SoxProspects News
|
|
|
|
Legal
Forum Ground Rules
The views expressed by the members of this Forum do not necessarily reflect the views of SoxProspects, LLC.
© 2003-2024 SoxProspects, LLC
|
|
|
|
|
Forum Home | Search | My Profile | Messages | Members | Help |
Welcome Guest. Please Login or Register.
WAR and More (...what is it good for)
|
Post by telson13 on Aug 1, 2018 13:30:15 GMT -5
Harper's underlying skills and rate stats and long-term production record mean that I would rather have him than Span the rest of the season. But are you asking, like retroactively? Who would I rather have had from April through July/ Obviously if an Aaron Hicks type of player out-produces a superior Bryce Harper for a period of time, I'd rather have had the superior production over that past period of time. But...so what? Worse players having periods of time where they outperform better ones isn't some newfangled statistical revolution stuff. That's gone on since the start of baseball and was reflected in traditional stats. I remember going to a Sox game when I was pretty young and they flashed player stats on the board. I thought Ed Romero was a god for batting .500, until my dad explained SSS to me.
|
|
nomar
Veteran
Posts: 10,793
Member is Online
|
Post by nomar on Aug 1, 2018 14:16:45 GMT -5
Which Dunkin swirl has the highest WAR? I’m out of coffee at home and will have to buy coffee tomorrow morning.
|
|
|
Post by James Dunne on Aug 1, 2018 14:19:39 GMT -5
Mocha, everyone knows that.
|
|
|
Post by fenwaythehardway on Aug 1, 2018 14:41:59 GMT -5
Still less wonky than judging a player based on HRs, RBI, BA and runs scored. I've never quite understood how a mathematical formula with a dozen random variables and constants is less wonky than watching the game and reporting what happened. I like baseball statistics ( I started a sports analytics group in grad school) and I think they play their part, but they get applied in ways that frustrate me sometimes. Simply adding WAR up and using that as the sole source of comparison is one of them. Take a more topical example, Cain and Martinez were free agents last year, here's their recent WAR stats. Cain Martinez WAR (2015) 7.2 5.0 WAR (2016) 3.0 1.7 WAR (2017 5.3 4.2 Which one would you rather have? More importantly, who got paid more? Front offices know what these guys are worth beyond a single summary number and they have more analytic tools at their disposal to determine who should get paid. Also, Cain signed first and I didn't hear one person anywhere say "OK, Cain got 5/$80 so that's the realistic ceiling for Martinez." Everyone knows Martinez is more valuable than Cain, even if you can't always write a number down to prove it. I think WAR is great, but flawed and the absolute certainty that tends to be associated with it is frustrating. It's a data point and generally a good one, but it's not the only data point. I'm sorry but people who want to poke holes in WAR are going to have to do a lot better than "well it just doesn't SEEM right".
|
|
|
Post by texs31 on Aug 1, 2018 15:19:26 GMT -5
RBI, Pitcher Wins . . .
If the same behavior(performance) can result in different outcomes then the outcome is NOT a legitimate measure of the performance. It's really that simple.
|
|
|
Post by malynn19 on Aug 1, 2018 15:34:48 GMT -5
I mean, knocking down the strawman that WAR is infallible and the only stat you should look at is another tired talking point that is so ten years ago. Literally no one thinks WAR is "all there is to know, that it is absolute, and there will be nothing added in the future." You want to criticize defensive metrics next? What about FIP versus RA for pitcher WAR? C'mon now. Thanks for the explanation your Highness.
|
|
|
Post by umassgrad2005 on Aug 1, 2018 15:52:04 GMT -5
I will say you can question war, its not perfect. Last year I called out Moncada bwar, especially his defensive value. It seemed way too high. This year it has gone down to a level that makes sense. So you can certainly have one year outliers, especially in regards to defensive value. Thing is a guy like Cain has a very long track record of being great.
