Accountability Corner: DFS Performance And Observations

It’s accountability week for yours truly. Yesterday, I wrote/gloated about my winning roster in the ottoneu league FanGraphs Staff Two. I consider that post my reward for winning an otherwise free league. Today’s conversation is decidedly less celebratory. My DFS season didn’t come close to replicating my 40 percent ROI from 2014.

I actually lost money this year. In over $5,000 of entries – entirely in GPP formats – I lost about $60. Am I disappointed? You betcha. But at least I didn’t hemorrhage money. In three years of DFS play, I’ve still turned a tidy profit.

Despite that my column is named The Daily Grind, I have no interest grinding out head-to-head and 50/50 style contests. The ROI for those is capped somewhere around eight percent. After figuring in the site rake, you’d have to win over 60 percent of the time. Nope, GPPs are my game.

The format is pretty simple – you play for the big score. One spiked lineup will guarantee a profit on an entire season of contests (assuming you don’t spend more per day after winning). Two spiked lineups guarantees a gaudy ROI like my 40 percent from last season.

This year, I spiked zero lineups. I came within a home run of a big payday nine times. There was one day late in the season where three points from Chase Utley was the difference between a triple-up and a 50x return (it was an unusually tight range of scores and Utley was virtually unowned).

My performance suggests one of two things. The glass half full view is that I simply missed my day. A glass half empty observer might say I got lucky last year (I did make $700 on Brad Miller’s two home run day). Personally, I think my general process is just fine, but I’m on the hunt for ways to distinguish myself from the pack.

Speaking of the pack, it’s my impression that DFS players have improved dramatically over the last calendar year. I’ve also noticed that the analysis provided by other sites has improved. Last year, the DFS sites that recruited me often required that I use batter versus pitcher data. I either talked them out of it or turned down the job. This year, I was approached by more sites and none of them believed in BvP data.

Theoretically, better analysis will produce greater parity. This was reflected most obviously in the overwhelming use of platoon matchups. It got to the point where I was looking for Eric Hosmer versus a lefty just to find somebody decent but rarely owned. Last season, my picks were rarely rostered over 20 percent of the time. This year, I frequently ran into 40 percent ownership rates on guys I thought were sneaky selections.

I did (re)learn something about my DFS play – I’m biased by the first lineup I set. I don’t use an algorithm simply because it’s more fun to play by my gut. However, if I set a lineup on FanDuel first, I’d find myself with a similar group on DraftKings.

The two sites tend to have wildly different prices at times. To give an example, Josh Reddick might cost $2,700 on FanDuel and $4,600 on DraftKings. He’d be a massive bargain on FanDuel – both relatively and on his own. If I picked him for both teams, it probably means I was doing something wrong.

My biggest mistakes came in April when I took an aggressive multi-lineup approach. Many GPPs are of the multi-entry variety. Every once and awhile, you’ll notice the winner has the same lineup five times and therefore wins the top five prizes. While it’s fun to dream about such a huge payday, it’s a good way to lose a lot of money.

That’s what I did in April. When I actually did win a small amount with one lineup, it just barely covered my non-winning lineups. Once I switched to a single lineup approach, I quickly fought my way out of a $200 deficit. Then I stalled out for four months.

All-in-all, I’ll continue to tout my career earnings and ROI. I do hope I can get off to a hot start next season to fully restore my credibility!