2017 MVP

A forum for introductions and casual chat about the baseball simulation hobby. Who are you? How do you like to use baseball simulations? Do you like league play, historical replays, drafting, or library construction? Come on in and join the conversation!

2017 MVP

Postby johnnybravo17 » Mon Oct 02, 2017 6:40 pm

This place has been eerily quite of late ... maybe this will help:

A couple of years ago, while developing DDBBXSV, I was seeking an "MVP Predictor", something similar to ESPN's Cy Young Predictor: http://scorecenter.espn.go.com/mlb/features/cyyoung

I'll let those numbers speak for themselves, though I'd be shocked if Jansen actually won the award. Early in the season it looked like Scherzer was a lock, but then he slipped behind Kershaw, but then Kershaw got hurt, but then Scherzer stopped being superhuman and got hurt also ... and then neither of them impressed much down the stretch when they could have put it away. I still think with the actual voters, however, that it will come down to one of those two.

Anyway, back to MVP ...

I soon discovered that there wasn't an "MVP Predictor", so I set about to create one. I finally did so (using linear weights from the performances of all players receiving MVP votes from the years 1960-2014, and then multiplying by a "scaling factor" based on the primary fielding positions of all players receiving votes). I was actually pretty happy with the results. The raw 2015 formula became the basis for the MVP system in DDBBXSV, and also correctly predicted Josh Donaldson's win that season over Mike Trout. To be fair, it also predicted (by 4 points) that Paul Goldschmidt would edge out Bryce Harper ... which, of course, didn't happen.

Within DDBBXSV, that base formula is far more complex now. The scaling factors have been modified slightly, and the program also takes advantage of a lot of other information that you just don't have when looking at raw season totals. Things like, performance down the stretch, playing well in games your team actually wins, etc. etc. all play a factor within DDBBXSV. It's actually quite sophisticated, and very involved. A lot of work went into it, and I think it's my favorite part of the program (except possibly the All Star Voting, that's pretty cool too!) As I watch DDBB seasons unfold (and therefore have a feel for the top players and what they've meant to their teams) it seems to always choose the "right" player ... regardless of what the raw numbers may be telling you.

But this season, "The Year of the Home Run", we once again have lots of strong offensive seasons to choose from in what figures to be a close vote in both leagues. So I thought I'd dust off that original formula, insert the statistical totals from this season's top candidates, and see where that gets us. Previously I included pitchers, as the system accounts for and includes them, but I don't think we have any pitchers in either league that are serious MVP candidates this season, so my criteria for making this particular list were the top 10 Offensive WAR finishers in each league (according to Baseball Reference). It was actually necessary to do so to include Goldschmidt, who did not make the NL's top 10 in total WAR ... despite having a fine season and arguably being a strong candidate.

First the AL:


This was Altuve's award to lose all season long, and with a slight downturn in the season's final month, coupled with a strong finish by Judge, the predictor seems to think he's done so. Altuve actually gets a boost here for playing second base, but it's still not quite enough. It will be interesting to see what the actual vote looks like. Dozier's finish so high on this list surprised me. But he benefits from being a second baseman also, as well as for getting plenty of AB's in, and making at least a reasonable showing in every stat category. Lindor and Ramirez had strong seasons also, and the trio of Trout, Springer and Correa were all hurt by missing significant time due to injury. Trout, in particular, could have made things very interesting had he stayed healthy for the entire season. That guy is truly amazing!

And finally, the crazy, and guaranteed to be controversial NL race, with at least 5 very strong candidates:


The predictor likes Blackmon in a squeaker over Stanton. If I just look at the numbers (and Blackmon has several other impressive ones that aren't included here) I'd probably give it to him. But the predictor doesn't know that Blackmon was absolutely Ruthian (.391/.466/.773) in Coors Field, and rather ordinary (.276/.337/.447) away from home. The voters do, however, and that's a huge difference to ignore. I'm guessing that Blackmon will not win the award. That still leaves us with Stanton, Votto, Goldschmidt, and Arenado (whose home/road splits are much more balanced). I guarantee that all five of them (including Blackmon) will get at least one first place vote. And had he gotten the elusive 60th HR, I think Stanton would have been a lock. As it is, I'll stick by the predictor's "Coors adjusted" findings and say that Stanton wins the award anyway. Votto has amazing numbers, but Votto always has amazing numbers and hasn't fared appreciably well in the voting since winning in 2010. It's hard to see him winning now over a guy with 59 HR. Goldschmidt slumped a bit down the stretch, and arguably wasn't even his team's MVP when it really mattered (that would be JD Martinez). And Arenado still has Coors working against him, as there are those who won't dig deep enough to actually look at his Home/Road splits and just assume that the stadium inflated his numbers the way they did Blackmon's. It should be a very interesting and close vote, however ... no matter which way it turns out.

By the way ... the "MVP" column denotes "MVP Points", used to predict the winner. These were purposely created to be on the same scale as the "Cy Young Points" that ESPN's predictor uses. I did this by considering actual MVP votes earned by pitchers in the 1960-2014 period, and then using their Cy Young Points as the baseline for the MVP Points scale. The express reason for doing so was to create a system where pitchers could be considered for the award also. So, for sake of argument, you could say that top finisher Corey Kluber would land between Ramirez and Upton in the AL, and that top finisher Jansen would land between Rendon and Pham in the NL. I actually think that's about right. A couple of fine seasons, to be sure, but no serious MVP candidates among the pitchers in either league.
User avatar
Posts: 849
Joined: Thu Apr 04, 2013 9:54 am
Location: Somewhere on the outskirts of sanity, and perhaps in your head, as the unofficial 'voice' of DDBB

Return to Casual Chat About Baseball Sims

Who is online

Users browsing this forum: No registered users and 1 guest