Sox advanced stats standings

[voodoochile]

Interesting how much these differ site to site but still kind of cool if you're into this kind of thing...

WAR:
ESPN has Luis Robert 3rd in MLB at 2.2 Jose Abreu is 5th 1.9 and TA is 15th.1.5

Fangraphs has Luis Robert 9th 1.9 TA 19th 1.4 Abreu 21st 1.4 and Moncada 29th at 1.2

BBREF has Robert 3rd at 2.2 and Abreu 5th at 1.9 TA 16th at 1.5 and McCann 36th at 1.1

OPS+
BBREF has 4 Sox in the top 30 Robert, Abreu, TA and Eloy with the team over all second in the majors at 121 behind.... that team out west that shall remain nameless...

I could keep going, but that's a good sample and I know some are going to rant about propeller heads and made up stats, but just love seeing the Sox all over these advanced stats boards

[FielderJones]

I know some are going to rant about propeller heads and made up stats

The fact that some advanced stats like WAR are proprietary, and each site has its own methodology for determining the coefficients for weighting the components, gives some credence to the "made up stats" denunciation. WAR is not an objective stat, and each organization fits the numbers to its agenda.

[voodoochile]

The fact that some advanced stats like WAR are proprietary, and each site has its own methodology for determining the coefficients for weighting the components, gives some credence to the "made up stats" denunciation. WAR is not an objective stat, and each organization fits the numbers to its agenda.

Right, but they can be good for looking at meta data. Every site has these players among the leaders in WAR (for example) though they vary on how high up the leader board they are. I'm not looking for them to tell me anything specific about their production, but they are useful and fun for comparing players across the league. Kind of like OPS. It's not a stat that tells you anything specific but when you see a player with an OPS of .900+ you know he's having a good offensive year.

[JazzyCyclist]

The fact that some advanced stats like WAR are proprietary, and each site has its own methodology for determining the coefficients for weighting the components, gives some credence to the "made up stats" denunciation. WAR is not an objective stat, and each organization fits the numbers to its agenda.

True.

[HomeFish]

The fact that some advanced stats like WAR are proprietary, and each site has its own methodology for determining the coefficients for weighting the components, gives some credence to the "made up stats" denunciation. WAR is not an objective stat, and each organization fits the numbers to its agenda.

I'm curious what you think the "agenda" is for group of nerds x versus group of nerds y.

[HomeFish]

Fangraphs has the Sox at a 49% chance to win the division (highest in the AL Central) and 99% chance of making the playoffs. The White Sox also lead the AL in run differential and have the 2nd best run differential in all of baseball behind the Dodgers...yes, 2 runs better than you know who in Southern California.

[ChiTownTrojan]

I'm curious what you think the "agenda" is for group of nerds x versus group of nerds y.

There are two subjective decisions that are part of the development of a composite statistic like WAR: (1) which statistics to feed into the model, and (2) what modeling approach they use. It is possible that one "group of nerds" thinks that a certain statistic is more or less necessary to include in the model than another statistic, or that one of the groups has come up with a way of summarizing the data that is highly predictive of wins that another group has not. But none of these groups are purposely biasing their models to prop up certain types of players. They are all trying to do their best to come up with a number that is as highly correlated with wins as possible.

[ChiTownTrojan]

The fact that some advanced stats like WAR are proprietary, and each site has its own methodology for determining the coefficients for weighting the components, gives some credence to the "made up stats" denunciation. WAR is not an objective stat, and each organization fits the numbers to its agenda.

All stats are made up. A long time ago someone made up a stat where they took the number of hits a player got and divided it by the number of at bats, and they called the resulting proportion a "batting average." The reason it was made up a long time ago was because hits and at bats were among the easiest data points to collect a long time ago. Even the decision to count the number of times the ball goes over the fence after it is hit and call it "home runs" is a made up stat - in statistics there is a term for such a statistic (a "counting process").

Over the years we have realized that batting average does not tell the whole story about a player - not all .270 hitters are created equal. Now we have access to a lot more data points that tell a more complete story, things like launch angle, exit velocity, defensive metrics, etc. So people made up stats that are more fine-grained, and because there are more stats available they needed ways to summarize them into composites that are easier to digest than looking at 25 different numbers.

[voodoochile]

I just thought it was cool we had so many high ranked players. Agree with the stats or not it's nice to see the Sox players at the top of any leaderboard.

[FielderJones]

All stats are made up. A long time ago someone made up a stat where they took the number of hits a player got and divided it by the number of at bats, and they called the resulting proportion a "batting average." The reason it was made up a long time ago was because hits and at bats were among the easiest data points to collect a long time ago. Even the decision to count the number of times the ball goes over the fence after it is hit and call it "home runs" is a made up stat - in statistics there is a term for such a statistic (a "counting process").

Yes, but you and I can calculate batting average and come up with the same result. You and I can calculate WAR and not come up with the same result.

[voodoochile]

Yes, but you and I can calculate batting average and come up with the same result. You and I can calculate WAR and not come up with the same result.

Yes but if you check the WAR standings on most sites they tend to be fairly close. Obviously there are some discrepencies as noted above, but that mostly is shuffling the same 20 guys around the top 20 slots. No player is #3 on one site and #150 on a different one.

[ChiTownTrojan]

Yes, but you and I can calculate batting average and come up with the same result. You and I can calculate WAR and not come up with the same result.

That's only because they don't release the methodology, presumably because they put a lot of person hours into developing it and they rightfully want to be compensated for those hours. If fangraphs or BA or whatever released their methodology then it could be duplicated (theoretically exactly) by anyone else, and then less people would visit their websites.

It does not mean that they have an "agenda," other than to create something that the public finds useful (so they in turn make more of a profit).

[FielderJones]

There's some sort of agenda involved when a stat ranks Leury Garcia above Lucas Giolito. Imagine my skepticism.

[Mohoney]

There's some sort of agenda involved when a stat ranks Leury Garcia above Lucas Giolito. Imagine my skepticism.

Don’t compare pitchers to hitters. Compare them to other pitchers. And stick to Fangraphs instead of Baseball Reference.

[A. Cavatica]

There's some sort of agenda involved when a stat ranks Leury Garcia above Lucas Giolito. Imagine my skepticism.

It’s called alphabetical order, but it’s a made-up stat.