Advanced stats aren't perfect, but they adjust for things like parks you play in. You want to judge a guys offensive numbers, you better be using OPS+ that adjusts for park effects. Otherwise the numbers are kinda useless. Just look at a guy like Beltre and his homers going from Dodgers, to Mariners, to Red Sox to Rangers. Now tell me park effect doesn't matter. If you have watched Manny and Hanley play the OF and are now watching Betts tell me D doesn't make a huge difference.
|
|
|
Post by dcb26 on Aug 1, 2018 16:06:04 GMT -5
But in these instances the inferior players didn’t out perform: Harper has better numbers this season than Span, Arenado than Cain, Martinez than Chapman... but that is not reflected in WAR. As I wrote above, I’ll leave it at that. We are likely talk across each other, and I don’t want to gum up the board. en.wikipedia.org/wiki/Begging_the_questionThis is such an early 2000s discussion. At this point, I just don't have the patience to explain how WAR works anymore. Wasn't there just a big thing ripping posts that contribute nothing on this site? As it relates to the WAR topic, I think this is actually a worthwhile discussion; we all know there are at least potential flaws with WAR, but it's become almost the default stat for discussing players, especially on this site. It's pretty fair to ask if that makes sense, and if it does, to look for more info as to why. I mean, we accept that the Sox front office knows more than any of us about their players yet we are still OK questioning their moves (as we should be) - I don't see why we can't question WAR or other stats just because they're more accurate than what we had in the past. Not every topic has to immediately become a referendum on "old school vs. new school stats" where everyone posts along their party lines. I know the active moderators and site staff tend to have a similar philosophy on this topic, and that's fine, but I don't think that's grounds for belittling a topic that questions it - this was a lot more nuanced than the old "RBI's make sense so they're better than WAR" (I think.) Anyway, I'm just a lurker so I'll go back to lurking, but just throwing it out there that 1) a discussion about how useful WAR really is could be worth having/revisiting (personally I'd love to see an ongoing discussion about changes/trends/advancements in baseball stats) and 2) as it pertains to this site, it would be nice if more people could participate in those types of discussions.
|
|
|
Post by jimed14 on Aug 1, 2018 16:13:21 GMT -5
I'm sorry but people who want to poke holes in WAR are going to have to do a lot better than "well it just doesn't SEEM right". If you want a specific example of WAR breakdown, here are a couple: - Lorenzo Cain has a dWAR (looking at baseball reference numbers) above 2 in four of the last six years. By their definition a mlb starter is anyone with a WAR above 2, so by that logic Cain could have a batting average of .000 and still be a legitimate, if not above average, major league starter. Literally 0-5 with 5 strikeouts every. single. night. and the math says he's a starter. We all know that's not true. If Cain had a batting average of .000, his defensive value would be more than wiped out and he'd have the worst negative WAR total in history if he played an entire season.
|
|
|
Post by jimed14 on Aug 1, 2018 16:15:27 GMT -5
I will say you can question war, its not perfect. Last year I called out Moncada bwar, especially his defensive value. It seemed way too high. This year it has gone down to a level that makes sense. So you can certainly have one year outliers, especially in regards to defensive value. Thing is a guy like Cain has a very long track record of being great. Advanced stats aren't perfect, but they adjust for things like parks you play in. You want to judge a guys offensive numbers, you better be using OPS+ that adjusts for park effects. Otherwise the numbers are kinda useless. Just look at a guy like Beltre and his homers going from Dodgers, to Mariners, to Red Sox to Rangers. Now tell me park effect doesn't matter. If you have watched Manny and Hanley play the OF and are now watching Betts tell me D doesn't make a huge difference. Absolutely, I agree with all of what you just said. Like I said earlier, I like statistics and some of the new stuff is great. Not suggesting we throw it all away, just saying that sometimes there are some holes and a broader discussion paints a more accurate picture. You should really take the time to learn what it is before you start criticizing it.
|
|
|
Post by jimed14 on Aug 1, 2018 16:24:11 GMT -5
If Cain had a batting average of .000, his defensive value would be more than wiped out and he'd have the worst negative WAR total in history if he played an entire season. Yeah, complete brain spasm on my end. Forgot that WAR goes negative. What I meant was if his offensive WAR was 0.0 is he still a mlb ballplayer? Yes. Why don't you consider preventing runs better than the average player to be valuable? Why don't you consider a superior base runner, stealing bases, taking extra bases and not getting thrown out better than the average player to be valuable?
