WAR calculations

[traveller79]

I noticed in previous seasons that WAR totals changed between the season and offseason, sometimes dramatically. I chalked it up to adjustments in replacement level at season's end.

But last night I took note of my Mookie Betts card's values and checked it against today's totals. The 2029 WAR had adjusted as expected, but so had all previous seasons, which is not how WAR is calculated--each component of WAR is weighted for replacement level in that particular season, not all players in all times. Babe Ruth's 1927 WAR is not weighted against Mike Trout's performance in 2019 and would not shift according to the performance of players before or after 1927; the same should be true for Perfect Team seasons.

I hadn't noticed this happening in my OOTP seasons, so I'm assuming that Perfect Team includes every new season in its overall calculations of replacement level within a particular league level (discrete replacement levels for Perfect, Diamond, etc), which would have a natural deflationary effect on previous seasons' WAR calculations as more packs are opened and better cards pulled (with allowances for new players and teams).

[Matt Arnold]

Yeah, this has been discussed before. Basically, in a regular league, we keep the proper league history so we can look up exactly the correct values for previous seasons.

However in PT, we don't include that league history, so past seasons always just use generic league averages, so they will change. We try to use generic league averages, so that values don't change between seasons, but looks like there might still be a couple that get adjusted based on your current league. We'll keep looking and see if we can clean that up some more. Values will still change after a season, but hopefully we can prevent them changing even further after that.

[One Great Matrix]

Originally Posted by traveller79

I noticed in previous seasons that WAR totals changed between the season and offseason, sometimes dramatically. I chalked it up to adjustments in replacement level at season's end.

But last night I took note of my Mookie Betts card's values and checked it against today's totals. The 2029 WAR had adjusted as expected, but so had all previous seasons, which is not how WAR is calculated--each component of WAR is weighted for replacement level in that particular season, not all players in all times. Babe Ruth's 1927 WAR is not weighted against Mike Trout's performance in 2019 and would not shift according to the performance of players before or after 1927; the same should be true for Perfect Team seasons.

I hadn't noticed this happening in my OOTP seasons, so I'm assuming that Perfect Team includes every new season in its overall calculations of replacement level within a particular league level (discrete replacement levels for Perfect, Diamond, etc), which would have a natural deflationary effect on previous seasons' WAR calculations as more packs are opened and better cards pulled (with allowances for new players and teams).

Nothing you say is false about how WAR is calculated but for them to use the lastest MLB WAR v. 2.2? in the game wouldn't really be correct either...because it's Perfect Team generated statistics in Iron, Bronze, Silver, Diamond & Perfect Leagues. The W in WAR stands for Wins so I think anyway that the number they are showing is a better number for using some formula related to the sims than what they've done with historic stats in MLB... the number they show should give you an idea of how well the player performed better if it is related to the leagues you are simming in & not the way real-life historic WAR is calculated. The stat is a nice broad one but kind of kooky IMO anyway...

Anyway, WINS above replacement is likely to change if the A. replacement or B. competition changes.

[One Great Matrix]

You say Ruth's WAR is not weighted against Trout's WAR but that's exactly what the statistic is doing. IF Ruth were to play in the majors next year at the age of 26 he likely wouldn't put up the same numbers he did back when he WAS 26.

Plus, like I said that's what the stat is doing, it's weighing (even if by weighing each player against an imaginary replacement) how well the player did relative to them.

Therefore if this imaginary replacement remains the same while you move from Silver to Diamond Leagues then the WAR will change as well. That part is obvious maybe...

Lost my train of thought but hey, there are various ways to calculate a conceptual statistic. If the number was that accurate in regards to "wins vs. Player B" you could add up the WAR of a team and get say their wins above .500 or what-have-you, it would reflect a little more directly.

Even if they don't compare Trout to Ruth, I argue they do. With HR Trout has a number and Ruth has a number. With WAR they both have a number but it's a rating of their performance. So they are not weighted against one another but the idea is to compare their overall performance.

