Fantasy Basketball Manifesto Part IV - why MRiS Rankings are Better

Obviously, since I went through all the work to develop these rankings I probably have a reason. In part I, I explained how most websites (Yahoo, ESPN, and Basketball Monster, among others I am sure) do their rankings (Standard Scoring). In part II, I explained the idea of Rarity Rankings. Finally, in part III, I explained in detail my ranking system, which I am calling Modified Rarity Scoring, or MRiS.

The concepts are the same - try to normalize production across all categories. In Standard Scoring, this is done by using standard deviations away from the average, and in MRiS it is based on the rarity of production above a minimum production threshold. Since both systems attempt to compare production in all different categories, it is easy to compare them. 

Standard Scoring assigns each player a negative starting point by subtracting from their production the average production in each category. This means that if they have a zero score overall, they are producing the average amount overall. In a 100 player league, they should be ranked around 50th. However, since every single player gets the same "average production" subtracted, we can ignore it. Like I explained earlier in the Manifesto, if both teams got 30 extra points to start a game, the winner is still determined in the 48 minutes of actual play.

The normalized scoring unit is the standard deviation. A single "Z score" is assigned for every standard deviation of production. In Modifed Rarity Scoring, Equivalent Fantasy Points are the normalized ranking units. So, by setting EFP equal to a Standard Scoring "Z score", we can compare the two systems and see which one makes more sense.

Before we get to the numbers, I'll make a pitch on purely theoretical grounds. This is the important argument - if we only agreed to accept results that we were hoping to see in science, not much progress would occur. Obviously, this is not a rigourous scientific endeavor (as much as I try!) but the idea is the same. In my mind, it does not matter how the statistics are distributed amongst players, only how many stats you can accumulate overall as a team. I see no logical reason that standard deviations should be added together and used to compare players. Fantasy basketball works by accumulating statistics, not players. Modified Rarity Scoring is based on the simple principle of equality of categories.

And now, on to the numbers...

Category	Standard Scoring	MRiS	Percent Difference (MRiS/SS)
Points	1.57	1.57	0.0%
3pt Made	8.76	10.25	17.0%
Rebounds	2.9	2.98	2.8%
Assists	3.45	4.01	16.2%
Steals	15.92	16.56	4.0%
Blocks	12.79	18.56	45.1%
FGOP	12.99	12.3	-5.3%
FTOP	20.28	31.3	54.3%
Turnovers	-9.4	-10.13	7.8%

As you can see, I 'anchored' the scoring systems to points. It looks like most of the categories are more valuable in my system, but really this is demonstrating that Standard Scoring does not account for the high True Zero of points. There is actually a wider difference than Standard Scoring allows between players since we expect even the worst player to score a significant number of points. 

The big differences between the systems are 3PTM, Assists, Blocks, and FT%. These are all categories with high standard deviations - some players score a lot of these, other players score few. The Standard Scoring method would have you believe that blocks are less valuable because blocks are more scattered among players. Does it really matter if you win a category with 1 player scoring 15 blocks one week? 

Standard Scoring is faulted because it does not account for the minimum expected scoring of the worst player worthy of being picked up, and it mistakenly assumes that the more tightly-grouped players are in a category, the more valuable that category is. Most seasoned fantasy players know that it only takes a couple good producers to win blocks for you every week, and a lot of times those same players will guarantee that you lose Free Throw Percentage as well! These two effects are discounted in the Standard Scoring method.

Stay Tuned to StatDance.com for our rankings pages, soon to come! I will post the MRiS rankings in standard leagues with their EFP in each category, along with some Free Agent ideas and players to target if you are tanking certain categories. A lot to come!