War is trash . Well not complete trash , but I'd prefer a lot of stats over that . For pitching , era+/- , FIP , hell even whip and k/bb all are better For hitting , ops+ , wrc+ , obp , babip etc Defensive metrics are no Bueno right now especially with all the shifting going on . They might be viable for outfielders .
Over time WAR and defensive metrics have really matched up well to what people were saying. WAR is a great stat. It doesn't make it perfect, but at the end of the year I bet it matches pretty well with who the best players are in baseball. There can be outlier years for defensive metrics, but over time they do eventually match up well with reputation.
All of the other stats on the WAR line are through last night, so I assume the WAR is current since everything else is. http://www.espn.com/mlb/player/stats/_/id/31815/dallas-keuchel
On the other hand, Keuchel is elite in GB% and LOB%. All things equal I'd prefer the Sale-type pitcher but Keuchel is doing some incredible things. His GB% and LOB% are even higher than his Cy Young year. So far the only thing significantly worse than that season is the HR/FB ratio. I'd argue preventing baserunners from scoring is as important as preventing base runners.
Doubt it. That page says the stats come from the Elias Sports Bureau (internal ESPN) but the WAR value comes from Baseball Reference and BR hasn't updated their stats to reflect last nights game yet.
I agree, but that goes to my reply to Joe Joe. That's highly influenced by the inputs of what you decide a replacement player is, which (unless I misunderstand) is not necessarily a hard and fast thing. So the inputs into the WAR formula dramatically affect the outputs of the WAR formula.
My problem with stats like ISO is that things like position matter a lot. When team WAR is plotted with wins, it has a very linear relationship. When preseason team winning percentage projections based on estimated WAR of individual players outperforms the first 50 games team winning percentages for predicting the next 112 games, I have a hard time thinking the problems in WAR just are that bad. Having some subjectivity...though I believe WAR models will remove more of this over time, is better than perfectly measured stats that are used more subjectivity to determine value of a player. I have a big problem with the argument that... Since we can't measure X perfectly, we should ignore it even though we know X is important and use Y. Y ignores X, but because we can measure it perfectly, it is less subjective...hence better. Most baseball stats are complex relationships between batter, pitcher, and defense. I hate expressions like the pitcher only allowed 2 runs today. He wasn't only one involved. Maybe he pitched well enough to allow 4 runs, but CF robbed a 2 run homer on a play that is only made once every couple of years.
PS just to get my thoughts out there and not directed at anyone specifically, I am not saying fWAR should be end of any discussion, but it is usually within the ballpark. There are things I do to see confidence in a WAR value like subbing in different defensive run models (DRS vs. UZR). A lot of people tend to think elite players not named Trout (who still is undervalued and likely will lead AL in WAR missing 2 months), are worth more than they are. Teams with WAR consolidated in star players win as many games on average (though some variance) as teams with similar WAR totals, but spread across the team. A 4.5 WAR player is better than a 3.5 WAR player, but it isn't some huge gap as some people think. Astros to me exemplify what happens when you have stars, but also get value from a bunch of 2nd tier players.
Sorry. To clarify, when I used it to make the argument that Correa was #5, it was in the context of offensive stats, so I was using Offensive WAR from ESPN: http://www.espn.com/mlb/war/leaders/_/year/2016/type/seasonal/alltime/false/count/1/sort/OFF He was ranked 19th in total WAR last year.
I will say offensive and defensive WAR values are weird. Should position value be included in offensive WAR?
Baseball-Reference WAR is actually abbreviated rWAR. The weird part is the r doesn't stand for reference, but rally...