Kinda getting what I mean?

How many HR did he hit? 52. Trout? 37

What was his WAR? 20.6 The whole idea of knowing the next player's is you're comparing them. Nothing independent.

[bonalditaylor]

Why can't a 'snap shot' of the data from the last season just be saved upon league end and posted to that card. Why the need to adjust things? Im obviously not understanding why this is an issue and why the need for the recalcs. Just simplify it and use saved data, make it "read only" or some similar option and stop the recalc which seems to be the issue.

[traveller79]

Originally Posted by One Great Matrix

You say Ruth's WAR is not weighted against Trout's WAR but that's exactly what the statistic is doing. IF Ruth were to play in the majors next year at the age of 26 he likely wouldn't put up the same numbers he did back when he WAS 26.

Plus, like I said that's what the stat is doing, it's weighing (even if by weighing each player against an imaginary replacement) how well the player did relative to them.

Therefore if this imaginary replacement remains the same while you move from Silver to Diamond Leagues then the WAR will change as well. That part is obvious maybe...

Lost my train of thought but hey, there are various ways to calculate a conceptual statistic. If the number was that accurate in regards to "wins vs. Player B" you could add up the WAR of a team and get say their wins above .500 or what-have-you, it would reflect a little more directly.

Even if they don't compare Trout to Ruth, I argue they do. With HR Trout has a number and Ruth has a number. With WAR they both have a number but it's a rating of their performance. So they are not weighted against one another but the idea is to compare their overall performance.

Kinda getting what I mean?

How many HR did he hit? 52. Trout? 37

What was his WAR? 20.6 The whole idea of knowing the next player's is you're comparing them. Nothing independent.

Like Matt said, they're using generic league averages rather than set-in-stone, (simulated) historical averages to calculate replacement level, which is why you see the fluctuations in Betts' WAR in past seasons. I see why they did do it with teams floating between leagues and levels, but I didn't find anything on it when I searched the forums.

But here in the real world, WAR--WAR never changes.

[One Great Matrix]

Man, I don't know if you're talking about War or the stat WAR at this point, lol...

But yes it does change...you are comparing it to players who (at first would give you a .320 winning %), and now would give you a .294 winning %.

Can you figure out how even if it remains relative to a player that would give you a .294 winning % in the MLB a player's WAR will change?

Actually, the statistic is an ever-changing one...

That Babe Ruth his 61 home runs in a season is a constant.
His high WAR just means he was many wins better than the back-up plan (player to replace Ruth) that the Yankees did not have in place in the case Ruth was unavailable.
The only way I can imagine you'd keep that WAR number constant is by constantly adjusting other numbers.

WAR is an approximation of wins added for a player's performance...
There are pretty obvious reasons why it wouldn't, couldn't be an exact number. And it's kind of for the same reason that it isn't an exact number that it's bound not to be a constant one.

[One Great Matrix]

Although...I do understand what you're trying to say.

I think the 2 posts I've sort of pointed out that WAR is a formula have been rather enlightening for me on how it is drastically different from most baseball statistics.

There may be an inherent problem with it, I don't know.

As much as I can appreciate what you said about the way in which Betts' WAR was adjusted, I noted something like that when the statistic was introduced AND BECAME WILDLY POPULAR...that's where I had some problems myself ...? Right off the bat. Wasn't something I really put into words but complaining about inconsistent WAR values is one example I guess of why I was disappointed with the way the stat became so popular.

Taking the time, I see what you mean but I'm not going to try and say it isn't a real stat. But it is a sort of rating, a concept, not a number that complements baseball, more like a stat concerning stats.

[One Great Matrix]

Although...I do understand what you're trying to say.

I think the 2 posts I've sort of pointed out that WAR is a formula have been rather enlightening for me on how it is drastically different from most baseball statistics.

There may be an inherent problem with it, I don't know.

