TotalMinors Ratings Discussion

[GForce22]

As a lot of you know, there was recently a...we'll call it a disagreement...about me giving the formulas used for my TotalMinors roster project since I'm not doing it this year. I declined, because I would need to develop new sheets in order to create the formulas appropriate for this year. And, since I'm not doing rosters, I'm not going to develop new sheets for a project I'm not doing.

That being said, I'm certainly open to spelling out what I can or providing what info I can for those interested in figuring out the ratings on their own. If guys tell me what ratings you want, I will try and steer you to the best of my abilities how to get there, at least as close as I can, without my having to develop new sheets.

if RGlass could incorporate formulas into MVPEdit in a much more orderly way than my spreadsheets (fact is, they are poorly designed from an interface standpoint, but they work), guys would be able to blow through ratings development pretty quickly. But I don't know what's on Robert's plate at the moment.

Anyway, ask away and I will give what I can...just understand that, because I'm not developing new sheets this year, there WILL be limits. I just can't devote the time this year to developing this stuff for a project I'm not doing. Hopefully, now that things have cooled down a bit, people understand that. That anyone questioned my willingness to help the community is, quite frankly, both hurtful and laughable, given my history.

GH

[Nomarhits400]

GForce,

Your work- and more specifically dedication- is what impresses me most about your roster sets. You do what you say you will- despite the intimidating size of the task. To make a minor league set is laudable; but to create one that accurately assimilates real attributes and progression is simply amazing.

Having said that, I am wondering if time wouldn't be better spent tweaking last year's set. Certainly new prospects have developed since last year, and some ratings need altered- but I think an overlooked bonus of your set last year was the reworking of the major leaguers- whom EA, in some cases, represented atrociously. I guess I'm just wondering that if there is an offeset in the ratings this year- it perahaps could be accounted for globally- at least to start.

Perhaps that would involve the manual input of ratings for Major Leaguers- so as to avoid entering career stats- but Minor leaguers don't have stat tracking- so this wouldn't be a problem...Importing them might be- but perhaps there is a way around all that.

Perhaps all the minute changes that you feel need to be implemented supercedes any benefits this method might involve- but it's all a moot point anyway if you don't grant permission for a project like this...Just wanted to get your thoughts. I don't think anyone is going to take the time (if you're not- and I have a family- so I don't blame you one bit. Frankly, I am in awe it ever got done) to do it the way it needs to be done and follow through...It's a lot to ask- and no offense- but not all those that are willing are capable

[rglass95]

How would you want the formulas to work? I guess I need a sample formula to see what I am dealing with.

[Goldensuitcase]

Hey Rglass, for some reason I am having problems finding your "How to use MVPedit" instructions. Can you point me in the right direction?

And GeForce, whatever you can offer would be of great assistance and much appreciated.

[qpontiac]

GForce,

You are the man, and yeah it would have been nice to have an MVP roster set by you since I am new to this game, at least I have your latest Franchise Stars Roster set for OOTP 6, and all of your other OOTP roster sets...

Congrats on your family!!! I'm expecting a son in June myself, so my days too are numbered!!!

[poonani]

i definitely would appreciate whatever formulas you did use last year. There are ways to guestimate ratings based on scouting reports as well as statistics (when considering age and level) and compare to the #s given to the prospects already in the game, but it is often a ballparked #

[GForce22]

How would you want the formulas to work? I guess I need a sample formula to see what I am dealing with.

Well, it would need a few things. Basically, if you were to take on adding it to MVPEdit, it would have to be a player generation tab of some sort and a tab/place/whatever you want to call it to store all the league totals and age/league modifiers that my formulas used. It would also need to have a set ratings scale, because my formulas spit out stats, not ratings. I transfer the numbers into ratings.

The data my formulas incorporate include:

-- League totals for every league in professional baseball for the last 3 years.

-- Average age for each league

-- Age modifier (penalty/credit for a player being old/young for the league)

-- MLEs for hitters (Based on minor league park factors, a spreadsheet provided to me by Dan Szymborski from Baseball Prospectus...I have not asked for an updated one, but you can find MLEs in books or use ZIPS projections if you can find them)

-- Pitcher projections (pick your source, Baseball Prospectus, for example, of projections for pitchers...this will be pitchers equivalents of MLEs in their formulas)

-- Player stats for the past 3 seasons (including splits, easily attainable for major leaguers, less so for minor leaguers though Stats Inc. used to include it in their books, which I loved)

Basically, in an ideal situation, the .mbe file would have a tab for all figures that would be referenced regularly (meaning the league totals, age modifer adjustments per league and level adjustments for each league, the ratings scales for categories). Then, each player would have its own generator tab where a user could enter the other data I mentioned for each individual player. Click a generate button and the current ratings for that player would be overwritten by the results of these calculations.

