Jump to content

MVP '99


MetsReyes777

Recommended Posts

  • Replies 743
  • Created
  • Last Reply
  • 7 months later...

Sounds good. Thank god for the Lahman database. I'm not quite sure what you mean though, baseball-reference has every 25-man roster since forever. Go look. Keep it up man, I'm gonna continue the uniforms shortly.

baseball-reference.com actually lists more than 25 players on a roster. For example, here's what it lists for the '99 Orioles:

C Charles Johnson

1B Jeff Conine

2B Delino DeShields

3B Cal Ripken

SS Mike Bordick

LF B.J. Surhoff

CF Brady Anderson

RF Albert Belle

DH Harold Baines

Will Clark

Jerry Hairston

Jeff Reboulet

Rich Amaral

Ryan Minor

Mike Figga

Gene Kingsale

Willis Otanez

Derrick May

Calvin Pickering

Lenny Webster

Jesse Garcia

Tommy Davis

Pitchers:

SP Scott Erickson

SP Sidney Ponson

SP Mike Mussina

SP Juan Guzman

SP Jason Johnson

SP Doug Linton

CL Mike Timlin

RP Jesse Orosco

RP Arthur Rhodes

RP Scott Kamieniecki

RP Ricky Bones

Doug Johns

Mike Fetters

Al Reyes

Gabe Molina

Rocky Coppinger

B.J. Ryan

Jim Corsi

Matt Riley

Heathcliff Slocumb

Brian Falkenborg

That's a lot more than 25 players. What I've done in cases is like this is make sure I have enough pitchers, and then start removing players based upon the number of games played. A number of guys only have 5 or 6 games played, so they'd be the first ones to be dropped.

Another thing to be aware of for your rosters is trades made during the season. A player may show up on multiple team rosters during that year. In that case, I put him on the team where he played the most games.

What do you do in september when rosters expand to 40?

Link to comment
Share on other sites

If the issue is rosters, I may be able to help. I've been attempting to collect and index minor league stats from baseball-reference, and it's a matter of time until I get to the 90s (I've been waiting for their new site to finish, and other projects to let up).

When I get the 1996-1999 stats, I'll run MLEs on them, and try to 'Marcel' it up to make 1999 'projections' (possibly with some actual 1999 stats factored in.

Also, Baseball America's prospect lists just happened to start in 1999 (http://web.archive.org/web/20000304110753/www.baseballamerica.com/features/top100/index.html). In retrospect, they were pretty good.

I might not have it posted right away, but given the timeline, I doubt that'll matter. Just let me know if you think I'd just be wasting my time.

Also, opening day rosters can be found at: http://sanpurdue.freesitespace.net/Goodies.htm

In this case, the Orioles opening day roster was: R.Amaral, B.Anderson, H.Baines, A.Belle, R.Bones, M.Bordick, W.Clark, J.Conine, S.Erickson, M.Fetters, J.Garcia, J.Guzman, D.Johns, C.Johnson, M.Mussina, J.Orosco, W.Otanez, S.Ponson, J.Reboulet, A.Rhodes, C.Ripken, H.Slocumb, B.Surhoff, M.Timlin, L.Webster

Link to comment
Share on other sites

If the issue is rosters, I may be able to help. I've been attempting to collect and index minor league stats from baseball-reference, and it's a matter of time until I get to the 90s (I've been waiting for their new site to finish, and other projects to let up).

When I get the 1996-1999 stats, I'll run MLEs on them, and try to 'Marcel' it up to make 1999 'projections' (possibly with some actual 1999 stats factored in.

Also, Baseball America's prospect lists just happened to start in 1999 (http://web.archive.org/web/20000304110753/www.baseballamerica.com/features/top100/index.html). In retrospect, they were pretty good.

I might not have it posted right away, but given the timeline, I doubt that'll matter. Just let me know if you think I'd just be wasting my time.

Also, opening day rosters can be found at: http://sanpurdue.freesitespace.net/Goodies.htm

In this case, the Orioles opening day roster was: R.Amaral, B.Anderson, H.Baines, A.Belle, R.Bones, M.Bordick, W.Clark, J.Conine, S.Erickson, M.Fetters, J.Garcia, J.Guzman, D.Johns, C.Johnson, M.Mussina, J.Orosco, W.Otanez, S.Ponson, J.Reboulet, A.Rhodes, C.Ripken, H.Slocumb, B.Surhoff, M.Timlin, L.Webster

what new site?