As much as I can appreciate what you said about the way in which Betts' WAR was adjusted, I noted something like that when the statistic was introduced AND BECAME WILDLY POPULAR...that's where I had some problems myself ...? Right off the bat. Wasn't something I really put into words but complaining about inconsistent WAR values is one example I guess of why I was disappointed with the way the stat became so popular.

Taking the time, I see what you mean but I'm not going to try and say it isn't a real stat. But it is a sort of rating, a concept, not a number that complements baseball, more like a stat concerning stats.

[traveller79]

Originally Posted by One Great Matrix

Although...I do understand what you're trying to say.

I think the 2 posts I've sort of pointed out that WAR is a formula have been rather enlightening for me on how it is drastically different from most baseball statistics.

There may be an inherent problem with it, I don't know.

As much as I can appreciate what you said about the way in which Betts' WAR was adjusted, I noted something like that when the statistic was introduced AND BECAME WILDLY POPULAR...that's where I had some problems myself ...? Right off the bat. Wasn't something I really put into words but complaining about inconsistent WAR values is one example I guess of why I was disappointed with the way the stat became so popular.

Taking the time, I see what you mean but I'm not going to try and say it isn't a real stat. But it is a sort of rating, a concept, not a number that complements baseball, more like a stat concerning stats.

The jargon I learned in the 1990s to describe newer standards like WAR was "metric," meaning a measurement norm. And while WAR got to be a Metric, Batting Average and OBP remained "stats," maybe because they were easy for kids to calculate on the back of their scorecards. Sometimes I think the math nerds just wanted to clear the field of us amateurs.

The inputs into WAR may not be easy to understand immediately if we're used to traditional assessment norms, especially for those of us old enough to remember when we got our stats from the Sporting News and the backs of baseball cards, but WAR is a useful (if not the best) reference once you parse everything that goes into it. Plus the process of developing the new metrics when everything was shaking out was interesting just from a politics and personalities perspective.

[One Great Matrix]

One way of looking at it is that instead of comparing, say, Betts to a .300 hitter or a .200 hitter (+.015 or -.020 or something), OR a hitter with 10 extra-base hits or 50 extra-base hits (+27 or -2 etc.)...

Is they are comparing Betts to a player who hits .200 with 10 extra-base hits vs. a player who hits .300 with 50 extra base hits in the same number of PA.
If people find that useful, OK but in terms of say transportation, it'd be like instead of comparing:

10 MPH to 55 MPH to 700 MPH
OR
17 MPG to 32 MPG to 130 MPG

Attributing a number to a car that's based on comparing (Vehicle 1) 10 MPH & 17 MPG to (Vehicle 2) 700 MPH & 130 MPG. Then the car gets a number not just based on how much better it is than 10 MPH & 17 MPG but how it compares to (the 10 MPH/17MPG vehicle vs. the 700 MPH/130 MPG vehicle)...

If that doesn't make sense, that's what doesn't make sense to me about the metric WAR. Besides anything about the math being complex.

Thanks.

[One Great Matrix]

I guess one flaw (besides the mis-understandings) the complexity of the number creates...

is that it may overlook a player being weak in one or more area but particularly outstanding in one (or 2 or 3)... Who couldn't use 35+ HR or a .420 OBP out of their clean-up hitter or lead-off hitter respectively even if the entire spectrum of their stats is not as good as the next player's.

Transportation vehicles...700+ MPH or outstanding efficiency in MPG even if the gas is a real downer on the fast car or the efficient one doesn't fly. Ha...

[One Great Matrix]

Another is that it's based on stats, stats, stats. There is nothing really wrong with this but if baseball is what we make it, and it seems to me it is outside the basic foundation of its rules and bats, balls and gloves (from there it's up to us to enjoy the game)...well, you keep measuring players' performances with this one formula, and baseball becomes about players performances by a single number. A player could win you a game with a extra-base hit in the bottom of the 9th, & go overlooked/released a week later because he hit 2 less extra-base hits total than another guy on the team that week.

I don't have a big problem with the stat or anything so I'll stop writing against it now unless you know I come up with something I MUST say regarding this.