Win Predictor Fun

[Guest holychicken]

How would one determine VORP?

[metsmarathon]

well, you could just go here or here and get the totals there!

granted, that's just not as much fun!

oh, and jose reyes is the NL leader, with a win percentage added of almost 0.800, while arod leads all players with a 1.01

[Guest Edgy DC]

How are you crediting the ptichers? Do they get full credit for all outs they get, or do they share it with the defense?

[metsmarathon]

and therein lies the flaw!

[Willets Point]

You guys are so wonderfully geeky. I don't think I could ever fathom what you're doing but more power to you.

[Benjamin Grimm]

Edgy DC wrote: How are you crediting the ptichers? Do they get full credit for all outs they get, or do they share it with the defense?

It seems like it would go to the pitcher for getting the out.

Players get credit for reaching base on an error. And I think the pitcher would be penalized for the error even if someone else made it.

[metsmarathon]

and yancy, take the info from the game graphs or the box score page, and the teams won't be intermingled. you'll still have to combine your pitchers' contributions with the bat into his contributions with his arm, but its a start...

[Benjamin Grimm]

I've taken this as far as I care to.

The intermingled players are easy enough to sort out.

[Guest Edgy DC]

If pitchers get 100% of the credit for all outs, they'll win big time.

[metsmarathon]

y'know, it doesn't seem to play out that way.

last year:

beltran - 5.21
wright - 4.64
delgado - 3.11
reyes - 2.11
...
pedro - 0.60
bannister - 0.53
glavine - 0.46
...
wagner - 3.52
sanchez - 1.95
heilman - 1.58

and actually, if i counted their offensive contributions as well, pedro and glavine would both be in the negatives!

the reasoning being that by holding your opponent scoreless early in the game, you're not really contributing to your team winning so much as merely holding serve. at least, that's the best i could make of it.

to clarify, if the score is tied at the start and end of an inning, the win expectancy should remain virtually unchanged. only once you have a run differential does the win probability change after a scoreless inning.

[Guest Edgy DC]

Unfortunately, they get 100% of the blame for hits and shit also.

[Guest holychicken]

metsmarathon wrote: well, you could just go here or here and get the totals there! granted, that's just not as much fun! oh, and jose reyes is the NL leader, with a win percentage added of almost 0.800, while arod leads all players with a 1.01

Nice to ruin our fun, jerk.

But anyway. . . as others have already illustrated

Imagine a game where two pitchers throw perfect games into the 9th and then in the bottom of the 9th with 2 outs the guy pinch-hitting for the pitcher hits a homerun.

He wasn't in the game at all until that point so he did almost nothing ESPECIALLY compared to the pitcher and he would get all of the credit for the win and the other pitcher would get the entire credit for the loss.

I am wondering if there is anyway to edit the win predictor % to account for this . . . or to have an entirely different graph that favors the pitcher more than this seems to favor the hitters.

[Willets Point]

holychicken wrote: Imagine a game where two pitchers throw perfect games into the 9th and then in the bottom of the 9th with 2 outs the guy pinch-hitting for the pitcher hits a homerun.

That sounds like a beautiful game that I would like to see.

[Guest Edgy DC]

The losing pitcher wouldn't get all the blame. The lion's share would go to his failing hitters.

[Guest Edgy DC]

Using 2001-2006 data, I calculate that the ill-begotten losing pitcher (or the defense) would have had .697 win shares (where a team nets 1.0 for a win and 0.0 for a loss) before the homer and a mere .27 on the day after the homer.

Subtract from that the -.10 his hitting would have garnered, and it hardly seems worth the effort.

But measuered against the perfomance of a theoretical replacement player, I'm sure his day would prove statistically worthwhile, if epically disappointing.

[metsmarathon]

Willets Point wrote: ="holychicken"] Imagine a game where two pitchers throw perfect games into the 9th and then in the bottom of the 9th with 2 outs the guy pinch-hitting for the pitcher hits a homerun. That sounds like a beautiful game that I would like to see.

actually, i've done more thinking...

the value of keeping the score tied for each inning is as follows, for home and away pitchers:

inning	Home P	Away P
1	0.051	0.058
2	0.049	0.049
3	0.049	0.057
4	0.065	0.061
5	0.066	0.073
6	0.078	0.080
7	0.089	0.088
8	0.114	0.113
9	0.134	0.136
total	0.695	0.715

home teams have about a 2% better chance of winning a given game!
oh, and i used the 1977-2006 data...

so the starters wouldnt get totally hosed... well, the losing starter would get kinda hosed.

[Guest Johnny Dickshot]

I like this stuff, even if I can't do it.

Here's a suggestion: Rather than screwing up "your vote" as to Schaefer PoTG by using things like this, why doncha create a "character" who will? You know what I'm saying?

You could call it MetBot. We could all judge our own interpretations against him, or perhaps, use him for guidance. MetBot could crunch all numbers here using whatever stuff you come up with.

[Guest Johnny Dickshot]

Scutaro breaks the meter:

[Guest Edgy DC]

They should paint that on the scoreboard and wear it on their jerseys.