The DHL Algorithm
Some of you have asked me to go into detail about how this website works, and why my projections are what they are. DHL came about because I was sick of spending hours every day reading websites and searching through statistics, in order to create a good DraftKings lineup. I wanted a way to cut down on the time it took to create a good lineup, without reducing the quality of the lineup.
How it started
One night I was running some numbers trying to figure out which player was more valuable, when I decided to write a short algorithm to do it for me. This algorithm looked at two player's cost and average points and returned the best valued player. This algorithm could only do comparisons between two players, but it still didn't save me enough time to make it worth using, so I decided I would work on the algorithm over time to make it better.
What it is Today
The DHL algorithm saves me a lot more time than it used to. Now it creates all possible 9-Player valid lineups, and then sorts through them looking for the top lineups. It chooses the best lineups based on each players expected value in the lineups.
How does DHL get a players expected points for the website? There are a bunch of variables that are looked at, and I change them and update them all the time to try and make predictions as accurate as possible. Sometimes I use reddit feedback to update the algorithm, and sometimes I will stumble upon an article that will make me change my variables.
The main statistic that is used to come up with a player's projected value is their average fantasy points per game this season. Once I get this statistic I apply multipliers to it based on that day's games in order to try and make the number more accurate.
For example, lets say Player A averages 5 points per game and is playing against the best team in the league. And player B averages 5 points per game and is playing the worst team in the league. Then the prediction would say that player A might be projected 4.5 points tonight and player B might be projected 5.5 points.
The above example is just an example, meant to give you an idea of how I apply multipliers in order to get a better idea of a player's output on a given night. I also go through the same process for many other variables such as, if the player plays on the power play, or how much time on ice a player gets.
Are DHL predictions accurate?
I work daily on making sure the predictions are as accurate as possible. But sometimes there are players with out-lying statistics that can show up in lineups. For example when Leon Draisaitl averaged 8.5 fantasy points in his first two games he showed up in every single lineup. So that day I went into the algorithm and reduced his value in order to get a more accurate top line. Leon Draisaitl still shows up in some DHL lineups but I think he is a good value pick even though his expected points are still higher than what I actually expect.
Using feedback from the reddit community along with my own experience, I can conclude that DHL lineups are accurate. I suggest playing DHL lineups in cash games, since the lineups seem to have more success there, but they are also okay in GPP's.
I hope DHL has reduced the amount of time you spend researching for your lineups. Since, that was the goal of writing the algorithm. Also, I have put in hundreds of hours of work into the algorithm, which I wrote from scratch. I update the website and algorithm daily, in order to maintain lineups that are as accurate as possible. So, if you have had success with DHL lineups, or you enjoy the website. Please feel free to donate using the donate button below. It really helps me out.