+- Sons of Ivy (https://sonsofivy.com/forum)
+-- Forum: Chicago Cubs (https://sonsofivy.com/forum/forum-4.html)
+--- Forum: The Friendly Confines (https://sonsofivy.com/forum/forum-8.html)
+--- Thread: WAR, what is it good for? (/thread-5701.html)
WAR, what is it good for? - leonardsipes - 02-21-2010
Not much.
One of the reasons people say Hendry is a sub par GM, is because he does not understand modern statistical analysis. For example, he will sign replacement level players to multi million dollar contracts. I think the weakness is in the modern stats.
Formulas for WAR (wins above replacement) or the like, use wOBP or OPS. These stats are most accurate for everyday players who hit in the middle of the lineup. They are not going to be that good at evaluating part time players who hit at the bottom of the lineup (borderline replacement level players).
WAR is also fine for 1bman and OFers. I don't think it is accurate in reflecting the difficulty of finding a 2b,3b,ss, or C that can both hit and field at a minimum level. The bullpen is a crap shoot, and often a minor lea. will come up and do fine, but a team can not count on the MLs to fill the pen and especially for rotation spots.
Even if you take WAR at face value, the formulas figure a replacement level player will cost a team 1 to 3 wins. A GM with a big payroll, need to not only sign and retain star players, but spend the extra million here and there to make sure replacement level players or worse, are not getting on the field.
WAR, what is it good for? - Ace - 02-21-2010
Fine points, I have one gripe:
Nothing wrong with using OPS for any player, as long as you're using it to make the right kind of comparisons. The fact that using OPS is useless to compare Albert Pujols and Andres Blanco means only that the comparison is worthless, not the stat.
WAR, what is it good for? - leonardsipes - 02-21-2010
OPS's value is that on average, it measures production about as well most more complicated stats, and better than many traditional stats. The key word is on average. So it is going to work great on some players, adequate on others and suck at evaluating some. It is also a weighted average. So, working well for one player who gets 600 PAs makes up for a poor job on 4 guys with 150 PAs. It is going to work best on the best players, who hit in the middle of the order and who need to have both SLG and OBP. It is not as good for players at the top of the line up, where OBP is the most important, or the bottom, where OBP is less important.
Most of the value from an elite hitter, is going to be from OBP or SLG. As you move down the scale and guys make more outs and have less productive hits, a larger part of their value is going to be measured in other factors like SB, bunting, clutch hitting and productive outs.
Most of the statistical analysis validating OPS, is going to be done on players worth evaluating. That is why it appears to work so well - and it does at evaluating middle of the order hitters. For evaluating Andres Blanco, it is no less a shot in the dark than using RBI.
WAR, what is it good for? - kbwsb - 02-21-2010
I wish there was a catch-all stat that we could use to rank players with, but one doesn't exist.
Even OPS+, which is a great stat, because it incorporates ballpark factors, measures players against their league counterparts, etc., isn't ideal.
For one, it only measures hitting, not fielding, baserunning, clubhouse personality, or anything else. Also, it's a doesn't give credit for counting stats; thus, a player who plays all 162 games and puts up a 135 OPS+ is considered the same as a part-timer who gets a 135 OPS+ in only 60 games. That's bull.
And none of this even mentions the obvious fact that OPS measures OBP and slugging as being equal, when it's well-established that OBP is far more valuable.
WAR at least includes defense. It also gives credit for defensive positions played; in other words, it's obvious that a SS is more valuable (and harder to find) than a left fielder, so a player is given credit for being able to handle MLB SS.
I don't know if it's the best stat out there, but it is the most comprehensive.
WAR, what is it good for? - Destined - 02-21-2010
Lowrider and Cisco Kid were decent hits.
WAR, what is it good for? - jstraw - 02-21-2010
<!--quoteo(post=80249:date=Feb 21 2010, 09:31 PM:name=Destined)-->QUOTE (Destined @ Feb 21 2010, 09:31 PM) <{POST_SNAPBACK}><!--quotec-->Lowrider and Cisco Kid were decent hits.<!--QuoteEnd--><!--QuoteEEnd-->
Yep. Slippin' Into Darkness was cool too.
WAR, what is it good for? - dk123 - 02-21-2010
Just realized what a clever thread title this was...
WAR, what is it good for? - Fella - 02-22-2010
<!--quoteo(post=80236:date=Feb 21 2010, 07:22 PM:name=KBwsb)-->QUOTE (KBwsb @ Feb 21 2010, 07:22 PM) <{POST_SNAPBACK}><!--quotec-->OPS+, which is a great stat, because it incorporates ballpark factors,<!--QuoteEnd--><!--QuoteEEnd-->
This is actually one of my biggest gripes about OPS+, not that there shouldn't be some adjustment based on the ballpark you play in, but I think they go WAY overboard in the OPS+ formula. You really think an .870 OPS in Boston is the same value as an .815 OPS in Seattle just because a player plays half his games in that ballpark? Seems pretty overboard to me.
