Where do player stats come from ?

[nightwulf]

Here's something puzzling me. The batting / base running abilities for every non-pitcher in RBI are determined by their contact, power, and speed. So how are these numbers determined?

For example, Vince Coleman is the quickest baserunner with a speed of 148. But where did this number (with respect to other players' numbers) come from? Coleman stole 109 bases in '87. Ozzie Smith stole 43, Tom Herr 19, and Jack Clark a whopping 1. Their speeds (as defined in RBI) are 148, 144, 136, and 122. The next-quickest baserunners are Davey Lopes and Tim Raines, both with speeds of 146. In '87, Lopes stole 2 bases, and Raines 50. Obviously the speed rating for a player isn't determined solely by their stolen bases. So what other stats figure into a player's speed? Somehow I don't think the players visited the Tengen office and were timed by programmers.

You could look at their slugging percentage. Quicker runners can stretch hits into extra bases while slower runners will have to stay on first. The same logic could be used with a player's doubles and triples. You might also use their on-base percentage (with respect to at bats). But those are all very subjective tests.

And if that boggles your mind (like it does mine), that's just a player's speed. We know exactly what the speed rating does; it determines how quickly they run the bases. Where do contact and power come from? We don't even know exactly what effect those numbers have, other than that a higher number is better.

Pitcher stats are even worse. Each pitcher has a different speed number for sinkers, curveballs, and fastballs. They also have numbers that control how "effective" their sinkers and curves are. Where the hell do these numbers come from? The MLB site lists all kinds of stats for pitchers. Hell, I don't even know what some of them are. But there certainly aren't stats that relate directly to a pitcher's throwing speed, especially for certain types of pitches. Then again, I can't imagine the RBI programmers sitting in a room with a pitcher and clocking sample pitches.

What I'm getting at here, is that even with RBI Manager or a hex editor, I find it difficult to put different players in the game with realistic stats. I wish there was some program which would take a player's stats for a year, run them through some complicated algorithm, and spit out realistic numbers for use with RBI Baseball. I'd love to write a program that did this, but I don't even know where to start.

Any ideas?

Nightwulf

[vgp100]

Sadly, it probably is a bunch of guys in a room saying--oh yeah  he's real fast. or --Of course Steve Bedrosian pitches submarine (sidearm). They probably give the benefit of the doubt to favorite players.

[Gwynn3k]

i would have to agree with vgp... i dont think the guys in japan spent an inordinate amount of time researching the teams... how else can you explain the astros...   must be cause they like bright colors.  

the only thing the numbers did for me is they proved what i had been stating all along...  don baylor is not fast even though he's listed as 'A' speed.

[Gantry]

Where do contact and power come from? We don't even know exactly what effect those numbers have, other than that a higher number is better.

Quick correction - a lower contact rating is supposedly preferred.  That being said, I think the contact rating has very little effect on a hitters effectiveness.    All power baby...

As for where they came from, I think the programmers studied the stats and made some quick judgements.  I'm sure they didn't have a pretty algorithm based on stats....

[sucka free]

Nightwolf,

    I am somewhat of a baseball stat guru so I'll try to shed some lighyt as to what happened.  Although you bring up some good points, I don't necessarily think that the stats are that far off.  First lets look at Raines, coleman, smith.  Obvisouly coleman was a great base stealer.  The best that year in shear volume.  He was also in a better position to steal becasue he was a leadoff batter for a 1st place team, causing him to get a lot of at bats and it gave his team alot of "incentive" to run to get into scoring position.  He also took opportunites away smith, who, when on base would be in A LOT less stealing opportunities.  Tim raines, who played on the third place expos was also had fewer opportunties to steal the colman.  Raines's team scored fewer runs than the cardinals  (798 to 714) and the expos weren't the running team the cardnials were that year.  So far, all three were base stealers but colman had more incentive and opportunites to run.
    Now lets look at stolen base percentage....
colman  109 out of 131 for 83%
raines     50 out of   55 for 91%
smith      43 out of    52 for 82%

Based on percentages Raines was the better base stealer, but since colman stole twice as many bases, colman gets to be the fastest in the game..but not by much.  Since ozzie smith was close to colman in percentage smith is also close to colman in the RBI game.  I conclude that for These Three players, RBI did i good job in reflecting their running standing in the game by factoring in sb percentage, total stoolen bases, and SB incentive.

As for lopes, herr, and clark I really have no clue.  My best guess would be to say that in order to balance the teams out certain players needed "a boost" to make their respetive team more competative.  One team needs top be faster, one needs the long ball, one needs better pitching, ect.  I bet they had some scale in which they ranked ecah teams abiltiy as far as speed, power, pitching, ect and averaged them out.  Maybe the cardnial speed needed to match bostons power so they made the cardnials a faster team by boosting herr and clark.

You were wrong about slg% and on base %.  Both are USELESS when judging a players speed.  Homerun hitting is what really boosting slg% and speed has nothing to do with homeruns (see babe ruth's stats for a prime example).  Homerun hitters get walked a lot more becasue of their power, not their speed.   Walks are the greatest enhancement to on-base percentage (see bonds, ruth, mcgwire).  Power hitters always have more doubles than a speedy guy becasue it is easy to hit a ball in the gap than to strecth a single into a double (jeff kent).  The only offense stat to measures speed besides stolen bases as the triple.  Maybe RBI used triples as a basis for speed?

Now on to pitching.  Pitching speed ranges are given to EVERY pitcher for EVERY pitch they throw.  Its all recorded and averaged out then complied in a scouting report.,  Although these stats won't show up in the paper every team has these stats on every pitch every pitcher throws.  No bullshit.  ANd the scouting reports aren't hard to get espically if your making a game promoting that teams players.  What probably happened was that MLB or each repsective team sent tengen a scouting report of their pitchers which included the speed range for every pitch their pitchers threw.  Then tengen created a scale to match MLB's averages.  But trust me, there are stats for every thrown ball.  curve/slider speeds are calculated not only everytime they are thrown, but also in every count.  Pretty crazy.  Example, Zito will throw three types of curves ranging in three types of speeds for various situations.  All of which are recorded in a report.

To end, i think it is possible to create a realistic game that is reflexive of MLB.  I think in large part RBI did a good job but had to create some ridiculous changes (don baylors speed) in order to balalnce out the game play based on some scale that they created.

let me know what you think.  Sorry for my long windedness and grammer.  I don't proof read.  its a problem