth 200 million because he played good defense which was completely absurd its also why Jeffy was saying Bader was better than Willie McGee lol its also why people were saying sign Oneil to a long term contract because he is a WAR god when in reality if you actually looked hes rarely on the field so it would be dumb to sign him to a long term contract but yet the WAR people were like hes the greatest! WAR is fine but if you look at it alone you will be wrong about a lot of players and make terrible decisions like signing heyward to 200 million or signing Oneil to a long term contract

[An Old Friend]

Posted this in a different thread, but perfect for here…..

“Heyward was by far the Cards best player in 2015”

Carp
R 101
2b 44
HR 28
RBI 84
.272
.365
.505
.871
135 OPS+

Heyward
R79
2b 33
HR14
RBi 60
.293
.359
.439
.797
117 OPS+

Defense

Carp
CH 370
PO 102
A 254

Heyward
CH303
PO 290
A 10

Carp 4.8WAR
Heyward 7WAR

Now we all know that Carp wasn’t Arenado at 3b, but it is a much more difficult position than RF, with many more chances producing many more outs.
Carps offense was appreciably better and yet theres a HUGE disparity in the WAR #. If you weren’t aware of the giant holes in WAR you’d think Heyward outhit Carp by an incredible margin and Heyward must’ve saved at least as many games with his glove that Carp lost…….

Of course it depends on which WAR model/stat you use. The differences between them, for position players at least, is mostly on the defensive side of things.

The numbers you listed above are baseball-reference's WAR stats, or bWAR. bWAR uses DRS as it's defensive component while fWAR, fangraph's WAR model/stat, uses UZR. DRS typically has a wider spread of defensive metrics than UZR so you will see larger outliers, and Heyward is definitely an outlier defender (at least in 2015 he was).

WAR at its core is batting runs (above or below average) + base running runs + defensive runs + positional adjustment. There's adjustments for league and playing time but those can mostly be ignored when comparing teammates from the same year.

Batting runs 2015:
Carpenter +30.6 (+28 BBR)
Heyward +15 (+14 BBR)

I included the baseball reference number in parenthesis, the two models are in agreement at least in terms of direction and magnitude. Heyward was above average at the plate (117 OPS+, 121 wRC+) while Carpenter was about 20 points above that in both metrics (135 OPS+, 140 wRC+). That's about double the gap to average (100 is average for both) so his batting runs above average is about twice Heyward's. That should make sense.

Base running runs 2015:
Carpenter +1.3 (-1 BBR)
Heyward +7.2 (+6 BBR)

Base running doesn't generate anywhere close to the value of hitting or defense, but Heyward was much better on the bases both by standard metrics (23 SB vs. 4 for Carpenter, 3 CS for each) and advanced (shown above). Advanced metrics show a 6-7 run advantage for Heyward, given he stole 19 more bases and was caught the same amount of times I'd say 6-7 runs seems about right. That reduces the "offense runs" advantage Carpenter had at the place from 14-15 runs down to 8-9 runs overall.

Note: Baseball reference also has a Rdp metric, runs saved by avoiding double plays (Heyward was +5, Carpenter was 0). Fangraphs includes this in their base running runs shown above, so I guess the models actually start to differ a bit here, by 5 runs or so. So BBR it was +11 for Heyward and -1 for Carpenter, a 12 run difference, not a six run difference (that's 0.6 WAR difference right there).

Offense runs 2015:
Carpenter +31.9
Heyward +22.2

This is simply batting runs + base running runs, and the gap, at fangraphs, has been reduced to 9.7 runs. At baseball reference the gap would be 14 - 7- 5 = 2 runs.

Now we get to the point that always drives differences and makes people question things, because it is really hard to quantify defense just watching as a fan.

Fielding runs 2015:
Carpenter -4.8 (-3 BBR)
Heyward +17.4 (+28 BBR)

These metrics are relative to your position. Both UZR and DRS thinks Carpenter was a little below average (5 runs, and 3 runs, respectively), and they both think Heyward was great defensively but DRS thinks he was really great (a whole 10 runs saved better than UZR does, 28 vs. 17 runs saved, respectively). 10 runs is a win, in terms of WAR, so this is a huge part of the gap between the two metrics.

