Jose Altuve and the Dangers of ESPN’s Player Rater
Sep 18, 2014; Houston, TX, USA; Houston Astros second baseman Jose Altuve (27) throws the ball to first base during the first inning against the Cleveland Indians at Minute Maid Park. Mandatory Credit: Troy Taormina-USA TODAY Sports
If I asked you who the top player was this past season in fantasy baseball, then your answer would probably be Mike Trout or Clayton Kershaw. While they were certainly the best hitter and pitcher in 2014, the answer according to ESPN’s Player Rater was Jose Altuve of the Houston Astros. Even though Altuve had a brilliant season, I am certainly not comfortable considering him the top fantasy baseball player this past year.
A standard fantasy baseball league uses a 5×5 format where the offensive categories are R, HR, RBI, SB, and AVG while the pitching categories are K, W, SV, ERA, and WHIP. The ESPN Player Rater is a (presumably) proprietary tool that boils down the fantasy performance of each player to a single number. Jose Altuve owned a 14.57 rating while Kershaw finished second with 13.52 and Trout posted a 12.84, which was third best in baseball.
From my assumptions, the Player Rater value is created by finding the sum of their player’s standard deviations above or below the mean in each of the five categories. For example, if the mean number of home runs in 2014 was 10 home runs and the standard deviation was 5, then a player with 15 home runs would have a standard deviation of 1 and 20 homers is a standard deviation of 2. A player with no home runs would have a standard deviation of -2. This process would be repeated for each of the offensive categories (or the five pitching categories if the player is a pitcher).
Furthermore, I want to point that I could be completely wrong in my assumption here because I do not know the formula that ESPN uses to calculate player values in their fantasy sports. However, I feel confident that I am on the right track.
Another assumption I am making is that each of the five numbers is a standard deviation and not a z-score. However, this is not very important because the implementation of standard deviations or z-scores does not favor a player one way or the other.
If the introductory statistics have bored you, then let’s get back to baseball and why I have a slight distrust of the ESPN Player Rater with respect to speedy ballplayers.
Jose Altuve won the AL batting title with a gaudy .341 batting average and stole 56 bases that tied him for second best in MLB with lightning quick Billy Hamilton of the Cincinnati Reds. As I said above, Altuve led all players with the highest fantasy baseball value according to ESPN’s Player Rater with a score of 14.57.
The vast majority of Altuve’s league-high score was derived from his average and stolen bases. Altuve earned a 6.51 for his 56 steals and a 5.05 for his league-leading .341 BA. In the other three categories, Jose Altuve recorded a 1.93 for 85 runs scored, a .04 for 7 home runs, and a 1.05 for 59 RBI’s. Those five numbers approximately add up to 14.57.
The fact that Jose Altuve was able to generate a score of 6.51 shows how the Player Rater formula skews towards base stealers. If you look at league leaders in each offensive category, then you will not find any other categories with players that are so many deviations away from the mean.
In this table, you will see the league leaders in each of the five standard fantasy categories, their stat in each category, and the player rater they received.
[table id=1289 /]
As you can see, the Player Rater considers 64 steals to be the most valuable fantasy asset. While I would prefer a player who hits 40 home runs or drives in 116 base runners, that is not the debate at hand.
The reason that the stolen base standard deviation score is so high is simply because the stolen base mean is so low. In fact, the mean seems to be roughly 4.5 with a standard deviation somewhere around 8. Therefore, 64 stolen bases is 59.5 more than the approximate mean, which is roughly about 7.5 times more than the standard deviation plus the mean. This is how Dee Gordon’s league-leading 64 stolen bases was valued at 7.52 points.
Now that we see how the ESPN Player Rater works and features a strong, but unintentional bias towards speedsters, let’s view the absurdity of using this system as the end all, be all when evaluating fantasy baseball players. According to the Player Rater, 64 steals is worth more than 115 runs and 116 RBI’s combined because 7.52 > 3.00 + 3.10.
I honestly do not know if anyone could truthfully prefer 64 stolen bases over 115 runs and 116 RBI’s. If you could choose to add either side of the equation to your team’s seasonal stats, then I could all but guarantee that choosing the R/RBI combo is the far superior choice. Seriously, their values are not even close, but 7.52 > 6.10.
This theoretical example almost came to life when you compare Jose Altuve and Mike Trout in 2014. Both players had terrific campaigns, but it is pretty apparent that Trout had the better year. Trout scored 115 runs, hit 36 deep balls, drove in 111, stole 16 bases, and hit .287. Jose Altuve scored 85 runs, hit 7 home runs, drove in 59 RBI’s, swiped 56 bags, and hit for a .341 BA. Despite massive advantages in three of the five offensive categories, Altuve outscored Trout 14.57 to 12.84 on the Player Rater.
I am just not buying it. The (assumed) standard deviations are in parenthesis. Somehow, 56 steals (6.51) and a .341 batting average (5.05) is worth 11.56 on the Player Rater while Trout’s 115 runs (3.00), 36 bombs (3.69), 111 RBI’s (2.92), and 16 steals (1.46) is worth 11.07.
Seriously, how could you even argue that Jose Altuve’s contributions in stolen bases and batting average are better than Mike Trout’s counting stats? There is simply no way for the Player Rater to even be close to accurate when it comes to valuing players with high stolen base numbers.
I think the flaws are quite obvious with respect to the SB category due to the fact that Jose Altuve finished first in the Player Rater, Dee Gordon finished eleventh, Ben Revere 22nd, and Billy Hamilton placed 34th. Revere and Hamilton each had two out of the five categories where they fell below the Major League mean, which gave them a negative score in those categories according to the Player Rater.
I understand why the formula values stolen bases so highly (stolen bases have an incredibly low mean and reasonably low standard deviation), but maybe they need to introduce a multiplying factor to reduce the unintentional bias of the stolen base category.
I want to point out that these stolen base performances are not historically impressive compared to the league leaders in other categories even with respect to the era. For example, a 100 steal season or a 70 home run season would warrant a very high category in this current era of slightly depressed offense and lower steal frequencies. The absurdly high value for stolen bases is simply due to the reason stated in the previous paragraph.
I think the lesson to be learned here is pretty obvious. Take the ESPN Player Rating with a grain of salt when it is comes to speedsters. Do not draft Jose Altuve first overall even if he could repeat as the top player according to ESPN’s Player Rater. Leave that pick for the player who truly gave you the greatest fantasy stats, Mike Trout.