H-Value

Cracking DraftKings: H-Value (From Chapter 4)

H-Value

By Dan Hindery

 

How do you figure out which players‘ project to be the best plays each week?

The answer: H-Value.  H-Value was developed as part of a collaborative effort to determine which players are the strongest daily fantasy options.H-Value combines overall value (points per dollar) and projected points in a novel formula that provides an accurate ranking of the top overall plays each week. Simply sorting the players byH-Valueprovides an easy ranking system, and the player with highest H-Value is the single strongest play of the week.

Why is H-Value better than traditional value measures?

The two most common ways to rank players each week from a DFS perspective — points projection and value projection — are ill-equipped to identify the players that are the best plays each week.

H-Value is a better measure than simply using projected points and focusing just on players projected to score the highest. The players projected to score the most points skew very heavily towards the most expensive options every week, so they are not necessarily the best plays when considering their salaries.

H-Value is also a better measure than simply using a common value formula based upon points per dollar (or other similar measures). Simple value measures will skew far too heavily towards the most inexpensive options. Why is this? Every week there will be players near the minimum salary of $3,000 that will be projected to score 10 or more points. Meanwhile, players at the highest end of the spectrum with salaries near $9,000 will very rarely be projected to score 30 or more points. In terms of points per dollar, the more inexpensive players have a clear advantage in terms of traditional value measures. A list of the top DraftKings plays of the week based on traditional value measures will almost exclusively include players priced under $4,000, which is not very helpful in assembling a great lineup.

Lastly, H-Valueis a better indicator of true value than using simple salary multiple calculations. For example, a player with a $5,000 salary projected to score 15 points is referred to as scoring 3x (3 points per every $1,000 of salary). A $3,000 player projected to score 12 points and a $9,000 player projected to score 36 points would both have a multiple of 4x. Should they be viewed as equals in terms of how strong of a play they are? Definitely not, as the player projected to score 36 points is a much more rare and valuable commodity and would have a much higher H-Value.

In real terms, it is simply easier to find the lower-priced bargains that are expected to hit desired scoring multiples than it is to find higher-priced players hitting the same multiples. But you need to build your lineups around at least some higher-priced players with high overall scoring projections because a focus solely on value will lead to not using the full salary allotment and leave a lot of points on the table.

To further explain why two players with the same points per dollar value should rank differently, let‘s use an analogy. Imagine that you are a C student in biology. What is more likely to help you earn a higher grade, an A on a quiz that accounts for 10% of your final grade or a B+ on the final exam that accounts for 40% of your final grade? If you assume that the rest of your grades are near average, then the B+ on the final exam will actually lead to a higher overall grade than the A on the quiz. Similarly, a high-priced player with a high multiple will be more helpful to your chances of winning than alower-pricedplayer with the same multiple.

How is H-Value calculated?

H-Valueis equal to projected points raised to the square root of three, divided by salary and then multiplied by 2,000.

It sounds a bit fancier than it really is but we used a collaborative effort and some trial and error to come up with a formula we felt best identified the most valuable players and best options each week. The idea behind the calculation is that points per dollar puts a little too much emphasis on the per-dollar part and not enough on the points part. So we tried squaring points, which was better, but put a little too much emphasis on points. Raising to a power of 3/2 went a bit too far back in the other direction. Through a process of trial and error, we figured out that raising to a power of around 1.7 to 1.75 gave the most accurate rankings. The square root of three (~1.73) is in the right range and seems more elegant than just saying ―1.73, so that‘s the formula. After dividing by salary, we can multiply or divide by any constant we want, and the values will stay proportional to one another. So we multiply by 2,000 just to get numbers that are easier toread—e.g.,24.3 rather than 0.01215. 

The formula works well and gives DFS players a great idea of which players are most likely to be strong plays in a given week. Simply sorting the Footballguys projections by H-Value for running backs, for example, and looking at the top five players provides a great list for the focus of your weekly research before choosing two running backs to build your lineup around.

A real example picked at random from the 2014 season illustrates the strength ofH-Valueas a way to rank the overall top plays at each position every week. Let‘s look at the projections going into Week 14 of the season and the top five RBs as ranked by the two most common measures:

Rank

Points

Value

1

Le'Veon Bell

Jonathan Stewart

2

Arian Foster

Stepfan Taylor

3

C.J. Anderson

Matt Asiata

4

Eddie Lacy

Ryan Mathews

5

Jamaal Charles

Daniel Herron

The top five RBs ranked based upon projected points yielded five high-priced runners with an average price of $8,520. It is difficult to build a strong overall roster with those prices and a couple of the top five were very poor values. Meanwhile, the top five RBs based upon value yielded a group of very low-priced options with an average price of $4,260. A few of these guys produced very poorly in Week 14, which was not unexpected since they ranked so low in terms of projected points. Neither of these measures used alone gave a very good idea of the top options. Let‘s look at the Week 14 top 5 RBs ranked by H-Value:

 

H-Value

Points

Value

 

Week 14 Player

Rank

Rank

Rank

Final Scoring Line

Le'Veon Bell

1

1

12

185 Ru. Yards, 6 Rec., 50 Rec. Yds, 3 TDs

Jonathan Stewart

2

16

1

155 Ru. Yards, 1 TD 

C.J. Anderson

3

3

21

58 Ru. Yards, 3 TDs 

Daniel Herron

4

9

5

26 Ru. Yards, 4 Rec., 45 Re. Yards

Eddie Lacy

5

4

24

73 Ru. Yards, 5 Rec., 33 Re. Yards, 2 TDs

H-Value did a great job of pinpointing which of the expensive players (all of whom ranked outside of the top 10 according to the traditional value measure) were worth paying up for and ranked as strong overall plays. It also helped to identify a player like Jonathan Stewart who ranked well down the projected points ranking but was still a very strong play overall for the week due to his low salary ($3,800). 

How should H-Value be used?

Everyone uses the Footballguys staff projections a little differently and adds their own research to the overall puzzle of choosing a lineup. One great way to leverage the fantastic projections of David Dodds, Maurile Tremblay, and Sigmund Bloom is to use them as a tool to greatly limit the potential player pool each week. This smaller player pool provides an efficient starting point for weekly research.

Most DFS players simply don't have the time to conduct exhaustive research of every matchup for every player and every team. But if you can quickly determine the 20 to 25 best plays each week by looking at the list of players with the highest H-Value, it is much easier to research just those players‘ specific matchups in more depth. Further research allows you to narrow down your list to five to ten players that constitute your top plays of the week. Those players become the primary building blocks for constructing weekly tournament, Head-to-Head, and 50/50 lineups.

The top of the H-Value rankings also gives fantasy owners a good idea of what price range the real value plays will lie in that week. Some weeks, the value will be on the extreme ends of the spectrum, with some high-priced players expected to post huge fantasy numbers and other low-priced players providing enough value to successfully fill out the rest of the roster. Other weeks, a more balanced approach will be ideal if the leaders in H-Value are players closer to the middle of the price range that provide both great value and high projected point totals. Quickly scanning the H-Value rankings will help you decide whether a stars and scrubs or mid-priced value approach is the strongest play of the week.

Lastly, H-Value serves as a strong tool to compare across positions. It can give you a good indication whether wide receiver, running back, or tight end offers the strongest options for the week, which can be quite valuable in deciding which position you will use as your Flex option.