PDA

View Full Version : Starting Rotation and Bullpen


doctor30th
01-27-2004, 08:46 PM
My Friend writes programs and did a program similar to those that predict Baseball stats for a season. He showed me the sox "starting" rotation and bullpen; this is how it came out. He did several 5th starters based on some probability thing that he showed me that I could not understand.

Loaiza: 19-8, 3.19 ERA*
Buehrle: 18-10, 3.89 ERA*
Garland: 14-9, 4.03 ERA*
Schoeneweis: 15-11, 4.52 ERA*

5th Starters (if are 5th starters all year)
Person: 9-10, 4.92 ERA*
Stewart: 8-11, 5.54 ERA
Cotts: 9-8, 5.23 ERA
Diaz: 5-11, 5.35 ERA
Rauch: 7-10, 5.98 ERA

Bullpen:
Wright: 3-7, 4.38 ERA, 1 Save, 3 Holds, 4 BS, 40 SO, 14 BB.*
Marte: 3-2, 1.98 ERA, 6 Saves, 20 Holds, 1 BS, 84 SO, 22 BB.*
Wunsch: 3-3, 2.68 ERA, 10 Holds, 2 BS, 38 SO, 15 BB.*
Politte: 1-4, 3.98 ERA, 5 holds, 3 BS, 30 SO, 20 BB.*
Majewski: 1-3, 4.50 ERA, 8 holds, 1 BS, 22 SO, 10 BB.*
Takatsu: 2-5, 4.15 ERA, 15 holds, 3 BS, 54 SO, 23 BB.**
Adkins: 0-4, 4.99 ERA, 1 Hold, 4 BS, 15 SO, 20 BB.

Closer:
Koch: 5-4, 3.74 ERA, 34 Saves, 1 hld, 5 BS, 68 SO, 35 BB.*

*is the original program ran, these are the 11 pitchers that he used.
**Takatsu came in late in his programing so he didn't have complete stats to use for him. He was not inculded in the original 11.


Of course this all really means nothing but its interesting.

SEALgep
01-27-2004, 08:57 PM
If you use a fifth starter win 9 wins, that's 93 wins. I'm pretty sure that would win the division. I'd take that. I wonder in Schoeneweis can really knotch up 15 wins. I know its all speculation, but that would be a successful season.

doctor30th
01-27-2004, 08:59 PM
The original one did not include takatsu, so the sox finished 91-71.

soxfan26
01-27-2004, 09:02 PM
Originally posted by doctor30th

Garland: 14-9, 4.03 ERA*
Schoeneweis: 15-11, 4.52 ERA*

Koch: 5-4, 3.74 ERA, 34 Saves, 1 hld, 5 BS, 68 SO, 35 BB.*

Of course this all really means nothing but its interesting.

Interesting Indeed! If Garland and Schoeneweis turn in seasons like that, and Koch saves 34 games, I'll shave my head for game 1 of the ALDS!!!

doctor30th
01-27-2004, 09:07 PM
Interesting Indeed! If Garland and Schoeneweis turn in seasons like that, and Koch saves 34 games, I'll shave my head for game 1 of the ALDS!!!

Keep in mind that this was a program done by a human, a Cubs fan but still a human. I wouldn't go bet on it with my bookie because of it.

soxfan26
01-27-2004, 09:10 PM
Originally posted by doctor30th
Keep in mind that this was a program done by a human, a Cubs fan but still a human. I wouldn't go bet on it with my bookie because of it.

Don't worry about that. I'm not a gambling man, but if I were the last thing I would bet on would be Schoeneweis winning 15 games. For that matter I wouldn't bet on him winning 9, and this club is NOT going to win 91 or 93 games. But computer games are fun.

RichFitztightly
01-27-2004, 09:12 PM
Can you get your friend to post on here with the formula he used? I'd be curious as to how he came up with what he did.

Daver
01-27-2004, 09:13 PM
Originally posted by soxfan26
Don't worry about that. I'm not a gambling man, but if I were the last thing I would bet on would be Schoeneweis winning 15 games. For that matter I wouldn't bet on him winning 9, and this club is NOT going to win 91 or 93 games. But computer games are fun.

As it sits right now this team would be lucky to win 80 games.

doctor30th
01-27-2004, 09:14 PM
But computer games are fun

Yes games are fun, but this wasn't from a game, this guys a programing junkie, He's a white as a sheet cause he's always programing something new. He's like a math wiz and he pisses me off cause he can do stuff like this.

soxfan26
01-27-2004, 09:19 PM
Originally posted by doctor30th
Yes games are fun, but this wasn't from a game, this guys a programing junkie, He's a white as a sheet cause he's always programing something new. He's like a math wiz and he pisses me off cause he can do stuff like this.