Heyward saved 22 extra runs compared to Carpenter, using UZR, and 31 runs using DRS.

Positional adjustment 2015:
Carpenter +2 (+3 BBR)
Heyward -6 (-5 BBR)

In both models there is a positional adjustment that adds 8 runs (almost a whole win) to Carpenter, relative to Heyward. That cuts those defensive gaps from 22 and 31 to 14 and 23 runs. That's 1.4 to 2.3 wins in WAR.

As we saw above on offense Heyward had a 9.7 run deficit on fangraphs, and a 2 run deficit on baseball reference.

Offense + defense (including position adj) runs 2015:
Carpenter +31.9 - 2.8 = 29.1 RAA (runs above average)
Heyward +22.2 + 11.4 = 33.6 RAA

Using baseball reference you get:
Carpenter +27 + 0 (+3 - 3) = 27 RAA
Heyward +25 + 23 = 48 RAA

This is runs above average. They had similar playing time so to drop the baseline from average to replacement level 18-20 runs (Carpenter got one extra run due to more playing time, he hit leadoff). Then you divide by 10 to convert runs to wins.

So fangraphs thinks Heyward was 3.5 runs or so better than Carpenter, or 0.3 fWAR (5.6 fWAR for Heyward, 5.3 fWAR for Carpenter).

Baseball reference thinks the gap is bigger because of how it views defense (11 extra runs saved, 2 extra for Carpenter) and base running (5 extra runs for avoiding GIDP by Heyward), leading to a gap about 1.5 WAR larger than fangraphs (it had an extra run in the batting gap, and another in regular base running as well) so the gap is about 1.7 WAR larger, plus the 0.3, which is really 3.5 runs difference, on fangraphs and you get an expected gap of about 2 WAR between Heyward and Carpenter in 2015, which is what baseball reference shows . . . 7.0 vs. 4.8 bWAR.

This is really well done and explained. Kudos.

Couple of notes on baserunning to add...
Carpenter took an extra base 44% of the time and made 11 outs on the bases
Heyward took an extra base 57% of the time and made 8 outs on the bases

All of these little things add up when quantifying the difference in value.

For context on defense, just 13 outfielders of 72 that played at least 750 innings saved 10 runs. By Fielding Bible, Heyward was 3rd overall at 26. Kiermaier blew everyone away with 38, and Ender Inciarte was 2nd with 27.

Of the 25 3B that played at least 700 innings at the position, Carpenter was 16th at -2. There were 4 3B that saved at least 10 runs, with Arenado and Machado leading the way with 22 and 18, respectively.

* took 750 and 700 innings to get roughly an equal positional player sample size *

Again, appreciate your post, very thorough.

[rbirules]

In 12,232 words, Melville did not answer my question (no surprise).

Oh, but I did.
You did not see it.
I will not judge as to whether that is a surprise.
By the way, your word count stat is incorrect.

No one sees an answer. What specific number or stat do you do to compare value in trades or drafts? That is what WAR attempts to provide. You may say it isn't a stat (true) but it uses stats to create a singular number for comparison and is as far as I can see the best thing out there.

The answer is that everybody uses WAR either directly by using the metrics available or indirectly by subjectively creating their own version of it. And the real kicker is, everybody has been doing this for decades.

It is exactly how we answer questions like the ones you asked. "who should we draft?" "Should I trade my ace for a MOTO bat who has 40 HR, 100 RBI but is a slug on the bases and a 1B, or should I go after the 20 HR, 40 SB GG SS instead?" Before WAR came around and objectively measured all components of baseball (with a repeatable but albeit imperfect method) we would have to weight power, RBIs, base running (usually steals), and defense (position and usually awards won) in our heads to determine if a trade was "fair value" or not. That is exactly what WAR does, just in a structured framework to allow you to reduce all those inputs to a single number, to make a decision. It's just measuring skills that lead to run differential and stacking them together. Then you put the stacks on a scale to see if a trade is fair or in your favor. As I mentioned prior to WAR the "weights" on the scale were GG defense, .300 hitter, 100 RBI bat, 30 SBs, 40 HRs, workhorse SP, shutdown reliever, etc.