WAR, what is it good for? - leonardsipes - 02-23-2010
<!--quoteo(post=80236:date=Feb 21 2010, 08:22 PM:name=KBwsb)-->QUOTE (KBwsb @ Feb 21 2010, 08:22 PM) <{POST_SNAPBACK}><!--quotec-->I wish there was a catch-all stat that we could use to rank players with, but one doesn't exist.<!--QuoteEnd--><!--QuoteEEnd-->
One can't exist. Different types of players are valuable in different ways. It would be impossible to have a stat that is good at evaluating one type of player also be good at evaluating others. A lot of evaluating energy is spent on the best players, and therefore, most of the catch-all stats work best at evaluating better hitters.
OPS is great at comparing players with high OPSs. Players with lower OPSs are not in baseball because too many GMs don't understand modern stats, but because their value is in other aspects of their game. I am not trying to say, other aspects of the starting 1bman's game can make up for a low OPS. If you could pick between 2 starting 1bman, one with a .900 OPS and another 10% better, you don't need any more information to know one of them is an MVP candidate.
If I tell you to pick between a SS with a .650 OPS or one with a .715 OPS, you will need more information. There are a lot of factors that can make the guy with the .650 OPS a lot more valuable
WAR, what is it good for? - kbwsb - 02-25-2010
<!--quoteo(post=80257:date=Feb 21 2010, 10:06 PM:name=Fella)-->QUOTE (Fella @ Feb 21 2010, 10:06 PM) <{POST_SNAPBACK}><!--quotec--><!--quoteo(post=80236:date=Feb 21 2010, 07:22 PM:name=KBwsb)--><div class='quotetop'>QUOTE (KBwsb @ Feb 21 2010, 07:22 PM) <{POST_SNAPBACK}><!--quotec-->OPS+, which is a great stat, because it incorporates ballpark factors,<!--QuoteEnd--><!--QuoteEEnd-->
This is actually one of my biggest gripes about OPS+, not that there shouldn't be some adjustment based on the ballpark you play in, but I think they go WAY overboard in the OPS+ formula. You really think an .870 OPS in Boston is the same value as an .815 OPS in Seattle just because a player plays half his games in that ballpark? Seems pretty overboard to me.
<!--QuoteEnd--></div><!--QuoteEEnd-->
I'm actually curious how they determine the actual numbers. But I do know that Fenway Park has been the best hitter's park in the majors for a long time, and by a <i>mile</i>.
Just one example of a thousand, but during Wade Boggs's prime, he routinely hit about .400 in Fenway, and about .300 on the road. When ranking a player, to not adjust for that kind of thing is bad front office work.
There are some parks, like Petco Park in S.D., that take away so many homers that it's simply not fair to compare Adrian Gonzo to a player in Philly or Cinci, who play half their games in home-run bandboxes.
WAR, what is it good for? - Coldneck - 02-25-2010
<!--quoteo(post=80634:date=Feb 25 2010, 12:43 PM:name=KBwsb)-->QUOTE (KBwsb @ Feb 25 2010, 12:43 PM) <{POST_SNAPBACK}><!--quotec--><!--quoteo(post=80257:date=Feb 21 2010, 10:06 PM:name=Fella)--><div class='quotetop'>QUOTE (Fella @ Feb 21 2010, 10:06 PM) <{POST_SNAPBACK}><!--quotec--><!--quoteo(post=80236:date=Feb 21 2010, 07:22 PM:name=KBwsb)--><div class='quotetop'>QUOTE (KBwsb @ Feb 21 2010, 07:22 PM) <{POST_SNAPBACK}><!--quotec-->OPS+, which is a great stat, because it incorporates ballpark factors,<!--QuoteEnd--><!--QuoteEEnd-->
This is actually one of my biggest gripes about OPS+, not that there shouldn't be some adjustment based on the ballpark you play in, but I think they go WAY overboard in the OPS+ formula. You really think an .870 OPS in Boston is the same value as an .815 OPS in Seattle just because a player plays half his games in that ballpark? Seems pretty overboard to me.
<!--QuoteEnd--></div><!--QuoteEEnd-->
I'm actually curious how they determine the actual numbers. But I do know that Fenway Park has been the best hitter's park in the majors for a long time, and by a <i>mile</i>.
Just one example of a thousand, but during Wade Boggs's prime, he routinely hit about .400 in Fenway, and about .300 on the road. When ranking a player, to not adjust for that kind of thing is bad front office work.
There are some parks, like Petco Park in S.D., that take away so many homers that it's simply not fair to compare Adrian Gonzo to a player in Philly or Cinci, who play half their games in home-run bandboxes.
<!--QuoteEnd--></div><!--QuoteEEnd-->
Your point is exactly correct Kb. Just one clarification. I used to think that Philly was a band box too. But after looking at the numbers it appears that Philly is nuetral when it comes to HRs.