Link to comment
Share on other sites

The Baseball Cube is a great resource for minor league stats, but it isn't practical to make an entire league at a time.

Baseball-reference's new site is still beta, but it's pretty functional: http://sandbox.baseball-reference.com/

I wrote a leech script a few months ago to get stats from their old minor league site (http://minors.baseball-reference.com/), but I've been waiting for their new site to finalize before converting it over.

The plan is to leech all relevant stats from the 1996-1999 range, pass them through MLEs (though minor league ballpark factors are unavailable, so I'll need to make some estimations), and long story short, come up with a spreadsheet with 3000+ players with relevant stats. From there, it should be easy enough to chain them in your (apparently closed source) ratings formulas, and get the majority of the work done in one fell swoop.

Using this technique, we can arguably make a total conversion mod for any season, as long as we can get MLE multipliers for each league in question. As usual, the toughest part will be fielding/pitch ratings, but those should be available for anyone with major league experience (TotalZone is available for all major league players from the Retrosheet era), and between the Neyer/James book on pitchers and the Mogul database, we should cover the pitcher ratings.

We can then try to run regressions to generate the other data, though anyone who never sees the major leagues is usually not important, and anyone who is can have the one-on-one research done as normal.

I'll try to get a sample up sometime this week, and if it's just optimizing old techniques, I can write scripts to do so.

Link to comment
Share on other sites

are they ever gonna have splits pre-1956?

I'd like to see some minor league splits for the late 80's-early 90's.

To get splits, you either need box scores for every game in a season, or ideally, full play by play events for every game in a season. For the major leagues, this is generally a thing of research. For minor leagues, these may never have been recorded, so it may not be possible.

But just 10 years ago, splits as we know them pretty much didn't exist. At best you had some home/road splits, and a HR log. That was generally it. Thankfully, most of the splits people are looking for don't give a lot of useful information, so it isn't a big loss. We'll have all of them for major leaguers anyway.

I can call up the major league splits for the seasons in question, and we can look at career splits if we need more info for those who weren't established major leaguers yet. The fun part of this kind of replay is that players who never made the majors are generally unimportant, so they don't need as much detail. Of course, I'll use as much info as possible.

Link to comment
Share on other sites

i like the splits they have already. they're very useful when trying to determine attributes. but the classic guys are tough to do since i can't get splits before 1956. i just assumed they existed because i've seen them in Baseball Mogul, but there's no other place i've seen them.

Link to comment
Share on other sites

No, if a split is needed, and isn't available, it's generally estimated. Mogul just does that in the background.

Generally speaking, most splits have no actual skill behind them and are just random deviation. The important splits are home/road (for park effects), and vsL/vsR (though very few players deviate from what their handedness suggests).

The major league data will be park adjusted, but minor league data might need a proxy. I can generate a listing of 1996-1999 vsL/vsR splits for batters, and that'll explain 80% of the data we need.

I assume there are no splits other than vsL/vsR, are there any other splits needed? It also appears that pitchers don't use them, which is a shame, since they show more significant platoon splits than hitters do.

Link to comment
Share on other sites

Platoon splits are a skill a batter can control, but if the information's not available, then using the league average platoon split is the best strategy. If you have several years of split info, it's still better to regress it towards the league average, since a lot of it will be random variance.

That's why the ratings will be based on 1996-1999. I'll Marcel together 96-98 to get the ratings, then I'll cheat and force in some 1999 info to "improve" the "projection". This'll inherently add regression to the mix to minimize the effect of luck on the ratings.

Either way, I'll try to work the spreadsheet one part at a time, when I have time. The major league stats are usually easy enough to do, I'll get the splits ready, and the long bit will be the minor league numbers.

However, all I can do is the numbers, if there's anything based on numbers, I can probably help automate it, but the visual stuff, I can't help with.

If this technique works, it could simplify the rating generation of future projects.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.


×
×
  • Create New...