WAR is fine unless you use it as the definitive stat. WAR is why people were saying Heyward who was a 13 homerun 60 RBI 79 Run outfielder was worth 200 million because he played good defense which was completely absurd its also why Jeffy was saying Bader was better than Willie McGee lol its also why people were saying sign Oneil to a long term contract because he is a WAR god when in reality if you actually looked hes rarely on the field so it would be dumb to sign him to a long term contract but yet the WAR people were like hes the greatest! WAR is fine but if you look at it alone you will be wrong about a lot of players and make terrible decisions like signing heyward to 200 million or signing Oneil to a long term contract

I'm not saying to use WAR, and only WAR, to make any and all decisions, that's a straw man argument, and I wouldn't suggest doing that with any stat. I'm saying it's a best starting point, of any metric that we have. It then helps to know what goes into it, where points of error (or difference of opinion) could come into play, as well as the sustainability of various components when using various metrics to answer specific questions. I don't think that's a particularly controversial stance to have.

[An Old Friend]

RBI, excellent breakdown of the equations.
Only thing is, baseball isn’t open to interpretation. An out is an out, a strike is a strike, a hit is a hit, a run is a run
This great disparity between the competing WAR houses completely invalidates WAR
Stats have Been recorded in the books since inception and those stats are what the player physically accomplished. Not what they might have or could have or a divider, multiplier, equator skewed to Offense or Defense.

I'm not sure how one could ever say that baseball isn't open to interpretation. Hits, errors, balls, strikes, out, safe... have all been an umpire's and an official scorer's interpretation since the beginning. All of those interpretations affect stats and outcomes.

[rbirules]

RBI, excellent breakdown of the equations.
Only thing is, baseball isn’t open to interpretation. An out is an out, a strike is a strike, a hit is a hit, a run is a run
This great disparity between the competing WAR houses completely invalidates WAR
Stats have Been recorded in the books since inception and those stats are what the player physically accomplished. Not what they might have or could have or a divider, multiplier, equator skewed to Offense or Defense.

I'm not sure how one could ever say that baseball isn't open to interpretation. Hits, errors, balls, strikes, out, safe... have all been an umpire's and an official scorer's interpretation since the beginning. All of those interpretations affect stats and outcomes.

Exactly this.

[Futuregm2]

In 12,232 words, Melville did not answer my question (no surprise).

Oh, but I did.
You did not see it.
I will not judge as to whether that is a surprise.
By the way, your word count stat is incorrect.

No one sees an answer. What specific number or stat do you do to compare value in trades or drafts? That is what WAR attempts to provide. You may say it isn't a stat (true) but it uses stats to create a singular number for comparison and is as far as I can see the best thing out there.

The answer is that everybody uses WAR either directly by using the metrics available or indirectly by subjectively creating their own version of it. And the real kicker is, everybody has been doing this for decades.

It is exactly how we answer questions like the ones you asked. "who should we draft?" "Should I trade my ace for a MOTO bat who has 40 HR, 100 RBI but is a slug on the bases and a 1B, or should I go after the 20 HR, 40 SB GG SS instead?" Before WAR came around and objectively measured all components of baseball (with a repeatable but albeit imperfect method) we would have to weight power, RBIs, base running (usually steals), and defense (position and usually awards won) in our heads to determine if a trade was "fair value" or not. That is exactly what WAR does, just in a structured framework to allow you to reduce all those inputs to a single number, to make a decision. It's just measuring skills that lead to run differential and stacking them together. Then you put the stacks on a scale to see if a trade is fair or in your favor. As I mentioned prior to WAR the "weights" on the scale were GG defense, .300 hitter, 100 RBI bat, 30 SBs, 40 HRs, workhorse SP, shutdown reliever, etc.

