OSFM ST

[VeranKai]

Wow, first of all thank you sooo much to everyone in the community who puts in so many hours each year to help make the games we love even better. I've been messing around with rosters more than ever this year, and the O.S.F.M. group gives such a great base for us to work with. Thanks again guys.

So I've lurked on these forums for many years but am just starting to give some input. The first place I wanted to start was with player ratings - and pitching specifically. Now there's a lot of great work already in this regard, Jimbo's Caluculators certainly come to mind as the most prevelant example. So I'm not here to say mine are better or worse. They're just different. Trying to quantify a human into a small set of calculated ratings is no easy task. There's certainly going to be a lot of subjective bias somewhere in the process (which is okay!)

Anyway, to keep making a short story long, I posted a set the other day in the vault called OSFM ST. The ST is kind of purposely vague, though Steamer and Spring Training are both very applicable in a lot of ways. The ratings in that set are based on Steamer(U) projections of the same date. There are a couple of lineup adjustments also to try and fit in accordance with the projected rosters before Spring Training started - but I openly admit I'm not the most knowledgeable in that regard and it's far from perfect right now.

Since downloading a roster and looking through all the ratings isn't an easy way to get a good overview, I'm posting a link to a pdf so you can see the ratings I've input:

My 2014 Steamer(U) Pitcher Ratings

I have done some batting calculations too, but between all of this and also offering my ratings for JSeward92's 2003 roster that's a lot of hours so far.

There are a ton of roster projects on the forums already. Even the split of historic vs current is large. Despite that though, I really wanted to give one more option to the community because options are great. If anyone is interested in my ratings for their team or season feel free to ask. If anyone is interested in collaborating with me to take these farther in accuracy (batting, fielding, transactions/lineups/rotations) I'm game for that too. I'm really looking forward to seeing the OSFM Hybrid rosters very soon. Since some like injury-free rosters, some like to replicate every day-to-day transaction manually, and everywhere inbetween, if I do keep working on my 2014 roster it will be for a little bit different of a crowd than the Hybrid set.

Oh, and I'll throw up a OSFM ST v2 set now that has a couple of seanjeezy's awesome pitch edits for 2013. I'm definitely really looking forward to his new data for this year.

[VeranKai]

Re: OSFM ST

Well OSFM ST v4 is uploaded now. A lot of pitchers have been rerated per more recent Steamer(U) projections. Even more importantly, a lot more pitch edits have been implemented. As anyone working on The Show rosters will tell you, it's seldom a good idea to base your judgements strictly off of an overall rating. That said, as I keep putting in more pitch repertoire edits, I get more and more pleased with the overall ratings.

There have been a handful of small position player adjustments too. A lot more injuries have been removed and rotations and the like a little closer to an 'ideal present day'. Once again, the general idea with these rosters is to play from the start of spring training with the roster outlook each team had then, but with ratings that reflect how the players are performing this year. (Edit: kind of a blend between simulation (the ratings) and fantasy (injuries and lineups/rotations more variable per the game settings and human decisions)).

[VeranKai]

Re: OSFM ST

Okay, I put up a v5 with a lot more pitch edits in it. I must say I'm running out of steam though as I've spent countless hours on this project and haven't been able to get any feedback at all.

As a perfectionist, I hate having the roster so far from perfect, but I'm glad other projects are still going strong.

[BegBy]

Re: OSFM ST

I'll grab these later and try them tomorrow. I'll let you know what I think since you're looking for feedback.

Cheers!

[VeranKai]

Re: OSFM ST

Thanks! My H/9 ratings for pitchers are definitely about 10 points too high I think. These struggle to get enough guys over .300 for batting average and 100 for RBIs. Going to work on that much more at least.

[rjackson]

Re: OSFM ST

Actually, this sounds like a darn good approach to a roster and is the first I've noticed it. How did you handle pre-existing injuries before the start of ST? Guys like Cory Leubke who still aren't back from TJ...

I haven't really looked at Jimbo's calculator. Was that the basis that you used? Curious on what you used for ratings such as vision, power, contact...and fielding ratings as well.

[VeranKai]

Re: OSFM ST

I pretty much made up my own formulas for everything. And while I do have batting and fielding formulas, I haven't input many of those at all just due to the sheer amount of work it is. I went towards pitching first [so that's gotten the vast majority of my work]. For MLB players though I use custom reports from FanGraphs to compile data. For MiLB, which I've only done for the 2003 roster so far, it's been Baseball-Reference to get statistics. So for vision I use K%, for discipline BB%, et al. If you're curious as to the exact way I calculate something, I don't mind divulging, but they can seem crazy at first glance lol. For instance, on vision, my formula is 104-(1000*(K%^2)). Oddly enough, for pitcher clutch ratings I was going in-depth with a compilation of LOB% and RISP ERA and the like, but I found a formula with straight-up ERA numbers matched what SCEA seemed to be trying to do the most.

