Year end awards are pretty borked

[Claybor]

I just got version 16, and wow I have never seen any version blow a year end award this badly.

13-12 3.10 wins the cy young, and 22-6 3.31 does not even make the top 3? I think someone made some poor choices when 'fixing' (breaking) year end awards.

[Kent814]

Originally Posted by Claybor

I just got version 16, and wow I have never seen any version blow a year end award this badly.

13-12 3.10 wins the cy young, and 22-6 3.31 does not even make the top 3? I think someone made some poor choices when 'fixing' (breaking) year end awards.

What were their peripherals and FIP? Ootps award selection is a lot more sabermetruc friendly than real life

[Bluenoser]

Show us all their numbers please.

OOTP doesn't award the Cy based on wins/losses.

[Claybor]

Sabermetrics rule the roost these days. Wins mean nothing. I think I was thrown by the fact that in every previous version of the game the 22-6 would have won, but with this version wins don't have any value at all.

[Déjà Bru]

Originally Posted by Claybor

Sabermetrics rule the roost these days. Wins mean nothing. I think I was thrown by the fact that in every previous version of the game the 22-6 would have won, but with this version wins don't have any value at all.

And, with due respect for what BN just said, is not right and not realistic, current mania for sabermetrics regardless. A 13-game winner does not win the CYA over a 22-game winner, sorry. FIP does not rule in real life.

[Honorable_Pawn]

Originally Posted by Déjà Bru

And, with due respect for what BN just said, is not right and not realistic, current mania for sabermetrics regardless. A 13-game winner does not win the CYA over a 22-game winner, sorry. FIP does not rule in real life.

I have to agree. As much as we discount wins around here as a legitimate stat they are the single-most important stat IRL.

[RchW]

Originally Posted by Déjà Bru

And, with due respect for what BN just said, is not right and not realistic, current mania for sabermetrics regardless. A 13-game winner does not win the CYA over a 22-game winner, sorry. FIP does not rule in real life.

Originally Posted by Honorable_Pawn

I have to agree. As much as we discount wins around here as a legitimate stat they are the single-most important stat IRL.

Felix Hernandez Pitching Statistics and History | Baseball-Reference.com

See 2010 where he beat out two 19 game winners and a 21 game winner quite correctly and note that his 19-win season in 2009 lost to:

Zack Greinke Pitching Statistics and History | Baseball-Reference.com

And Greinke's mate

Clayton Kershaw Pitching Statistics and History | Baseball-Reference.com

beat out two 19 game winners in 2013.

Wins matter when two pitchers are close but there are much better and real ways to evaluate all pitchers. When wins come with performance then you get winners like:

Roy Halladay Pitching Statistics and History | Baseball-Reference.com

in 2010. However that is offset by 2005 where the winners in both leagues should have finished 4th or lower and Johann Santana got robbed.

2005 Awards Voting | Baseball-Reference.com

[NoOne]

lol someone beat me to it! felix fernandez was in my head the whole time reading.

also, not only would we need their info and all the competition's too, but please don't think i am making requesting.

------

wins aren't meaningless to the team, but multiple factors go into winning or losing that are not within the pitcher's control. therefore, it makes sense to discount the traditional importance placed on a starter's win/loss record.

wins do not hold a consitent value relative to a starting pitcher's record. so, it's innacurate to compare them between players as if they mean the same thing.

ERA is similar, although to a lesser extent. so, they created FIP. FIP only attempts to factor out defense and ballpark. that doesn't mean it does a good job. at minimum it's a valid effort. i'm only saying it is not perfect in nature. it still has good uses, like wins or ERA.

wins are still useful for pitching awards, but they should not carry their traditional importance. imo, a tie-breaker is a fitting use for wins.

[magritte]

But if you're trying to model how Cy Young award winners were picked, wins and losses were always very important. So for historical reasons, no, I don't want to see .500 pitchers chosen. Even if they really maybe were the best pitchers in the league, like Dave Stieb in 1984 and 1985. Ideally, you'd look statistically at past votes and figure out which statistics drive the awards.

I'm not a fan of FIP except for relief pitchers or for examining year-to-year fluctuations of individual pitchers. It's less volatile than ERA and less affected (though not unaffected by ballparks), it's systematically biased by style of pitcher, and you can demonstrate that some pitchers consistently post E.R.A.'s significantly better than their FIP's through several thousand innings, and several changes of team. Even if you really believe that whether a ball is turned into an out or not is mostly fielding, there's all kinds of other stuff you're losing: the impact of a pitcher's own fielding on his the runs scored against him, his ability to hold runners, and pitch from the stretch. Not to mention how the pitcher performed under pressure. I realize the sabermetricians want to believe that everybody performs equally well regardless of circumstance, but even if it's true that whether a pitcher gives up more hits or fewer with men on base is totally random, it still has an impact on whether or not the team wins the game.