WAR is fine unless you use it as the definitive stat. WAR is why people were saying Heyward who was a 13 homerun 60 RBI 79 Run outfielder was worth 200 million because he played good defense which was completely absurd its also why Jeffy was saying Bader was better than Willie McGee lol its also why people were saying sign Oneil to a long term contract because he is a WAR god when in reality if you actually looked hes rarely on the field so it would be dumb to sign him to a long term contract but yet the WAR people were like hes the greatest! WAR is fine but if you look at it alone you will be wrong about a lot of players and make terrible decisions like signing heyward to 200 million or signing Oneil to a long term contract

Who called O’Neill a WAR God besides Joel? He has 2 seasons with a WAR above 1.5 in his career. Only 1 of which happened with us which was when people would have been calling to sign him to a long term contract. WAR would say stay away from a player like TO just like his other stats would as well.

[renostl]

In 12,232 words, Melville did not answer my question (no surprise).

Oh, but I did.
You did not see it.
I will not judge as to whether that is a surprise.
By the way, your word count stat is incorrect.

No one sees an answer. What specific number or stat do you do to compare value in trades or drafts? That is what WAR attempts to provide. You may say it isn't a stat (true) but it uses stats to create a singular number for comparison and is as far as I can see the best thing out there.

The answer is that everybody uses WAR either directly by using the metrics available or indirectly by subjectively creating their own version of it. And the real kicker is, everybody has been doing this for decades.

It is exactly how we answer questions like the ones you asked. "who should we draft?" "Should I trade my ace for a MOTO bat who has 40 HR, 100 RBI but is a slug on the bases and a 1B, or should I go after the 20 HR, 40 SB GG SS instead?" Before WAR came around and objectively measured all components of baseball (with a repeatable but albeit imperfect method) we would have to weight power, RBIs, base running (usually steals), and defense (position and usually awards won) in our heads to determine if a trade was "fair value" or not. That is exactly what WAR does, just in a structured framework to allow you to reduce all those inputs to a single number, to make a decision. It's just measuring skills that lead to run differential and stacking them together. Then you put the stacks on a scale to see if a trade is fair or in your favor. As I mentioned prior to WAR the "weights" on the scale were GG defense, .300 hitter, 100 RBI bat, 30 SBs, 40 HRs, workhorse SP, shutdown reliever, etc.

IF we can debate the value of BA, RBI, OBP, HR, etc. then debating WAR is an automatic.

Defensive metric have been argued since their inception and then given position weight to
each will only widen such a gap in the discussion.

The oddity here is that Classico never ever, ever said that he only uses WAR, in fact he states the opposite,
and Mel somehow can call a player a 4th outfielder without using any numbers.

[Ozziesfan41]

In 12,232 words, Melville did not answer my question (no surprise).

Oh, but I did.
You did not see it.
I will not judge as to whether that is a surprise.
By the way, your word count stat is incorrect.

No one sees an answer. What specific number or stat do you do to compare value in trades or drafts? That is what WAR attempts to provide. You may say it isn't a stat (true) but it uses stats to create a singular number for comparison and is as far as I can see the best thing out there.

The answer is that everybody uses WAR either directly by using the metrics available or indirectly by subjectively creating their own version of it. And the real kicker is, everybody has been doing this for decades.

It is exactly how we answer questions like the ones you asked. "who should we draft?" "Should I trade my ace for a MOTO bat who has 40 HR, 100 RBI but is a slug on the bases and a 1B, or should I go after the 20 HR, 40 SB GG SS instead?" Before WAR came around and objectively measured all components of baseball (with a repeatable but albeit imperfect method) we would have to weight power, RBIs, base running (usually steals), and defense (position and usually awards won) in our heads to determine if a trade was "fair value" or not. That is exactly what WAR does, just in a structured framework to allow you to reduce all those inputs to a single number, to make a decision. It's just measuring skills that lead to run differential and stacking them together. Then you put the stacks on a scale to see if a trade is fair or in your favor. As I mentioned prior to WAR the "weights" on the scale were GG defense, .300 hitter, 100 RBI bat, 30 SBs, 40 HRs, workhorse SP, shutdown reliever, etc.