WAR, what is it good for? - Fella - 02-25-2010
<!--quoteo(post=80634:date=Feb 25 2010, 10:43 AM:name=KBwsb)-->QUOTE (KBwsb @ Feb 25 2010, 10:43 AM) <{POST_SNAPBACK}><!--quotec--><!--quoteo(post=80257:date=Feb 21 2010, 10:06 PM:name=Fella)--><div class='quotetop'>QUOTE (Fella @ Feb 21 2010, 10:06 PM) <{POST_SNAPBACK}><!--quotec--><!--quoteo(post=80236:date=Feb 21 2010, 07:22 PM:name=KBwsb)--><div class='quotetop'>QUOTE (KBwsb @ Feb 21 2010, 07:22 PM) <{POST_SNAPBACK}><!--quotec-->OPS+, which is a great stat, because it incorporates ballpark factors,<!--QuoteEnd--><!--QuoteEEnd-->
This is actually one of my biggest gripes about OPS+, not that there shouldn't be some adjustment based on the ballpark you play in, but I think they go WAY overboard in the OPS+ formula. You really think an .870 OPS in Boston is the same value as an .815 OPS in Seattle just because a player plays half his games in that ballpark? Seems pretty overboard to me.
<!--QuoteEnd--></div><!--QuoteEEnd-->
I'm actually curious how they determine the actual numbers. But I do know that Fenway Park has been the best hitter's park in the majors for a long time, and by a <i>mile</i>.
Just one example of a thousand, but during Wade Boggs's prime, he routinely hit about .400 in Fenway, and about .300 on the road. When ranking a player, to not adjust for that kind of thing is bad front office work.
There are some parks, like Petco Park in S.D., that take away so many homers that it's simply not fair to compare Adrian Gonzo to a player in Philly or Cinci, who play half their games in home-run bandboxes.
<!--QuoteEnd--></div><!--QuoteEEnd-->
I understand that, I just think its too overboard. Also, not everyone is a HR hitter. Ichiro has gotten HUGE boosts to his OPS+ playing in SafeCo, even though you could argue pretty easily that playing in a huge stadium is actually very good for his slap/speed style of hitting.
WAR, what is it good for? - AnnoCatuli - 03-01-2010
I have to admit, I have no idea what the hell WAR is, or how its calculated. Sure, its Wins about Replacement, but I do not fully understand it.
I look at OBP and SLG more so than batting average. Batting average basically says a single is the same as a HR. Slugging, On base and on base plus slugging are among the most important stats in the game.
WAR, what is it good for? - Destined - 03-01-2010
<!--quoteo(post=81139:date=Mar 1 2010, 06:51 PM:name=AnnoCatuli)-->QUOTE (AnnoCatuli @ Mar 1 2010, 06:51 PM) <{POST_SNAPBACK}><!--quotec-->I have to admit, I have no idea what the hell WAR is, or how its calculated. Sure, its Wins about Replacement, but I do not fully understand it.
I look at OBP and SLG more so than batting average. Batting average basically says a single is the same as a HR. Slugging, On base and on base plus slugging are among the most important stats in the game.<!--QuoteEnd--><!--QuoteEEnd-->
The most important stat is WS Titles. All of these new useless stats stem from over-analyzing. WAR, OBP+, ERA+, etc. are just annoying ways of trying to make laws that don't exist. Bloop singles have won World Series Game 7s and Cy Young winners don't usually win the World Series.
WAR, what is it good for? - Fly - 03-01-2010
<!--quoteo(post=81140:date=Mar 1 2010, 09:57 PM:name=Destined)-->QUOTE (Destined @ Mar 1 2010, 09:57 PM) <{POST_SNAPBACK}><!--quotec--><!--quoteo(post=81139:date=Mar 1 2010, 06:51 PM:name=AnnoCatuli)--><div class='quotetop'>QUOTE (AnnoCatuli @ Mar 1 2010, 06:51 PM) <{POST_SNAPBACK}><!--quotec-->I have to admit, I have no idea what the hell WAR is, or how its calculated. Sure, its Wins about Replacement, but I do not fully understand it.
I look at OBP and SLG more so than batting average. Batting average basically says a single is the same as a HR. Slugging, On base and on base plus slugging are among the most important stats in the game.<!--QuoteEnd--><!--QuoteEEnd-->
The most important stat is WS Titles. All of these new useless stats stem from over-analyzing. WAR, OBP+, ERA+, etc. are just annoying ways of trying to make laws that don't exist. Bloop singles have won World Series Game 7s and Cy Young winners don't usually win the World Series.
<!--QuoteEnd--></div><!--QuoteEEnd-->
Depends on what you're using stats for. If you're using stats to determine the value of players, then the "useless" stats are the way to go. If you're using stats to determine the best and/or hottest teams in October, then use World Series titles. Of course, the entire purpose of statistics is to try to determine the value of a player.
If you're using WS titles as the "most important stat", I hope you are prepared to say that Scott Brosius was a better player than Barry Bonds.