The One Percent: How to Analyze Your DFS Results Without Being Results-Based
As the daily fantasy sports (DFS) landscape continues to evolve, analyzing your results has never been more important or more challenging. The contests offered by operators are increasingly large, increasingly top-heavy, and a higher percentage of play has shifted away from cash games and towards guaranteed prize pool (GPP) formats. To go alongside this, across the industry there is easier access to optimizers, projection sources, and overall strategy content that has increased the skill of your DFS opponents. All of this has lead to a more competitive environment that also has an inherently higher degree of variance, and particularly so for those DFS players who are chasing the dream in lottery-style GPPs with six or seven-figure 1st place prizes. It has also made it more important than ever to understand your results and whether you are are a positive expected value (+EV) player.
At a minimum, most DFS players should be able to perform basic analysis. Getting your 1099 tax form each year is nice (hey congrats on being profitable!) but it doesn’t tell you much about your performance. Were you on the right or wrong side of variance? Maybe you won $1000, but actually played like a $10,000 player, or maybe you were a fish on a heater and actually played like a $5000 loser. If that was the case, wouldn’t you want to know so you could adjust your play?
Where to Find your DFS Results
If you haven’t done so already, you can download your CSV from DraftKings or your Fanduel History Tab into an excel spreadsheet. From there, lots of different data points to consider. Most of this analysis is possible in excel, but personally, I use RotoTracker to track my DFS results and will be sharing analysis based on that. This is not an ad, and RotoTracker is expensive for a player with a smaller bankroll, but it allows you to spend more time analyzing your play and less time writing formulas or constructing pivot tables.
The sample size is a caveat with all of this. It is important not to over-react to small samples or slice your data into too nuanced of a lens, but I’ll discuss in a few examples why I think even small sample data should be considered.
Return on Investment (ROI) and In the Money (ITM) Percentages
Beyond raw dollars, the next logical step towards measuring your play is by analyzing ROI and ITM percentage. While sample size issues are important for both, analyzing the ROI lets you better gauge skill. If you have achieved a $1,000 profit on $100,000 of total entry fees (1% ROI) that tells a much different story than if you have achieved a $1,000 profit on $10,000 of entry fees for a 10% ROI.
Similarly, ITM percentage helps you understand what percentage of your entries are “cashing” in various contests. If you are playing large field GPPs, a typical payout structure in 2020 may pay 25% of contestants, while smaller field tournaments like 3-max may pay 20% of the field. As such, a perfectly average set of results may cash 25% of the time in tournaments. If those are your results, that is great! But wait, you say, if the field is achieving 25% and I am achieving 25%, how am I supposed to beat the rake?
In GPPs it is all about being in that one percent. The best way to track your DFS performance is not by profit but rather based on what percent of your lineups you put into the top one percent.
Analyzing the One Percent
Before diving into the one percent, it is worth exploring the basic deciles of your performance. Generally speaking, my goal is to put more than 10% of my lineups inside the top 10% of tournaments, and more than 1% of my lineups inside of the top 1% of tournaments. The hypothesis behind this is that who ultimately wins or loses in a given slate is sadly determined by luck more often than not, but by consistently putting shots on goal we are putting ourselves in a position to be lucky.
#Humblebrag but the chart above is generally what you would see in a winning DFS player over a long period of time. In addition to consistently winning at the Top 10% (12.5%) it is a progression down, also winning at the Top 20% (11.1%) and Top 30% (10.5%) while breaking even at the 40th and being underweight all of the losing positions. Generally, I think if you can put 11.5% of your entries in the Top 10% of a GPP you’re in a position to overcome the rake which is a great first step.
Still, if you enter the largest field GPP for $20 and finish in the Top 10% you would often make a whopping $35 or $40, far from what dreams are made of, and that is without considering you may have 150 total entries. Putting 12.5% in that mix seems good, but how does that lead to a mid-40s ROI level? What we are really searching for is that top one percent.
The chart above shows 1.4% of entries in the Top 1% which helps better explain the strong ROI, in addition to winning at all levels within the top percentage of GPPs. Not as clearly noted, is that I have put 1635 shots on goal into that top-level which have helped to achieve 38 wins of 5-figures or more. Only 2.3 percent of my shots on goal have translated into big wins, highlighting the importance of getting a lot of shots on net.