|
|
|
Post by ajs1994 on Aug 1, 2018 16:30:04 GMT -5
If Cain had a batting average of .000, his defensive value would be more than wiped out and he'd have the worst negative WAR total in history if he played an entire season. Yeah, complete brain spasm on my end. Forgot that WAR goes negative. What I meant was if his offensive WAR was 0.0 is he still a mlb starter? Guys like you described are JBJ this year, Jose Iglesias, Billy Hamilton. They all hold onto jobs and retain value due to their baserunning and defense outweighing their replacement level production as hitters.
|
|
|
Post by umassgrad2005 on Aug 1, 2018 16:38:14 GMT -5
I will say you can question war, its not perfect. Last year I called out Moncada bwar, especially his defensive value. It seemed way too high. This year it has gone down to a level that makes sense. So you can certainly have one year outliers, especially in regards to defensive value. Thing is a guy like Cain has a very long track record of being great. Advanced stats aren't perfect, but they adjust for things like parks you play in. You want to judge a guys offensive numbers, you better be using OPS+ that adjusts for park effects. Otherwise the numbers are kinda useless. Just look at a guy like Beltre and his homers going from Dodgers, to Mariners, to Red Sox to Rangers. Now tell me park effect doesn't matter. If you have watched Manny and Hanley play the OF and are now watching Betts tell me D doesn't make a huge difference. Absolutely, I agree with all of what you just said. Like I said earlier, I like statistics and some of the new stuff is great. Not suggesting we throw it all away, just saying that sometimes there are some holes and a broader discussion paints a more accurate picture. For me I just looked at Moncada's fielding percentage, errors and range factor. Then compared that to gold glove 2b like Pedroia and Kinsler. It made no sense that Moncada's defensive value was on pace or better than those guys GG seasons, when he was horrible in fielding percentage way below league average, he was on pace to be near the league leaders in errors and his range factors was just slightly above league average. Pedroia and Kinsler were good across the board, it made zero sense. Moncada would have had to have a range factor through the roof to make up for the fielding percentage and errors and he didn't. That's were combaring bwar to fwar comes in, because fwar didn't value him as a gold glove defender, just slightly above average. That's why its not a bad idea to use both. Most times they are close, but sometimes they aren't.
|
|
|
Post by jmei on Aug 1, 2018 16:42:09 GMT -5
Wasn't there just a big thing ripping posts that contribute nothing on this site? As it relates to the WAR topic, I think this is actually a worthwhile discussion; we all know there are at least potential flaws with WAR, but it's become almost the default stat for discussing players, especially on this site. It's pretty fair to ask if that makes sense, and if it does, to look for more info as to why. I mean, we accept that the Sox front office knows more than any of us about their players yet we are still OK questioning their moves (as we should be) - I don't see why we can't question WAR or other stats just because they're more accurate than what we had in the past. Not every topic has to immediately become a referendum on "old school vs. new school stats" where everyone posts along their party lines. I know the active moderators and site staff tend to have a similar philosophy on this topic, and that's fine, but I don't think that's grounds for belittling a topic that questions it - this was a lot more nuanced than the old "RBI's make sense so they're better than WAR" (I think.) Anyway, I'm just a lurker so I'll go back to lurking, but just throwing it out there that 1) a discussion about how useful WAR really is could be worth having/revisiting (personally I'd love to see an ongoing discussion about changes/trends/advancements in baseball stats) and 2) as it pertains to this site, it would be nice if more people could participate in those types of discussions. Fair enough. I don't think it's a worthwhile discussion, but I suppose most people haven't been an internet baseball fan as long as I have. No one thinks WAR is perfect. But for those situations where, out of necessity, you need one stat that captures a player's context-neutral historical value, WAR is almost always the best option. If anyone thinks that there is a better option to fit that niche, I'm all ears.
|
|
|
Post by jimed14 on Aug 1, 2018 16:44:47 GMT -5
The problem with WAR is mostly due to inefficiencies in calculating dWAR. Team analytical departments likely have much more advanced defensive metrics than UZR and DRS, neither of which adjust for shifting.
