The Numbers Game: Analyzing Daily Fantasy MLB Data

rg misc mlb analytics

The realm of daily fantasy sports has been growing at an unprecedented pace of the past few years bringing in fresh blood everyday for a shot at fantasy greatness. Some take a whimsical approach simply guessing at good plays, some do research and base plays off splits, and some of us go above and beyond letting our inner number junkie run free.

I started playing DFS with NFL this past September and continued through the NBA season and now transitioned to MLB. The one philosophy I have held true to this point is “numbers don’t lie.” The only question: Which numbers are you really looking for?

DFS is quite comparable to many other games of skill (all you online poker fiends can nod your head) and can even be treated as a means of investment. Who doesn’t love the chance to watch the game you love and make a little extra cash on the side? In taking this game of skill to the next level, I bring you “the numbers”.

The Numbers Game: Sport vs. Sport

With the current MLB season in full swing, I took a good chunk of data from Fanduel to paint a numbers picture in search of a means for long term profitability by analyzing head to head, 50/50s and double ups. 1

The first thing I wanted to confirm was the variance between sports in daily fantasy. Looking specifically at head to head match ups, I broke it down by score percentile to give an idea where you would have expectations to be scoring as to hold down a long term winning percentage. Please note that any data from any one user on one given day was only counted once in the sample size (i.e. a user had 10 H2H contests with a score of 31 in MLB on Friday → this counted as one unique data point, not 10). This approach gives the best representative sample for the data accumulated.

TABLE 1

BY SPORT PERCENTILES STATISTICS
ALL H2H SAMPLE SIZE 30th 50th 70th 90th Average Standard Deviation Std Dev % of Mean
NFL 338 107.94 118.88 133.92 149.49 120.44 22.11 18.36%
MLB 871 26.58 31.58 37.45 47 32.44 10.82 33.35%
NBA 498 235 252 272 296 253.53 31.48 12.42%

Take what you will from the numbers posted but the most interesting are the standard deviation and the standard deviation percentage of the mean (average). By breaking it down this way, it is easy to see MLB has the highest score variance followed by NFL and then NBA, as expected. It is just crazy to think just 68% (one standard deviation from mean) of scores for MLB on FD fall between 21.62 and 43.62! With such a large variation, it WILL be more difficult to make a major profit margin in MLB but with patience and consistency, it is still possible.

The Numbers Game: GPPs, 50/50s, and H2Hs

TABLE 2

Percentile $1 – 50/50s $2 – 50/50s $5 – 50/50s $5/$10 Double-Ups
30th 26.52 26.91 27.21 28.99
50th 31.48 31.97 31.85 33.99
70th 36.37 36.93 36.82 39.42
90th 44.46 44.18 45.14 48.18
Sample Size 92 82 66 22

I would say the most common misconception about DFS is that the only way to hit it big or make money is to win a GPP. For all you new guys, spare yourself the heartache and start by catching on with H2H and 50/50s. Take a look at the 50/50 data above. There isn’t much difference between the 50th and 70th percentile scores for each entry value, $1, $2, or $5. The sample is a bit smaller than the H2H data, but it is still somewhat representative. With more data on the $10+ 50/50s, I will almost guarantee you will see the scores increase. This is both a result of the smaller player pool and the quality of the players.

Another misconception is that 50/50s and double-ups are the same contest type, when in fact they have slightly different pay structures and varying field sizes. The small sample for the double ups doesn’t do it justice but there is one point I still want to make. You will see higher scores at every percentile as the very low newb scores will be by diluted by the larger play fields (223 for $5 double ups and 112 for $10 double up on most nights). Based on the numbers I would recommend sticking to 50/50s over double ups.

The Numbers Game: Stakes vs. Opponent Quality

TABLE 3

Percentile $1 – H2H $2 – H2H $5 – H2H
30th 26.5 26.18 27.25
50th 30.83 31.54 32
70th 36.7 37 37.91
90th 44.6 45 48.85
Sample Size 147 396 328

The H2H show what most of us would expect. That is, the scores do increase as the entry value increases. The 50th percentile score from $1 to $2 increases by 2.3% and from $1 to $5 the score increases by 3.8%. Although this may seem to be quite a small increase, it boils down to being a few extra outs by hitters, a strikeout for a pitcher, an inning pitched, or any variety of hit. We have all been on the winning and losing ends of a H2H or bubble a small tourney by less than one point. The fact remains that this small increase can have a long term effect on your profitability if you ride the fine line of mediocrity when it comes to consistent MLB play. Just remember that if you look at the right numbers, the numbers NEVER lie!

1Thanks to those that helped contribute data in the forums

About the Author

nohitter48
nohitter48

Dan has been an avid full season fantasy player in NFL and MLB since 2005. He discovered DFS during the 2012 NFL season and had great success which carried on to NBA and soon MLB. Dan’s background as an engineer and derivatives trader lends to his unique style of a truly statistical approach to DFS, allowing self written programs to choose his lineups.