For the Spring Training rosters, I openly admit this part is really lacking in my roster. I've tried to put in some good effort there, but that's a lot of research and manpower swapping around rosters/lineups/rotations too. Mostly, if they didn't injured until after Spring Training started, or have return from a preseason injury already, I've moved them into the MLB rosters. Guys who aren't back are still as Class-A per the O.S.F.M. Opening Day roster; and at the least a guy suspended the full season was removed from the 40-man roster.

I should note that the lineups/rotations listed are in many cases more akin to how they are now. From watching the forums, it seems most casual players just want to use the best players on the team regardless of if John Doe missed 3 weeks for a pulled muscle. Hardcore simulators on the other hand, they'll put in a lot of work replicating day-to-day transactions no matter their base, so I suppose I've catered to them a tad less apart from the ratings adjustments. Then you throw on that with the CPU controlling teams in franchise they do all kinds of crazy roster/lineup moves, being too much of a stickler there can feel futile at times.

[VeranKai]

Re: OSFM ST

Good news! Adjusting my H/9 formula did seem to alleviate my batting average issues. So I went through really fast and put up a OSFM ST v6 with the fix.

[rjackson]

Re: OSFM ST

This is something I have wondered as it is pretty standardized to use K% for vision but I wonder how much of a difference it would make to use contact % instead? It speaks more to the intent of the ingame rating. Contact is another one--LD% or maybe BABIP would be more accurate than AVG, but there is a relationship there between all three. Plus, does steamer even offer these numbers? I have been leery of using rosters adjusting the SCEA players for these reasons but I just got home so I will take a peek at what you have so far.

[VeranKai]

Re: OSFM ST

Contact % isn't listed in the lump projections - not sure if I could get for each individual on their own page (tons of extra time), but you can definitely get it easily for previous years. I toyed with averaging ratings from K% and contact % at first.

There's always a tradeoff though. Some statistics are much harder to get quickly; and some don't seem to translate into the game the way SCEA has their ratings. This is my first year doing this, but as I've worked on each formula I try to get my result spreads to match the ratings spreads similarly to SCEA.

It's far from perfect. I mean narrowing a person down to a few in-game ratings to replicate them is no easy feat. You're going to have bias and error no matter who is calculating and from what base of information. With my pitching I've relied on Steamer(U) so they moreso reflect what's happened this year. With the batting changes I've done, I've had to rely more on previous season results. Considering how much flux you can get from guys having great years or prolonged slumps you'll be hard pressed to find two people who totally agree on a rating.

Unlike some others I guess, I'm not against the default SCEA ratings. I just always like doing my own too and give people another option.

Oh, and as you'll see in other threads too, if a roster messes with the rating spreads too much then future draft prospects will come in and be out of whack comparatively. So is it worth it to come up with a new rating setup that doesn't disrupt the metagame too much? At this point I'm not sure lol. If a few people like the approach though then it's all worth it.

[VeranKai]

Re: OSFM ST

If you are a Padres fan, Cashner, Kennedy and I think Ross have some slight improvements according to the latest data I'm doing right now even. I've put up so many versions already I know I need to tone down on that though. Mostly I was just trying to get some feedback and when you're doing all this alone, after enough hours you just want people to have the better version.

[rjackson]

Re: OSFM ST

I only found one version in the vault. A cursory look really showed nothing glaringly wrong. I didn't think about the incoming prospects and that is a good point. What makes it hard is that thia game diws such a good job of emulating up and down years, so testing becomes an adventure. I'll have to go back in and check out Cueto--I am sure you hit him up already.

[VeranKai]

Re: OSFM ST

Yeah I deleted all the previous versions this afternoon after figuring out my mistake in H/9 that was impacting sim results.

Even if you dislike my rosters, I highly encourage checking out any roster with pitch edits. Those guys put in a ton of work and I think it's pretty fun to have (they go through pitch repertoires for everyone). Most of what I've input is based on 2013 data, not Cueto's favorite I'm sure, so I'm hoping with 2014 data he'll get a little more control on a pitch or two ;-)

[VeranKai]

Re: OSFM ST

I pretty much just worked on the rosters all day, and yet another version is up (v7)...

Mostly it addressed the pitching aces and fixed some of the H/9 ratings on a few minor leaguers.