WAR is fine unless you use it as the definitive stat. WAR is why people were saying Heyward who was a 13 homerun 60 RBI 79 Run outfielder was worth 200 million because he played good defense which was completely absurd its also why Jeffy was saying Bader was better than Willie McGee lol its also why people were saying sign Oneil to a long term contract because he is a WAR god when in reality if you actually looked hes rarely on the field so it would be dumb to sign him to a long term contract but yet the WAR people were like hes the greatest! WAR is fine but if you look at it alone you will be wrong about a lot of players and make terrible decisions like signing heyward to 200 million or signing Oneil to a long term contract

Who called O’Neill a WAR God besides Joel? He has 2 seasons with a WAR above 1.5 in his career. Only 1 of which happened with us which was when people would have been calling to sign him to a long term contract. WAR would say stay away from a player like TO just like his other stats would as well.

Jeffy was all about drooling over Oneil and Bader constantly and he was the WAR and analytics champion on here and he had his lock step people in line with him saying sign heyward for 200 million bader is better than willie and all the other nonsense he and his WAR champions would say

[Goldfan]

RBI, excellent breakdown of the equations.
Only thing is, baseball isn’t open to interpretation. An out is an out, a strike is a strike, a hit is a hit, a run is a run
This great disparity between the competing WAR houses completely invalidates WAR
Stats have Been recorded in the books since inception and those stats are what the player physically accomplished. Not what they might have or could have or a divider, multiplier, equator skewed to Offense or Defense.

I'm not sure how one could ever say that baseball isn't open to interpretation. Hits, errors, balls, strikes, out, safe... have all been an umpire's and an official scorer's interpretation since the beginning. All of those interpretations affect stats and outcomes.

Exactly this.

Sure and official outcome is the record……not some interpretation of the record……or maybe it could have been a run or an out and somehow extra credit is given…..
And as far as 2015 Carp/Heyward, even you give a wash to offense/defense skews which I doubt really exist. It’s still a wash and NOT 7WAR v 4.8WAR

[imadangman]

I don't "use" WAR but I just look at it and basically react to it based on if I think it matches up with how I look at the player.

For the Heyward/Carpenter example I'm just gonna say yeah I get it, because Carpenter's offense was pretty good but Heyward that season in particular was outstanding in defense and baserunning in particular. So it clearly checks out why WAR liked Heyward that season.

But WAR is not an end all be all stat.

Baseball reference says Nolan Arenado was worth 7.9 war in 2022, edging out Paul Goldschmidt who came in at 7.7. Does that mean Arenado was definitively better by exactly 0.2 wins? No it means they were both about equally outstanding that season. Maybe Arenado was a touch more valuable because of the 3b defensive value. I'd buy it.

Really, where WAR correlates most to an end all be all is on the business side of baseball. When you talk about 1 War being worth 9mil or 10mil whatever on free agency (but in reality 1 War is worth more like 6mil across mlb baseball including team control and arb players along with free agents). So that's the one stat that can quickly give a estimated value. And then you're going to use more stats from there.

As said above if you want to spend 180mil on a roster and expect 90 wins (which is 42 wins above a replacement level team) then you essentially need to find a way for your roster to generate 42 wins for 180mil which is more like 4.3mil per War.

So that's the equation a competent front office is looking at. How are we going to take this roster, both team control and free agent players, and squeeze those 42 wins out at a cost of 4.3mil per War. And how are we going to plan and construct for future teams to be able to do it as well. How much production can you forecast in surplus from JJ Wetherholt, etc before their salaries go up? That's the type of thing front offices are trying to figure out..

[imadangman]

As for the OP, good work Melville. I doubt ClassicO will be able to put the work you have into jacking off all over the forum with your creative writing. I'm glad you took the extra time.

[An Old Friend]

RBI, excellent breakdown of the equations.
Only thing is, baseball isn’t open to interpretation. An out is an out, a strike is a strike, a hit is a hit, a run is a run
This great disparity between the competing WAR houses completely invalidates WAR
Stats have Been recorded in the books since inception and those stats are what the player physically accomplished. Not what they might have or could have or a divider, multiplier, equator skewed to Offense or Defense.