[NoOne]

ah, historical league... i didn't know aobut that little morsel of info.

get in there by hand and pick based on era/wins (or whatever you want), if you can be objective and not favor your team or certain historical names... and if you can't, that shouldn't stop you either.

exactly! sometimes people take these derivative stats as the word of god. e.g. you can say objectively that a .299 hit more than a .280 hitter (same AB, and didn't say 'better'), but you cannot say a 9war is more valueable than an 8war player(same games, ab, position or whatever). war isn't absolute in nature - it is not perfect. so, when it's close, it doesn't tell you much in that situation. what constitutes close? good question. ... <cricket noises>

you can find examples of imperfect attempts to neutralize factors, subjective data, or even simply flawed logic to begin with in more than one advanced metric. even the standard statistics have quite a few logical flaws and general uselessness.

i am not "for" sabermetrics. i am for improved data collection and data mining techniques, which allows for greater understanding of what factors correlate with success and their weight. how can anyone not want that? (magritte: i know you werent saying this, i just ranted a bit)

however, it's not right to say sabermetricians believe that everybody performs equally regardless of circumstances. this sounds like someone took the definition for the law of independent resutls and took a dump on it. (i assume and hope you read that line somehwere else and took their word for it, otherwise reword that last sentence as you re-read it.)

e.g. what you said about the idea of whip and random cluttering of them etc... the cluttering is random but it is more likely to happen with a pitcher who gives up more hits/bb per inning. each pitcher influences this random effect. each situation has a different probability based on a plethora of factors, and they do change constantly relative to the situation.

that concept is incontrevertable. it is our implementation, perception or calculation of a probability or statistic that might be flawed.

sabermetrics is about comparing apples to apples. it is about understanding each individual part and how it adds up to the whole, while maintaining a proper perspective on what success is for any given situation.

the same things standard baseball stats try to do... give you a way to compare players. it's really silly to distinguish between the two, imo. the same reasons for flaws exist on both sides of the line. i mean, explain why the weighting of OPS is logical... you can't. it's a poorly designed stat. there are many of them, whether 100 years old or 10 years old.

[RchW]

Timely article.

The Greatest Non-Cy Young Seasons in History | FanGraphs Baseball

[magritte]

@No one, I don't think you've quite caught my meaning. I think FIP is an overrated stat because it throws away too much baby with the bathwater, losing some valuable signal that ERA captures in attempting to lose noise. And it bothers me how statistic is getting used. Let's look, for a concrete example, at Zack Greinke's season. He had an extraordinarily low ERA and an FIP that was the best of his career but not that much better than in previous years. So if you're asking me if Greinke's likely to post a similar ERA next year, I'd have to say no. He has consistently posted ERA's that are better than his FIP's for several years but the spread is huge this year and likely unsustainable. So if I were being asked to predict who will have a better year next year, I would pick Kershaw over Greinke.

But choosing the Cy Young award isn't about projecting the future, it's about who had the better year this year. And the fact is, while batters were more successful in making contact against Greinke, they did much less with it. I'm extremely dubious of the implication of FIP that the difference between these two pitchers had much to do with the defense behind them. They played for the same team and I find it hard to imagine that the defense was that much better behind Greinke. It likely had more to do with where and when those balls were hit. And that difference mattered; the opposing team scored significantly fewer runs against Greinke and the Dodgers won more games behind him. Whether it was likely to fall out that way is to my mind irrelevant, that's what did happen. And I think that sabermetricians touting the lesser year-to-year volatility of FIP as an indication of greater accuracy at measuring pitching ability are erring. ERA is more volatile at least in part because it's based on runs which are rarer events than walks and strikeouts, but it comes closer to what actually matters in terms of winning baseball games than FIP.

Yes, when we look at situational stats like batting average with runners in scoring position, they bounce around a lot from year to year due to small sample sizes. It's very difficult to establish whether somebody is really a good hitter under pressure from the data of a year or two unless they manage something really freakish statistically. Whether a home run was hit with the bases empty or the bases loaded, whether it was a close game or a blow-out, may not be under the player's control, but it mattered. And in assessing a player's value this year (not what he might do next year), it's important. A lot of people now dismiss RBI's because we know that typically players get a lot of RBI's because they come up to bat with a lot of runners on base. We have the stats to check for this now, and see whether the player with more RBI's had more opportunities or was more efficient. But rather than looking at more data, it seems like a lot of people want to just look at WAR.

I think the fashion has swung too far toward the synthetic stats and a lot of sportswriters are treating the numbers uncritically, not considering what is lost in trying to boil down player performance to a single magic number. And it bugs me that--at least as used by sportswriters--none of these stats are ever presented with any kind of standard error estimate.