Shots on goal are more important than immediate results. My $15,000 in NBA losses will help you understand why.
NBA Results vs NBA Percentiles
I have a bit of gamble in me, but I was still surprised to find out that I’d wagered more than $50K on NBA DFS contests and flushed a used car down the toilet in the process.
My ITM% was solid at 30% but I was still kicking out a “not just losing to the rake you are worse than the rake” level of ROI across contests. However, digging into the performance percentiles is a bit more encouraging.
My sample size is a bit small for the NBA, but I simply have not taken enough shots on goal. In total, I had only put 57 shots on goal in the NBA which based on my all-sport one percent to GPP win conversion would yield 1.29 expected GPP wins. I may have been running bad (or may still be bad) but my results were only running one win below expectation, a win that given GPP prize pools for NBA DFS would have turned this negative ROI into a profitable one.
Bink or Bleed GPP life is real and it is a thing. That life is illustrated perfectly in my NBA GPP play but if your results are similar it doesn’t need to discourage you. Keep grinding.
A Fish on a Heater and How to Spot It
My lesson on how to spot a fish on a heater begins with the newest addition to my DFS play (CSGO) and ends with a discussion on some run pure from my first true DFS love (NHL DFS).
When DraftKings first launched eSports I dabbled with CSGO before League of Legends since the smaller CSGO prize pools would keep the whales focused elsewhere, and I had access to data courtesy of my friend Adam. I proceeded to peel off GPP wins on back-to-back slates and was convinced I had found a new gold mine before I bled some profits back into the ecosystem.
After gathering a bit of a larger sample there were a few things that were clear:
- My ITM percent likely reflected that of a losing player unless I had an elite 1% rate.
- My one percent rate was either ELITE, or more likely, FRAUDULENT.
- Results like this couldn’t sustain with an accidental contrarian strategy and more often than not playing the worst plays.
Unlike the results chart for my overall play which won not only at the one percent but each additional level in a natural progression, my CSGO chart more likely reflected a fish on a heater. An elite one percent rate but no clear pattern of skill, and a fairly clear pattern of putting up bottom end performances.
I was a fish on a CS:GO heater. That doesn’t mean I need to quit playing but should expect to adapt my play or die.
My First True DFS Love
When people talk about games getting tougher, or the DFS ecosystem, or how rule changes impact GTO strategy I always think about my first true DFS love, NHL DFS.
Lifetime NHL DFS Graph
One of my more consistent DFS sports was NHL DFS which allowed you to recognize your skill even faster given smaller fields and nightly contests. In 2019-20 things changed. DraftKings changed the scoring system and it took me a while to adjust my strategy despite clear signs that things were different. Combining that with more distractions (NFL DFS Showdown Slates M-TH-SUN) and things were set up to be a challenge. One that I did not respond to particularly well.
2019-20 NHL DFS Graph
I was still putting 10.7% of my lineups into the Top 10%, and winning at 20% and 30% levels. But it wasn’t good enough to sustain long-term results. With the change in scoring and strategy, increased opponent skill, or lack of focus, I had went from the profile of a winning to losing NHL DFS player.
Of course, you can also just run hot!
Despite my most profitable NHL DFS season to date and an FHWC Championship, my NHL DFS results were poor and improved play will be required to sustain profit.
If you have a desire to be in the one percent of DFS players, start by analyzing what percentage of your lineups are top one percent lineups. From there, dive into different areas where there could be data that influences your play. Some important elements to analyze include the following:
- Sport: An obvious starting point, analyze each sport you play to determine where your best results are.
- Field Size: Is your strategy more optimal in the smaller fields or larger field tournaments?
- Game Style: Classic, Single Game Showdown, Pick’Em, and more. Results can vary widely by game style, or game style and sport.
- Contest Dates: Past performance is not a guarantee of future results. Study results for different years or seasons to understand if you are still performing at a high level.
- Website: Are you having materially better results on DraftKings, Fanduel, or smaller providers?
- Entry Fee / Stakes: How do your 1% rates look as you move up in stakes?
- # of Entries: Are your results better in a single entry, 3-max, 20-max, or 150-max contests?