Without merit there should be no reward.
—Ancient Chinese proverb, huh?
A man’s true merit ‘tis not hard to find;
But each man’s secret standard in his mind,
That casting-weight pride adds to emptiness,
This, who can gratify, for who can guess?
In our glimpse into the major baseball awards, we have already looked at the worst players ever to receive any support and the best candidates who were completely overlooked in the voting. So what does it all mean? Is voting improving with all our fancy ciphering today or are the Mo Vaughn and Pete Vuckovich-type results the rulr rather than the exception?
In the final leg of this bloated study will take a stab at grading the voters. How well do there votes relate to reality?
Using Win Shares as the standard for player performance throughout baseball history (I know that it has certain eccentricities but it's the best we have), we will compare the votes apportioned to the players and determine how well the two correlate. The better the correlation is the better the voting reflects reality, at least that's the assumption.
In order to normalize the voting across eras with different voting rules, I have assigned each player an award share based on the points allotted divided by the maximum possible for points. I believe this is the system used by Baseball-Reference.com for their award shares.
Keep in mind that even though the results fit in a small table, the query was a bear that nearly chewed up and spat out my Access database. It got even worse in the second phase, that I'll discuss in a minute, with two main queries wed together with multiple subqueries for conditions. Anyway, that's neither here nor there: I just needed someone with whom to commiserate. Oh, and it also explains the delay in presenting the results—like their releasing the Hall of Fame ballot wasn't enough.
So here goes. First, let's look at the MVPs. Here are correlation coefficients per decade for all vote-getting player population. This reflects just those players who received at least one vote in a given year, broken down by decade. It attempts to test how well award shares correlate to Win Shares for each player-year:
The first thing I notice is that the voting doesn't correlate well to Win Shares at all.
Also, it seems that since the Fifties the voting had been devolving into an even paler reflection of actual performance. Ergo Mo Vaughn and Juan Gonzalez. However, that is until this decade. So far in the current decade (whatever we call it), the voting has been the closest to reality in MVP voting history.
So have the voters wised up finally? Are they listening to the sage advice of the sabermetric scolars?
Well, there's one problem with this approach, however. It ignores worthy players who were overlooked in the voting and just looks at vote-getters and how well their votes were meted out.
Let's broaden the scope to look at all players with greater than zero Win Shares (this is where the query gets hairy). How well does the correlation hold up now?
Vote Getters Coeff
This tells a completely different story. The voters seemed to turn away from performance as the main criterion as baseball expanded in the Sixties and have yet to go back. Even if they have suddenly improved in ordering the players they do deem vote-worthy, they still ignore worthy players the same way that they had for the previous three decades.
But why? It seems that the voters are doing a better job of seeding the candidates by actual performance, but there was a change earlier on that is still the more influential. It seems that as baseball expanded, the voters changed their way of thinking about the MVP award. It wasn't about the best players anymore. It was about the "most valuable" players on the winning clubs, and they took most valuable to mean whatever they wanted. This devotion to players on winning teams caused more worthy candidates to be completely ignored, and it continues until today.
So the writers may be getting better at evaluating the players that they will consider for the award. However, they still collectively turn their backs on equally worthy candidates because they are not on contending teams.
I guess there are worse thing (like giving Mo Vaughn the award). At least they are getting it right for the players they review. It may just mean that empty votes are thrown at Chone Figgins rather than to a super-sub on a non-contending team (Sorry, Melvin Mora).
Next we'll look at the Cy Young award. Even though it's supposed to go to the best candidate, does it have a similar "contender's only" bias? We shall see.