[Honorable_Pawn]

The Rays keep track of a stat that measures the speed of the ball coming off the bat in mph. When batters hit the ball hard they are seeing the pitch better and it is a good indication of whether or not a pitcher is effective.

In fact, the Rays do not like to let pitchers be seen 3+ times in a game for this very reason. Batters tend to have a higher "speed of contact" after seeing a pitcher multiple times. FIP, BABIP, WHIP, and ERA are all important stats. But sometimes we try to glean too much information from a specific stat. No stat is 100% comprehensive. Sometimes you just have to watch the game. Sometimes you see a pitcher getting weak contact time and time again--the batters just aren't getting good wood. Sometimes the mph off the bat is very soft. Does this mean that the defense did all the work?

Hardly.

Sometimes a good pitcher pitches to contact and it has nothing to do with defense. The defense is not the reason the batters couldn't get good wood. In this situation FIP is highly-overrated. Sometimes a pitcher that just wins games is a better pitcher than a pitcher that has good indicators. Sometimes not. But this modern idea of completely discounting wins is absolutely absurd.

A good pitcher wins games--stats be damned.

Just saying.

[Honorable_Pawn]

Do I think Wins is the most important stat?

Nope.

Do I think it is one of the most important stats?

Nope.

Do I think it is a worthless stat?

Nope.

[RchW]

Originally Posted by Honorable_Pawn

The Rays keep track of a stat that measures the speed of the ball coming off the bat in mph. When batters hit the ball hard they are seeing the pitch better and it is a good indication of whether or not a pitcher is effective.

In fact, the Rays do not like to let pitchers be seen 3+ times in a game for this very reason. Batters tend to have a higher "speed of contact" after seeing a pitcher multiple times. FIP, BABIP, WHIP, and ERA are all important stats. But sometimes we try to glean too much information from a specific stat. No stat is 100% comprehensive. Sometimes you just have to watch the game. Sometimes you see a pitcher getting weak contact time and time again--the batters just aren't getting good wood. Sometimes the mph off the bat is very soft. Does this mean that the defense did all the work?

Hardly.

Sometimes a good pitcher pitches to contact and it has nothing to do with defense. The defense is not the reason the batters couldn't get good wood. In this situation FIP is highly-overrated. Sometimes a pitcher that just wins games is a better pitcher than a pitcher that has good indicators. Sometimes not. But this modern idea of completely discounting wins is absolutely absurd.

A good pitcher wins games--stats be damned.

Just saying.

That's not always true. A good pitcher on an offensively challenged team may pitch better than anyone and not be in the lead when he comes out. OTOH a bad pitcher may get 8 runs scored for him while giving up 6 and get wins that had nothing to do with his performance.

Exhibit 1

Drew Hutchison Statistics and History | Baseball-Reference.com

13-5, 5.57 ERA, WHIP 1.483 and 6.9 runs scored for him per start. That's a pitcher that had little to do with the wins he got.

[magritte]

The Toronto Blue Jays must have agreed with you, since they sent him to the minors at one point last year. It looks like they don't put a lot of faith in FIP either, since his FIP (4.42) was better than R.A. Dickey's (11-11, 3.91 ERA, 4.48 FIP) and nearly the same as Marco Estrada (13-8, 3.13 ERA, 4.40 FIP).

[saturn2187]

I just finished a season and it gave both "Reliever of the Year" awards to middle relievers, rather than closers. I get that you don't need to be a closer to have the best stats, and that SV are overrated, etc but I thought the intent of the award was to simulate baseball's reliever award? Here are their picks:

Dave Pagan - 6-2, 3.04 ERA, 1.4 WAR, 1.26 WHIP, 2 SV
Steve Foucault 7-7, 18 SV, 2.34 ERA, 1.0 WAR, 1.00 WHIP
Ed Glynn : 5-1, 4 SV, 2.65 ERA, 1.2 WAR

Randy Moffitt - 11-5, 2 SV, 3.22 ERA, 2.0 WAR
Joe Sambito - 4-5, 1 SV, 3.27 ERA, 1.4 WAR
Dave Tomlin - 0-4, 35 SV, 1.0 WAR

Some guys they DIDN'T pick
Bruce Sutter 4-3, 23 SV, 2.17 ERA, 1.2 WAR
Warren Brustar 5-2, 24 SV, 1.14 ERA, 1.0 WAR
Kent Tekulve 3-5, 28 SV, 2.26 ERA, 1.5 WAR
Rollie Fingers 3-6, 26 SV, 2.95 ERA, 1.4 WAR
Tom Johnson 11-6, 20 SV, 2.61 ERA, 1.2 WAR

Dave pagan is on my team, but I feel like he didn't earn that award over some of those bottom names. 2 SV and a 3.04 ERA? His peripherals weren't even great. I think the voters would never even consider him for Reliever of the Year in real life, past, present, or future, over some of these names down bottom.