Any new Saber stats in OOTP 06?

[flubug31]

Originally Posted by DaMook

Markus I mean no disrepect but how can you claim that when Win Shares is highly regarded in the SABER community? Is probably the single most useful stats we have.

It isn't highly regarded. The consensus is that the replacement level is way too low and that the defensive system is poor. The single most useful stat we have is custom linear weights. Practically speaking, the single most useful stat we have is VORPD.

[mrbill]

I just read a quick argument against the basic calculation of WS that I thought was interesting.

Basically, win shares are split up based on wins a team actually had in a season. But, its done on run production stats, so why is the base number of win shares the *actual* number of wins a team got, rather than the pythagorean record of the team?

For instance, if a team outperforms by 10 wins, why do they get 30 extra win shares to split up? The abstraction seems to get broken when you work off stats and runs for almost everything, except the base number of WS for a team, which comes from the actual output of wins.

[DaMook]

Originally Posted by flubug31

It isn't highly regarded. The consensus is that the replacement level is way too low and that the defensive system is poor. The single most useful stat we have is custom linear weights. Practically speaking, the single most useful stat we have is VORPD.

What is VORPD? I have never heard of it and I cant find it mentioned anywhere.

[SandMan]

Originally Posted by mrbill

Basically, win shares are split up based on wins a team actually had in a season. But, its done on run production stats, so why is the base number of win shares the *actual* number of wins a team got, rather than the pythagorean record of the team?

WIN SHARES! Not pythagorean record shares..... just a way of getting the value of a player based on his contribution to the team. A good player will get a good amount of winshares regardless if he is on a good team or a terrible team. It is based on what a team or player did not what he could or should have done........ this has been clarified by Bill James in both the New Historical Abstract and the Win Shares book..... READ both books and you will see that his method of defining shares over team and eras is very accurate and probably the best system that is out there including Pete Palmers Linear Weights and other systems with serious flaws.....

[andymac]

*sigh*

Now we are on the way to a SABR vs. SABR flame war.

[boilermaker]

I can't follow anything more complicated than WHIP.

I wouldn't mind the inclusion of these fancy nancy numbers as I think it would be an opportunity (maybe excuse is a better word) for me to learn about them. But I won't lose sleep over not having them. Avg., HRs, RBIs, ERA, K/BB does right by me.

[Russ]

Originally Posted by DaMook

What is VORPD? I have never heard of it and I cant find it mentioned anywhere.

Check this out:

www.stathead.com/bbeng/woolner/statglossary.htm

[disposableheros]

nm. what i was going to say has been said by others. i just didnt read the whole thread before replying.

[tysok]

Originally Posted by DaMook

Um no dude. Win Shares arent like AVG, OBP and SLG where you can have high numbers with low ABs. With Win Shares you have to play. If you dont play, you arent contributing. A Backup isnt going to have more win shares than a starter.

That's why my post said "right now", like May 18 this season in MLB. At the end of the season that back up will be a blip on your win shares list. But because of possible injury and he's red hot right nowhe could have big win share numbers at the moment.

It can tell you who's contributing right now to your team, but not really any more than looking at the other stats. It's a tool to sort out the best of the best at the end of the season with a 30+ pool of top tier players. It isn't useless midseason, but I think it would be next to useless.

[Doughnuts?]

Can anyone explain why there is a large section in the league leaders section for WINSHARES if Markus is so opposed to it. Will this be removed before release?

[bababui]

Doughnuts? who the hell is that in your avatar?

[Malleus Dei]

Originally Posted by flubug31

It isn't highly regarded.

Bullcrap.

[mrbill]

Originally Posted by SandMan

WIN SHARES! Not pythagorean record shares..... just a way of getting the value of a player based on his contribution to the team. A good player will get a good amount of winshares regardless if he is on a good team or a terrible team. It is based on what a team or player did not what he could or should have done........ this has been clarified by Bill James in both the New Historical Abstract and the Win Shares book..... READ both books and you will see that his method of defining shares over team and eras is very accurate and probably the best system that is out there including Pete Palmers Linear Weights and other systems with serious flaws.....

Just to be clear, I only posted that as an interesting interpretation that I read in a review of one of the books, not that I feverishly stand by that point.

In fact, I just ordered a few books, not the Win Shares one to start, but his Historical Abstract.

I hope to form my own opinions soon enough, just wanted to see what the response was. My first impression of all this sabermetric stuff is that its all about "what a player should do" rather than what stats you actually saw. So, just because you saw 95 wins, doesn't mean they should have won 95 games, and statistics in the saber-world are supposed to show you what should happen. But, hey, thats why I'm getting the books.

