In my previous post regarding Alex Rodriguez, I used Nate Silver’s model
presented in Baseball Between the Numbers
to evaluate how much Rodriguez is worth. After reading complaints about Jimmy
Rollins’ NL MVP award and annual complaints that only players on winning teams
have a shot, I thought that I should try using that model to determine the
value of all those that received votes in the MVP voting. So that’s what I’ve
done. First, we’ll look at the AL MVP award, where I thought A-Rod should have
been a unanimous selection. Since I’m using a model based on historical data,
there are downfalls in this analysis, so here’s a list (most likely not a
complete list): the player’s replacement would produce at exactly replacement
level, the player has no value beyond what he produces on the field, no
consideration of market size, no consideration of actual year-end standings for
determining playoff teams (Cubs get hurt due to poor record as division winner,
Tigers, Mariners, Mets, and Padres get helped due to good record despite not
making the playoffs), and dependence on player’s teammates in order to field a
playoff caliber team. So here are the AL MVP vote-getters:
From this table, it is easily seen that Carlos Pena, Torii Hunter, Justin
Morneau, Chone Figgins, and Frank Thomas were not worthy of an MVP vote. (I
actually had estimate for Carlos Pena because the graph doesn’t go that low in
wins, so I assumed that a simple linear interpolation would do.) Other than
Figgins (114 games), these players were on teams that weren’t really
contenders.
Looking at the other end of the spectrum, the actual AL MVP came out on
top, but he didn’t finish much higher than John Lackey and Fausto Carmona. I
guess the real question is why did Lackey and Carmona finish no higher than 18th
in the voting?
On to the NL:
There are plenty more names that show up as not worthy of votes on the NL
side, but that’s mostly due to the league’s (and mostly the Central’s)
mediocrity. I won’t list them all, but there are 11 players in this list under
$17 million. After their removal, there are 15 players left.
Matt Holliday, Brandon Webb, and Troy Tulowitzki top the list, but I was
surprised to see Jimmy Rollins that close behind them. Rollins is tied for
fourth with Eric Byrnes (yet another surprise) at $23 million, only $3 million
behind Holliday.
In the end, the fact that Alex Rodriguez was not a unanimous AL MVP
selection and the fact that Jimmy Rollins was the NL MVP selection isn’t as bad
as I originally thought. While this process determines just how much a player
was worth in terms of getting his team over the hump and into the playoffs,
does it really make sense for Albert Pujols to not be the MVP because his
teammates weren’t good enough? Another question: aren’t the “real” most
valuable players those that are producing more than they’re getting paid? If
so, why don’t the writers consider how much each player is being paid?
I think you have the right article, but the wrong audience. The BBWAA awards the MVP to the most outstanding player, regardless of salary. In a perfect world, they would adhere to value in its truest sense - which is really to say the highest run contribution relative to the team's performance. So let's get them using VORP or WARP before we even worry about salary...
If you're audience is a GM, then they'd eat this stuff up.
Oh, and perhaps I've misread your intent, but there is a flaw. The $ value of those wins are nominal values. They should be expressed in real terms, i.e. percntage of the value of all the team's wins. A-Rod's 31 million is nominally the highest, but how does his percentage of the Yankees billion dollar revenue compare to Carmona's percentage of Cleveland's revenue?
I wasn't considering my audience to be the BBWAA. I just thought it would be an interesting way to look at the MVP race, and since I had most of the work done from the A-Rod article, I thought I'd put the pieces together to make it happen.
One thing that it did make me realize is that maybe the writers are on to something when they only consider those players on winning teams.
$34 million is more than $31 million no matter where you live, so I'm not sure why revenue percentage would be better than nominal values.
As for using the model to do so, whichever player's team sits closer to the sweet spot of Nate Silver's graph in Baseball Between the Numbers would end up with the larger percentage. In this case, Rodriguez sits closer to the sweet spot, so he beats Carmona 43% to 38%. Obviously, these percentages aren't correct either because the model isn't linear. If you were to do this for an entire team, how would you determine which players get the big-time revenue from the sweet spot of the graph and which players get the modest revenue from the ends?
Using your model to evaluate VALUE, percentage of revenue contribution is relevant. $34 million (A-Rod) in revenue is more than $32 million (Lackey), but $32 million is more valuable to the Angels than $34 million is to the Yankees. There are different economic ballparks (figurative) in which these teams play.
This is baseball. Let's have some fun.
Recommended Websites:
MLB Trade Rumors, Baseball Prospectus, Cot's Baseball Contracts, Fan Graphs, Football Outsiders