Example: JDM last year went from one of the worst outfielders in baseball (Detroit was last in the league in shifting) to slightly above average in Arizona (they were close to the top of the league in shifting).
That is not evidence to say WAR is a terrible judge of talent. It's the best we've got. It is evidence that it needs to be improved.
How logical is it to ignore defense and baserunning when determining a player's value? Because that's basically the argument against WAR.
|
|
|
Post by manfred on Aug 1, 2018 17:04:21 GMT -5
Wasn't there just a big thing ripping posts that contribute nothing on this site? As it relates to the WAR topic, I think this is actually a worthwhile discussion; we all know there are at least potential flaws with WAR, but it's become almost the default stat for discussing players, especially on this site. It's pretty fair to ask if that makes sense, and if it does, to look for more info as to why. I mean, we accept that the Sox front office knows more than any of us about their players yet we are still OK questioning their moves (as we should be) - I don't see why we can't question WAR or other stats just because they're more accurate than what we had in the past. Not every topic has to immediately become a referendum on "old school vs. new school stats" where everyone posts along their party lines. I know the active moderators and site staff tend to have a similar philosophy on this topic, and that's fine, but I don't think that's grounds for belittling a topic that questions it - this was a lot more nuanced than the old "RBI's make sense so they're better than WAR" (I think.) Anyway, I'm just a lurker so I'll go back to lurking, but just throwing it out there that 1) a discussion about how useful WAR really is could be worth having/revisiting (personally I'd love to see an ongoing discussion about changes/trends/advancements in baseball stats) and 2) as it pertains to this site, it would be nice if more people could participate in those types of discussions. Fair enough. I don't think it's a worthwhile discussion, but I suppose most people haven't been an internet baseball fan as long as I have. No one thinks WAR is perfect. But for those situations where, out of necessity, you need one stat that captures a player's context-neutral historical value, WAR is almost always the best option. If anyone thinks that there is a better option to fit that niche, I'm all ears.It isn’t context neutral in at least one essential way: position. If one says that Cain is worth what he is worth relative to CFers, and JDM is worth what he is worth relative to DHs, but then uses WAR as a single number to evaluate position players across the board, it is flawed. As I wrote before, it makes more sense to have separate categories. Because as it stands, a decent player at a weak position can appear more valuable than a great player at a stronger position.
|
|
|
Post by Oregon Norm on Aug 1, 2018 17:18:19 GMT -5
Wins above replacement is really a measure of the economic value of a player. Like many economic metrics it isn't perfect. GDP is one example. Two countries may have the same GDP but the one with fewer people has a higher GDP per capita. Same with WAR. Two players with the same amount of wins above replacement should have the number of games they've played be part of the equation. Defense is the real issue with day to day game metrics as jmei pointed out. It's just much more difficult to properly value that.
The problem with the eyeball test is that there's not enough eyeballs to cover a season's worth of games for a given player, and we forget stuff anyway because we're human. The inability to properly evaluate the worth of a player has kept a lot of great talent from being properly recognized. WAR was a real step forward in doing that, a somewhat objective measure of how good players really are and were.
I still use my intuition watching guys play but I'll admit it fails me sometimes.
|
|
|
Post by manfred on Aug 1, 2018 17:23:16 GMT -5
The problem with WAR is mostly due to inefficiencies in calculating dWAR. Team analytical departments likely have much more advanced defensive metrics than UZR and DRS, neither of which adjust for shifting. Example: JDM last year went from one of the worst outfielders in baseball (Detroit was last in the league in shifting) to slightly above average in Arizona (they were close to the top of the league in shifting). That is not evidence to say WAR is a terrible judge of talent. It's the best we've got. It is evidence that it needs to be improved. How logical is it to ignore defense and baserunning when determining a player's value? Because that's basically the argument against WAR. I think this is partly true: there are players whose value is heavily defensive in a way that seems potentially overblown. I think defense is obviously of great value... but only to a point. I find it hard to imagine that any defender is the equivalent in run prevention (over a different, competent defender) of a 80+ RBI guy. I’m a big baserunning guy — I think SBs should be factored into OPS somehow (a big basestealer may have a low slugging but end up on second more than others hit doubles, for example). And running into outs is very frustrating. But, again, over 162 games, I suspect most players fall somewhere in the no significant net gain or loss range on the bases. I guess I look at Harper again... 3rd in home runs, 7th in runs, 11th in RBI... and I wonder if his productivity is really barely more than the average guy in his position? (And this with 8 steals and average defense). What do I learn when I see that Zach Greinke has a higher player WAR than Bryce Harper this year (just batting! .7 to .6)?