If you don't click the generate button, then obviously the current ratings for that player would hold.

May be explaining myself poorly here, but I'm at work so I'm kind of rushing through it.

GH

[wikitundra]

Can you supply your formulae for these:

Bunting

Steal Aggressiveness

Baserunning

Speed

If you do not wish to post, please email them to wikitundra@yahoo.com Whatever you can supply will be greatly appreciated.

Thanks...

[basballfanatik9]

Hey GForce,

Your rosters last year were awesome, do whatever you want to do. But I would really appreciate it if you did do the rosters, because you're probably the only person on this board who made a Total Minors roster, which to me was AWESOME.

[GForce22]

Hey GForce,

Your rosters last year were awesome, do whatever you want to do. But I would really appreciate it if you did do the rosters, because you're probably the only person on this board who made a Total Minors roster, which to me was AWESOME.

Thanks for the kind words. I absolutely will not be doing a roster, however. I just don't have the time to do it the way I would need to. However, I will try and provide at least some detail as to how I develop ratings. It won't be exact, for reasons I mentioned above, but it'll at least be a guideline so people can do them on their own if they so desire.

GH

[basballfanatik9]

Thanks for the kind words. I absolutely will not be doing a roster, however. I just don't have the time to do it the way I would need to. However, I will try and provide at least some detail as to how I develop ratings. It won't be exact, for reasons I mentioned above, but it'll at least be a guideline so people can do them on their own if they so desire.

GH

OK Guidlines are fine, I will be sure to use them because Snash13 and I are trying to follow in your foot steps to make excellent rosters.

[wikitundra]

Can you provide formula for Fielding and Accuracy?

Thanks...

[GForce22]

Wiki,

I'll get to your requests but it may not be until tomorrow. Accuracy I never really did a formula for, more going on scouting reports. I just don't think it can really be measured formulawise.

I didn't have a bunting formula either, but I should be able to come with one for that pretty easily. Not many variables to work with.

I am actually redeveloping the speed/stealing/baserunning ratings formulas as I do my Franchise Stars roster project for OOTP. It's what I'm right in the middle of now, so I will post those formulas when I'm done.

GH

[GForce22]

RGlass,

Did what I posted in response to you make any sense, or do you need me to detail it a bit better. Just let me know what you would need from me.

GH

[rglass95]

Kind of, but I have no idea what a MLE is.

[GForce22]

MLE is major league equivalent, a sabermetric tool that says how a minor leaguer would have performed last year if he was in the majors. BUt you wouldn't need to dig that up on your own. There would just have to be space for the user to enter the MLE data.

GH

[GForce22]

OK, let's look at speed.

What can we use to measure speed...Steals, certainly. That's kind of a no brainer. Triples are a good one (Doubles are sometimes speed but really hard to discern between what was a gapper and what was a bloop a guy busted *** to make a 2B...triples are almost exclusively speed related.) Double plays. A fast guy doesn't ground into many, usually. Doing my Franchise Stars roster project, I found Mickey Rivers averaged less than 4 a season in his career, which is just plain silly. Boog Powell on the other hand...

What else can we use? I've traditionally used a formula the old Wizard program for APBA Baseball for Windows used to use, or at least a variant, but the more I've used it the more I question it....

It's runs scored/time on base (excluding home runs). The logic being, fast guys make more happen on the bases, are more likely to score on a sac fly, from second on a single or first on a double. They're more likely to be too fast to catch on a fielder's choice, remaining on the basepaths to score. So from those standpoints it makes sense. But it's also very much dependent on the lineup you're in and where you hit in it. Especially in a NL lineup, the 7 or 8 guys just aren't going to be driven in as much regardless of speed because of weaker hitters following them. That said, I've always used it, I'm just not sure I would any longer.

The other reason I do is because, for minor leaguers, grounding into DPs is very hard to come by. And I didn't want to work off just steals and triples to calculate their speed. It's not without it's flaws, but it works more often than not, I think, to paint a good, if not great, picture.