I'm not sure how one could ever say that baseball isn't open to interpretation. Hits, errors, balls, strikes, out, safe... have all been an umpire's and an official scorer's interpretation since the beginning. All of those interpretations affect stats and outcomes.

Exactly this.

Sure and official outcome is the record……not some interpretation of the record……or maybe it could have been a run or an out and somehow extra credit is given…..
And as far as 2015 Carp/Heyward, even you give a wash to offense/defense skews which I doubt really exist. It’s still a wash and NOT 7WAR v 4.8WAR

Are you somehow disagreeing with my assertion that interpretation has always existed in baseball?

Cause you could just say, "that's true" rather than this short rambling whatever it is.

[Goldfan]

Posted this in a different thread, but perfect for here…..

“Heyward was by far the Cards best player in 2015”

Carp
R 101
2b 44
HR 28
RBI 84
.272
.365
.505
.871
135 OPS+

Heyward
R79
2b 33
HR14
RBi 60
.293
.359
.439
.797
117 OPS+

Defense

Carp
CH 370
PO 102
A 254

Heyward
CH303
PO 290
A 10

Carp 4.8WAR
Heyward 7WAR

Now we all know that Carp wasn’t Arenado at 3b, but it is a much more difficult position than RF, with many more chances producing many more outs.
Carps offense was appreciably better and yet theres a HUGE disparity in the WAR #. If you weren’t aware of the giant holes in WAR you’d think Heyward outhit Carp by an incredible margin and Heyward must’ve saved at least as many games with his glove that Carp lost…….

Of course it depends on which WAR model/stat you use. The differences between them, for position players at least, is mostly on the defensive side of things.

The numbers you listed above are baseball-reference's WAR stats, or bWAR. bWAR uses DRS as it's defensive component while fWAR, fangraph's WAR model/stat, uses UZR. DRS typically has a wider spread of defensive metrics than UZR so you will see larger outliers, and Heyward is definitely an outlier defender (at least in 2015 he was).

WAR at its core is batting runs (above or below average) + base running runs + defensive runs + positional adjustment. There's adjustments for league and playing time but those can mostly be ignored when comparing teammates from the same year.

Batting runs 2015:
Carpenter +30.6 (+28 BBR)
Heyward +15 (+14 BBR)

I included the baseball reference number in parenthesis, the two models are in agreement at least in terms of direction and magnitude. Heyward was above average at the plate (117 OPS+, 121 wRC+) while Carpenter was about 20 points above that in both metrics (135 OPS+, 140 wRC+). That's about double the gap to average (100 is average for both) so his batting runs above average is about twice Heyward's. That should make sense.

Base running runs 2015:
Carpenter +1.3 (-1 BBR)
Heyward +7.2 (+6 BBR)

Base running doesn't generate anywhere close to the value of hitting or defense, but Heyward was much better on the bases both by standard metrics (23 SB vs. 4 for Carpenter, 3 CS for each) and advanced (shown above). Advanced metrics show a 6-7 run advantage for Heyward, given he stole 19 more bases and was caught the same amount of times I'd say 6-7 runs seems about right. That reduces the "offense runs" advantage Carpenter had at the place from 14-15 runs down to 8-9 runs overall.

Note: Baseball reference also has a Rdp metric, runs saved by avoiding double plays (Heyward was +5, Carpenter was 0). Fangraphs includes this in their base running runs shown above, so I guess the models actually start to differ a bit here, by 5 runs or so. So BBR it was +11 for Heyward and -1 for Carpenter, a 12 run difference, not a six run difference (that's 0.6 WAR difference right there).

Offense runs 2015:
Carpenter +31.9
Heyward +22.2

This is simply batting runs + base running runs, and the gap, at fangraphs, has been reduced to 9.7 runs. At baseball reference the gap would be 14 - 7- 5 = 2 runs.

Now we get to the point that always drives differences and makes people question things, because it is really hard to quantify defense just watching as a fan.