|
|
|
Post by jimed14 on Aug 1, 2018 18:16:18 GMT -5
Doing a little research, as requested. Anyone ever hear of this guy, Bill James? (My italics aren't working, assume an appropriate amount of playful sarcasm) He has an interesting take on WAR and the 2017 MVP race: “Aaron Judge was nowhere near as valuable as Jose Altuve. Why? Because he didn’t do nearly as much to win games for his team as Altuve did. It is NOT close. The belief that it is close is fueled by bad statistical analysis — not as bad as the 1974 statistical analysis, I grant, but flawed nonetheless. It is based essentially on a misleading statistic, which is WAR. Baseball-Reference WAR shows the little guy at 8.3, and the big guy at 8.1.” Sort of what I was getting at, but he's smarter and more articulate than I am so it's probably better to hear it from him. Interesting article, goes over some pros and cons: www.sportingnews.com/us/mlb/news/wins-above-replacement-war-what-is-it-how-to-calculate-baseball-reference-fangraphs-sabermetrics-analytics/1emmivy1vroy01owwnc9zj6mvrFocus on the bold, because that's where you're at. Bill James is looking forwards, not backwards.
|
|
|
Post by Oregon Norm on Aug 1, 2018 18:51:29 GMT -5
The best measure, for my money, is win probability added and it does exactly what you want it to do. You can find it at Baseball Reference under a player's advanced stats. In a nutshell, it does that by determining from historical data - literally every game that's ever been played and for which play by play data is available - what an at bat is worth for a given situation.
What's a situation? A given base state, score, and inning. It does not take into account the competition, but that's about the only thing missing. That's what data mining has allowed analysts to do. By that measure, Altuve was better than Judge.
Your job, should you choose to accept it, is to convince the MVP voters that it should be used. Given that at least a few of them still think AVG is a better stat than OBP, that might be a tough but to crack.
|
|
|
Post by fenwaythehardway on Aug 1, 2018 20:39:35 GMT -5
Anyway, Lorenzo Cain is having an excellent season.
|
|
|
Post by manfred on Aug 1, 2018 20:50:15 GMT -5
Another specific question — Boston related: why is Andrew Benintendi a negative defensive WAR? Is he really a below average defender? I am not asking to continue the battle — I don’t get how that can be so.
|
|
|
Post by fenwaythehardway on Aug 1, 2018 21:08:09 GMT -5
|
|
|
Post by manfred on Aug 1, 2018 21:16:34 GMT -5
This really surprises me. He seems pretty good to me. Obviously the third -best on the Sox, but that is the tough side of playing with JBJ and Mookie... you’ll always be the worst outfielder!
|
|
|
Post by philsbosoxfan on Aug 1, 2018 21:29:50 GMT -5
Another specific question — Boston related: why is Andrew Benintendi a negative defensive WAR? Is he really a below average defender? I am not asking to continue the battle — I don’t get how that can be so. You'll drive yourself crazy looking at specific examples. For example, I can't see much discernible difference between Mookie Betts and Mike Trout offensively. However, Betts is at 5.5 oWAR and Trout is 7.3 (per baseball reference). Apparently, 23 at bats and 20 OPS points is worth 1.8 oWAR. Because I learned something today (WPA) I'm going to point out that Betts is 4.0 WPA (leads the league) and Trout is 3.5 WPA. Still learning what that means, but thought it was interesting. Betts plays RF, Trout plays CF. That's a major factor.
|
|
|