So what are we looking at:

((Runs/(TOB-HR)*100)+(Steals per 162G*2)+(3B per 162G))-(GIDP per 162G)

I think 1 year is fine for speed, but you can use 3 years of data if you think it paints a more complete picture. I generally use 3 for hitting/pitching stats.

Let's spit out a few examples and see what we've got.

Ichiro: ((33.3)+(72)+(5))-6=110.3-6=104 so Ichiro would get 100...fair enough.

How about a guy who runs well, better than average, scouts say, but isn't a blazer. Mets 3B David Wright

Wright: ((53.2)+(28)+(2))-16=67

That's likely a bit low for Wright, hurt by 2 factors a) a tough start to his rookie year, and the Mets being a poor offensive team, which hurt the first number.

These are the things you have to know...NO formula is perfect. None. All you can do is take enough variables into play that you factor in as much as you can. Biggest thing about formulas is to be able to recognize an outlier, someone where the formula just doesn't work for whatever reason.

Ideally, in this speed example, a team's offense should be factored into the equation to compensate. This would raise Ichiro (though he'd still be 100), and Wright, as both teams had offenses that sputtered a bit last year.

Let's drag out a real slow guy...keeping with my Mets, Mike Piazza:

Piazza: (28+0+0)-18=10

Allow for his poor team offense and he'd get a boost, but Piazza would, rightly, probably be about the slowest player you would see.

I'll spit out a couple more of these if you want, but hopefully this is enough to show these variables are a useful way to go about determining speed.

I'll try and figure how to add team offense as a factor to this...shouldn't be hard.

Questions or comments, fire away.

GH

[llcmac]

Holy Crap GH, do you belong to Mensa?

[GForce22]

Holy Crap GH, do you belong to Mensa?

LOL!!! Thanks, but hardly. It's really just math. Find numbers that matter, factor them accordingly and calculate. The key is finding what numbers matter and how to weigh them...for speed, it's relatively easy.

GH

[GForce22]

A quick one:

Starters stamina: IP/GS minus 12.

GH

[qpontiac]

I've always been meaning to ask you on the OOTP forums GForce, but I'll ask here...Have you ever thought about applying at Baseball Prospectus or something?

[GForce22]

I've always been meaning to ask you on the OOTP forums GForce, but I'll ask here...Have you ever thought about applying at Baseball Prospectus or something?

Thanks for the compliment...those guys are well beyond me as far as this sort of stuff goes. I'm interested in sabermetrics, very much so, but those guys, as well as some guys over at OOTP, are beyond me in terms of sabermetric knowledge. I know a bit, but really should know more.

GH

[ZMan]

Maybe, maybe not.

Not all their guys are pure stat heads. Your demonstrated understanding of the minor leagues would be an asset to them, as I've never thought that BP was as strong on their minors analysis as their pro analysis.

[llcmac]

GH, is there any way to figure out ARM STRENGTH? It would be nifty to have a formula.

[GForce22]

Arm strength...there really isn't a formula for it. What I've usually done is list out the ratings the game gives from best to worst and assign scouting report terms (i.e. great, good, average, below average) to different numbers.

Then I took the DP/Arm runs column in TangoTiger's UZR (Ultimate Zone Rating) workup and adjusted accordingly. That number is a +/- single digits figure...this was one of the last things I did last year when I really wanted to work out defensive ratings to make them mean something. The game, and videogames in general, tend to rate defensive numbers too close to each other to really make a difference...the fielders all feel the same. TotalMinors was like that last year, even, until I made this adjustment. Then you could really feel the difference in the players on defense as well as offense.

That said, this was a very imperfect measure...for example, David Eckstein was a -1, and you would expect him to be a negative in any arm-related category. But Jeter was a -3. Jeter has long been vastly overrated as a shortstop, in fact, until last year, he was pretty much the worst by any statistical measure...that said, he certainly has a significantly better arm than Pee Wee Eckstein. Again, NO formula is perfect...you want to come up with something you can trust but know enough to say "hey, that's not right." RECOGNIZE THE OUTLIERS!!! The key to any formula for roster development or, pretty much, for anything.

Unfortunately, a quick search has not yielded a sheet for this year. Unlike Dan Szymborksi and others at Baseball Prospectus and Baseball Primer, I have not traded e-mails with Tango, so I can't just get the info from him. I believe he's working on a book or something actually of the sabermetric variety.

A long-winded answer to a simple question

GH