VORP as a measuring stick

[Syd Thrift]

Originally Posted by dougaiton

Well, it's built to give good grades to people historically who are considered to have been good defensive players, or, to put it the proper way round, built to grade well those that history records as having been particularly good defensively. Basically, the human eye could pretty much tell that Bill Mazerowski was better than Jose Offerman at 2B. More importantly than that, however, the statistics strongly implied it. We know that Rey Sanchez and the Mets infield of 2001 was excellent because it didn't make any errors, because the pitching staff had good ERAs, and because the BABIP were low.

For example, in the Abstract, James details quite clearly that his thoughts on how to rate catchers were based around 'Why do current statistics not rate Bench highly?'. We know that DPs turned is an important part of WS allocation, despite the fact that DPs turned are as context dependent as RBIs, for the most part, but Mazerowski was pretty darn good at them, so RACK that up in the formula. And so it goes: strikeout staffs get magically good defenses behind them, flyball pitching-staffs have remarkably good outfields behind them etc. etc.

Most importantly is really the thing that keeps UZR far ahead of WinShares: what are the bricks the building was constructed from? WinShares uses the same ol' bricks that we are told to believe gives an unreliable account of a player's defence. If I was to take the team DERA, and then compare it with the team ERA, and then assigned each portion of that ERA (-0.5, +1) to infielders and outfielders based on a basic OOTP fielding report, then I'd be just as theoretically sound as WS! Sure, we don't have better bricks from the past, but still...

I'm being unkind, and I take the point the WinShares, barring Rate of course, may be the best historical record on the market of players defensive record. But it doesn't mean I'd rather have WS than VORP!

Obviously, I'm preaching to the enlightened, but for those interested in the maths behind WinShares, try this:
http://walksaber.blogspot.com/2005/1...ough-pt-5.html
Parts 1-4 are the relevant ones, but this one links you to them all.

I totally agree that when it comes to all the stats that we have available to us right now, UZR and other stats are better than Win Shares. When it comes to comparing 2 shortstops from 1925 and all you have are putouts, assists, and errors, WS is a little better than FW. When you're comparing anybody but middle infielders, WS is a LOT better than FW. Given that, and given that the game is mostly a historical sim (I think, anyway), there's a really good argument for Win Shares.

[cabot41]

I'm running into some outsized OOTP-style VORP calculations in my fictional sims because of stolen bases. I've got everything set to modern day standards except for stolen bases (I'd like my 20/20/20 speed guys to be able to steal 100 bases ... just fun for me). But now I've got 1.100 OPS guys who are slow losing MVP battles to the .890 OPS gusy with 80 steals.

.420 obp/470slg with 80 steals is a really nice leadoff hitter ... but a measurment that grades him as more valuable than the .420/.680 60 HR guy has got to be slightly off, right?

[Carplos]

I'd say it's off some... but to be fair, if stolen bases are that much a part of your league... it does make sense to consider them a bit more than modern MLB.

[dougaiton]

Originally Posted by Syd Thrift

I totally agree that when it comes to all the stats that we have available to us right now, UZR and other stats are better than Win Shares. When it comes to comparing 2 shortstops from 1925 and all you have are putouts, assists, and errors, WS is a little better than FW. When you're comparing anybody but middle infielders, WS is a LOT better than FW. Given that, and given that the game is mostly a historical sim (I think, anyway), there's a really good argument for Win Shares.

YEAH OKAY SO YOU WIN THAT ROUND BUT ILL BE BACK FOR ROUND2.

[tysok]

Originally Posted by dougaiton

With men on base, leverage changes massively, and pitching five innings, starters are going to face far more high-leverage situations, because men on base is a far greater indicator of leverage than the point in the game you're at. Because starters pitch more innings, they'll pitch with more men on base, and those innings (where there are baserunners) are far more important to deciding a game than a 9th.

Again, Rivera may face a higher-percentage of high-leverage innings, but a starter will pitch many more high-leverage innings in a season, and as Win Shares is a counting stat, then starters should get more because they do more.

The thing no one ever mentions is that starters never face a situation where every pitch in this at bat could mean the difference between winning the ball game or losing it.... in the 5th inning a batter can't get a walk off RBI single, or a walk off HR.

A closer rarely has any room to slip up... if a starter slips in the 5th he can struggle through and try and regroup for the 6th... and his team still has at least 4 more times to try and recover. If a closer slips he has to regroup immediately, and there is no room for his batters to try and recover.

[Bay Area Bob]

Originally Posted by tysok

Win Shares.... Win Shares... the name tells you why it takes into account save, wins and RBIs... Win Shares...

What would that be measuring? How much your performance affected winning. Getting a win means you pitched well enough to earn it... getting a save means you "saved" the win.

No. Save and Win are both arbitrarily defined. They are flawed metrics in terms of determining how much a player did to contribute to the win.

[dougaiton]

Originally Posted by tysok

The thing no one ever mentions is that starters never face a situation where every pitch in this at bat could mean the difference between winning the ball game or losing it.... in the 5th inning a batter can't get a walk off RBI single, or a walk off HR.

A closer rarely has any room to slip up... if a starter slips in the 5th he can struggle through and try and regroup for the 6th... and his team still has at least 4 more times to try and recover. If a closer slips he has to regroup immediately, and there is no room for his batters to try and recover.

Well, the exciting thing about win expectancy is that it tells us the opposite! Different pitchers pitch massively important innings in every game, often ones that can decide the game that early. Sure a closer can lose a game in the 9th, but a starter can lose a game in the 1st-6th. If you want a simple check, have a gander at the decisions: Ws and Ls are the preserve of starters, right?

Many baseball games are exciting, and come down to the last AB. Most, however, are not, and are decided long before the 9th, even if they end a little closer. I was listening to Padres/Phils recently on MLBAudio, and the Pads ran up a 9-2 lead early before it came down to 9-7 late. It ended close, ace releivers in the game, but the game was lost early.

That's overly simple, but there is no doubt that Paul Quantrill getting a DP with bases loaded, one out, one run lead in the 7th is more important than Rivera coming in bases empty with a one-lead in the 9th. That's why win expectancy is so exciting - it challenges us to think of not 'innings' (like the 9th) which change games, but 'situations' that change games.

[CBL-Commish]

Originally Posted by tysok

A closer rarely has any room to slip up... if a starter slips in the 5th he can struggle through and try and regroup for the 6th... and his team still has at least 4 more times to try and recover. If a closer slips he has to regroup immediately, and there is no room for his batters to try and recover.

Well, except for the fact that a large percentage of saves in today's game are three outs, up by two or three runs, starting the 9th with no one on. The closer can give up two or three solo bombs and still escape without a loss. He can give up a homer or two and still get the save while recording only three outs. Most saves aren't one-mistake-and-you're-done. Otherwise guys like Mike Willams and Jose Mesa and Antonio Alfonseca wouldn't have 85% save percentages.