Before I cast any judgement I would like to see the program used. I don't give "projection" programs like this a whole lot of credit anyway, but this one obviously produced some questionable results.

doctor30th
01-27-2004, 09:20 PM
Can you get your friend to post on here with the formula he used? I'd be curious as to how he came up with what he did.

I can ask him, but it's not really a formula, this was a computer program with alot of lines of code. I don't think he could post it on here but I'll ask him to come on and see if he can spit out what he did in laymans terms. All I know is that the thing used mathmatical formulas for chance an probability, but I can't be sure it's been 6 years since I took a math class.

As it sits right now this team would be lucky to win 80 games.

I agree and as it is in the programming run he did with Rauch I believe they didn't even get 80.

doctor30th
01-27-2004, 09:25 PM
but this one obviously produced some questionable results

All prediction programs produce questionable results. The only non-questionable results are when they actually finish the 2004 season. Like I said he did several runs with several combinations, I just posted the one that had the most wins because I found it interesting, there was one run under 80 wins and the rest landed in the 80-88 win range.

ChiWhiteSox1337
01-27-2004, 09:27 PM
Is he going to do it for offensive stats??

Madbilly
01-27-2004, 09:28 PM
I think Schoeneweis will surprise some people this year. The Angels used him as a 1 or 2 starter when he was in there rotation. While he wasn't a total disaster there, he isn't suited for that. I think he will be more adept to being a 5th starter. Possible he could be a 4th starter if he develops that cut-fastball that KW talked about in one of his interviews (I forget which one)and finds a way to better against righties. I don't think he'll be good for 15 wins, maybe 12-13. In 2002, before he was dropped to the pen in June/July, I think he was 6-6 with a 5 something era. While the era is a little high, I think by July if he's 6-6 for us he wouldn't look to bad. That year the Angels played "small-ball" and didn't have the power the sox have(they had Anderson and Glaus, who isn't the same player he was a few years ago), with a little run support he could be a better pitcher then most give credit for.

Put him at the 5th starter spot, pick up a decent 4th starter (trade/Fa). If Person goes back to his preinjury form give him the 4th spot or maybe Wright if he goes back to 2002 form not 2003. Then let Rauch go to then pen to get some experience. By the break if that doesn't work out flip those two see how that works. Keep Cotts in AAA to work on his control.

doctor30th
01-27-2004, 09:35 PM
Is he going to do it for offensive stats??

I'm pretty sure that he programmed in for offesive stats to get these. But he didn't program in a print function so I had to write them down. After getting a hand cramp from writing down all the pitching stats I found interesting I had to get going so next time I visit him I'll tell him 3 things:

1. Go on to Whitesoxinteractive.com and post about the program
2. do some runs for offensive stats for me.
3. Program a print function so my hand doesn't cramp.

RichFitztightly
01-27-2004, 10:20 PM
I bet Jerry Manuel is sitting at a computer somewhere in Orland saying, "DAMMIT where was this a year ago!?" This sounds like the type of thing that's right up his alley.

:jerry

lowesox
01-27-2004, 10:24 PM
It sounds like a lot of wishful thinking. When I look at Koch's projections, it looks like your friend takes Career numbers into account because Koch rebounds. But when I see Loaiza's numbers it only looks like who used last season as a barometre.

Sorry, but that and Showenweiss' numbers totally discredit this system for me. THanks for posting it though. It's interesting.

Can you ask him what the numbers were based on?

doctor30th
01-27-2004, 11:16 PM
Can you ask him what the numbers were based on?

I know that all season stats were used, and he also took into account fluke season whether good or bad. From what I understand he used a formula to predict the chances of the person repeating a fluke season, or having a fluke season this up coming year. He calculated the odds for different situations, and like I said he ran it several times, this was probably the fluke one because it was the one with this many wins and the the best overall stats for everyone.

I wouldn't say that the system is discredited, I would just say that this run should have been discounted because the median result should have been taken for the answer. Just like every other prediction system it will give you different results everytime you run it, I just thought this one was the most interesting. No one would even want to talk another of the threads where the results were:

Loaiza: 14-8, Buehrle: 13-8, Garland 12-13, Schoeneweis 6-12, Rauch: 7-10.

In that one Koch only had 24 saves.

it's just like any other system it doesn't mean its going to happen, its just a prediction system.