Fielding runs 2015:
Carpenter -4.8 (-3 BBR)
Heyward +17.4 (+28 BBR)

These metrics are relative to your position. Both UZR and DRS thinks Carpenter was a little below average (5 runs, and 3 runs, respectively), and they both think Heyward was great defensively but DRS thinks he was really great (a whole 10 runs saved better than UZR does, 28 vs. 17 runs saved, respectively). 10 runs is a win, in terms of WAR, so this is a huge part of the gap between the two metrics.

Heyward saved 22 extra runs compared to Carpenter, using UZR, and 31 runs using DRS.

Positional adjustment 2015:
Carpenter +2 (+3 BBR)
Heyward -6 (-5 BBR)

In both models there is a positional adjustment that adds 8 runs (almost a whole win) to Carpenter, relative to Heyward. That cuts those defensive gaps from 22 and 31 to 14 and 23 runs. That's 1.4 to 2.3 wins in WAR.

As we saw above on offense Heyward had a 9.7 run deficit on fangraphs, and a 2 run deficit on baseball reference.

Offense + defense (including position adj) runs 2015:
Carpenter +31.9 - 2.8 = 29.1 RAA (runs above average)
Heyward +22.2 + 11.4 = 33.6 RAA

Using baseball reference you get:
Carpenter +27 + 0 (+3 - 3) = 27 RAA
Heyward +25 + 23 = 48 RAA

This is runs above average. They had similar playing time so to drop the baseline from average to replacement level 18-20 runs (Carpenter got one extra run due to more playing time, he hit leadoff). Then you divide by 10 to convert runs to wins.

So fangraphs thinks Heyward was 3.5 runs or so better than Carpenter, or 0.3 fWAR (5.6 fWAR for Heyward, 5.3 fWAR for Carpenter).

Baseball reference thinks the gap is bigger because of how it views defense (11 extra runs saved, 2 extra for Carpenter) and base running (5 extra runs for avoiding GIDP by Heyward), leading to a gap about 1.5 WAR larger than fangraphs (it had an extra run in the batting gap, and another in regular base running as well) so the gap is about 1.7 WAR larger, plus the 0.3, which is really 3.5 runs difference, on fangraphs and you get an expected gap of about 2 WAR between Heyward and Carpenter in 2015, which is what baseball reference shows . . . 7.0 vs. 4.8 bWAR.

This is really well done and explained. Kudos.

Couple of notes on baserunning to add...
Carpenter took an extra base 44% of the time and made 11 outs on the bases
Heyward took an extra base 57% of the time and made 8 outs on the bases

All of these little things add up when quantifying the difference in value.

For context on defense, just 13 outfielders of 72 that played at least 750 innings saved 10 runs. By Fielding Bible, Heyward was 3rd overall at 26. Kiermaier blew everyone away with 38, and Ender Inciarte was 2nd with 27.

Of the 25 3B that played at least 700 innings at the position, Carpenter was 16th at -2. There were 4 3B that saved at least 10 runs, with Arenado and Machado leading the way with 22 and 18, respectively.

* took 750 and 700 innings to get roughly an equal positional player sample size *

Again, appreciate your post, very thorough.

Again great dissection RBI
But even you narrate by saying “Fangraph thinks” “Baseball Ref thinks”…..these are opinions and most of their outcomes don’t align. So how can anyone look to this manipulated opinion analyses as reliable?

[renostl]

RBI, excellent breakdown of the equations.
Only thing is, baseball isn’t open to interpretation. An out is an out, a strike is a strike, a hit is a hit, a run is a run
This great disparity between the competing WAR houses completely invalidates WAR
Stats have Been recorded in the books since inception and those stats are what the player physically accomplished. Not what they might have or could have or a divider, multiplier, equator skewed to Offense or Defense.

I'm not sure how one could ever say that baseball isn't open to interpretation. Hits, errors, balls, strikes, out, safe... have all been an umpire's and an official scorer's interpretation since the beginning. All of those interpretations affect stats and outcomes.

The beauty of talking ball relies on it.

To Goldfans example, not all outs, hits, and runs are the same