For future reference, capitalizing "READ" entirely throws the tone of your post off and makes me dislike reading it, rather than feeling enlightened by your clear rebuttal to the reasoning I posted.

[tysok]

There's not one nice little box to put sabermetrics into. Some show what should have been done, some deal with what was done.

The pythagorean wins is a simple tool to do a simple job, just to give you an idea.

Win Shares is a very complex tool to do a complex job, to show you what (and who) happened. In the book, if I recall correctly, he spends most of his time (when he's not explaining the formula that is) relating it to the MVP award. For that use you HAVE to take into account what really happened, not what might/should/or could have happened.

I think there are more of these sabermetric stats that deal with reality as well, so it's like two different races of the same species.

It is obviously a highly regarded stat, and makes good sense even though it's not 100% perfect. I had it programmed into a database for a long time (lost it when I forgot to copy a directory over when I rebuilt my computer) for another internet based baseball game. I used it once with OOTP5, and once with Baseball Mogul... but getting the stats moved from game to database was a monstrous pain in the *** so I never did it again.

[SandMan]

Originally Posted by tysok

That's why my post said "right now", like May 18 this season in MLB. At the end of the season that back up will be a blip on your win shares list. But because of possible injury and he's red hot right nowhe could have big win share numbers at the moment.

It can tell you who's contributing right now to your team, but not really any more than looking at the other stats. It's a tool to sort out the best of the best at the end of the season with a 30+ pool of top tier players. It isn't useless midseason, but I think it would be next to useless.

I agree 100%, I made this type of quote last year and was blasted for saying it, and I like WINSHARES....

I have made a post early last year that I thought win shares was a useless stat during the season but is quite valuable at the end. I stand by this since a player can have cold and hot spells as well as the team having long winning or losing streaks in the year.

While I don't think it is needed in game it would be nice to have it at the end of season in the Almanac. I use WINSHARES to help determine my HOF players. Anyone over 400 is automatic, 300 gets a strong look and 250 plus gets a look... The are kind of like Bill James old standards of A, B and C type HOF players... just using numbers that should get them in. Look at the players in and the list of players over 350 are in or will be when they retire and a lot of bad choice HOF selections still have 250 or more Win Shares for their career.

[nestajones]

Originally Posted by DaMook

RBI, Errors, are beyond being flawed, they are just plain useless.

Completely ignorant statement.

[nestajones]

Originally Posted by tysok

Win Shares during the season are useless anyhow. Right now Derrek Lee of the Cubs won't have many win shares, even though he's obviously one of the best players they have. Can't think right now of someone who's struggling, or someone who always starts off slow but gets on fire in the second half... but both of those would show similar results. He doesn't look all that good now (WS wise) but at the end of the season he may end up being the best player the team had.

Homeruns during the season are useless anyhow. Right now Derrek Lee of the Cubs won't have many homers, even though he's obviously one of the best players they have. Can't think right now of someone who's struggling, or someone who always starts off slow but gets on fire in the second half... but both of those would show similar results. He doesn't look all that good now (HR wise) but at the end of the season he may end up being the best player the team had.

[disposableheros]

Originally Posted by tysok

There's not one nice little box to put sabermetrics into. Some show what should have been done, some deal with what was done.

The pythagorean wins is a simple tool to do a simple job, just to give you an idea.

Win Shares is a very complex tool to do a complex job, to show you what (and who) happened. In the book, if I recall correctly, he spends most of his time (when he's not explaining the formula that is) relating it to the MVP award. For that use you HAVE to take into account what really happened, not what might/should/or could have happened.

I think there are more of these sabermetric stats that deal with reality as well, so it's like two different races of the same species.

It is obviously a highly regarded stat, and makes good sense even though it's not 100% perfect. I had it programmed into a database for a long time (lost it when I forgot to copy a directory over when I rebuilt my computer) for another internet based baseball game. I used it once with OOTP5, and once with Baseball Mogul... but getting the stats moved from game to database was a monstrous pain in the *** so I never did it again.

its a lot less highly regarded than it was 5 years ago. lots of good thorough rebuttals have come out in the last couple years to, IMO, rightfully knock it off the grand pedestal it was on for a long time.

im not saying its useless, but its definitely got glaring flaws that, again IMO, leave it far from the 1 stat that NEEDS to be in the game.

[andymac]

Originally Posted by nestajones

Completely ignorant statement.

You are right, he should have said almost useless.

[Malleus Dei]

Originally Posted by disposableheros

its a lot less highly regarded than it was 5 years ago

Now that statement is not too far from the truth - if you removed the "lot" from it and just wrote "it's less highly regarded these days than it used to be and other people have even improved on it" I doubt if anyone would argue with you.

But it is still highly regarded.