Another ESPN Standard League Projections Update

Updated Aug 29, 2017.
ESPN Standard Leagues are NonPPR leagues, with 2RBs, 2WRs 1Flex

This is the table for the top running backs. Click here to see the full projections for all positions.
RBs
WRs
QBs
TEs

 

Name Mean Average Points Actual Points 10team Auction Price
(% of budget)
12team Auction Price
(% of budget)

Devonta Freeman
200.2 0.0 69.9 68.2

David Johnson
196.1 0.0 55.5 63.1

Ty Montgomery
172.8 0.0 30.7 37.9

C.J. Anderson
163.7 0.0 26.3 26.5

Todd Gurley
162.8 0.0 23.5 29.0

DeMarco Murray
161.8 0.0 22.6 30.2

Le’Veon Bell
159.4 0.0 23.4 26.0

Carlos Hyde
159.3 0.0 19.6 27.5

Isaiah Crowell
159.0 0.0 22.0 27.3

Kareem Hunt
157.5 0.0 20.7 25.7

Leonard Fournette
156.2 0.0 18.0 29.9

Dalvin Cook
156.2 0.0 22.1 25.1

Mark Ingram
151.1 0.0 17.7 22.1

Jordan Howard
150.1 0.0 16.2 22.0

Frank Gore
147.7 0.0 14.0 18.5

LeSean McCoy
145.7 0.0 15.2 21.9

ESPN Standard Fantasy Football Projections Update

Updated Aug 24, 2017.
ESPN Standard Leagues are NonPPR leagues, with 2RBs, 2WRs 1Flex

This is the table for the top running backs. Click here to see the full projections for all positions.
RBs
WRs
QBs
TEs

 

Name Mean Average Points Actual Points 10team Auction Price
(% of budget)
12team Auction Price
(% of budget)

Devonta Freeman
205.6 0.0 61.4 60.6

David Johnson
203.1 0.0 51.9 57.2

Ty Montgomery
176.7 0.0 32.8 35.3

DeMarco Murray
170.6 0.0 27.8 29.2

Todd Gurley
169.5 0.0 21.2 31.0

Jordan Howard
164.9 0.0 27.9 30.5

Le’Veon Bell
162.1 0.0 25.6 25.9

Leonard Fournette
158.2 0.0 23.1 24.0

C.J. Anderson
157.8 0.0 20.1 27.1

Isaiah Crowell
155.9 0.0 17.6 21.4

Carlos Hyde
155.2 0.0 21.9 21.5

Dalvin Cook
154.4 0.0 19.0 19.1

Mark Ingram
151.3 0.0 17.1 20.5

LeGarrette Blount
149.1 0.0 19.6 17.9

LeSean McCoy
147.6 0.0 13.2 16.0

Melvin Gordon
146.3 0.0 13.5 17.2

Paul Perkins
146.2 0.0 12.7 17.4

Quincy Enunwa, and Why Projections should not be Normal

Quincy Enunwa had a breakout season last year, and most people expected him to continue to improve in 2017, his second year as a number 1 receiver. Those expectations were wrong, because today the Jets announced that Enunwa injured his neck, and now will miss the entire season. It is a very unfortunate break for Enunwa. It shows the violent and unpredictable nature of professional football, where players can get severely injured even in scrimmages. It also shows a flaw in many projection systems which assume normally shaped distributions and thereby neglect the possibility of a catastrophic event.

In fantasy football, and in most future looking things, there is a non-zero chance of a zero outcome. This should be reflected in projections in that system, otherwise the projections are over-confident.

I try to follow this guideline for projection systems I make; for example, here is the distribution for expected points for Enunwa (standard league with no PPR), which features the possibility of 0 points:

Some other projection systems you might find have sort of a box and whisker plot describing the variance of their projections. However, none of these reflect the possibility of a 0 score. Regression based projection systems assume a Gaussian, or normal distribution, which is better: there’s a non-zero chance of a zero outcome. However, a Gaussian distribution probably understates the probability of a zero outcome. (There are other problems, for instance these distributions allow for large negative-value outcomes, sometimes when such outcomes are impossible).

Now, other systems will be quick to adjust Enunwa’s projections to 0. Players with Enunwa on their roster will be quick to grab another receiver. This responsiveness is arguably more important. But, if you can be both responsive and properly calibrated in your precision metrics, this is the best of both worlds.

Fantasy Football Projections for ESPN Standard Leagues

Here are some projections for ESPN Standard Scoring leagues. These leagues give 0 points per reception, which increase the value of running backs considerably.

The starting requirements for these leagues are such that you start 1 QB, 2 RB, 2 WR, 1 TE, and 1 flex.

The auction values reflect a percentage of bankroll, so if your budget is $200, you would multiply the values in the table by 2. The calculations assume a 12 team league. These values differ from many other auction values, so you might take them with a few grains of salt. In particular, it seems QBs are overemphasized compared to other valuation systems. If you do choose to use these valuations, it is smart to draft or auction with the intention of maximizing difference between the amount you pay and the amount our system projects.

For instance, in many ESPN mock drafts you’re usually able to get Devonta Freeman (or David Johnson) in the first round and then Leonard Fournette in the second, and are wise to do so, despite the system judging Fournette more valuable. You are also usually able to get Phillip Rivers (and other highly valued QBs: David Carr, Russell Wilson, fairly late in drafts).

You can see the full table of projections here: 2017 ESPN Projections

I have also put up projections for the past years’ season, 2016, here: 2016 ESPN Projections

Both systems were trained on data from 2011-2015, and you can look at how the model did last year to judge how much you might trust it this year.
(updated Aug 1, 2017)

Name Mean Average Points Auction Price (% of budget)

Russell Wilson
274.5 38.0

Matt Ryan
273.9 29.2

Aaron Rodgers
264.3 23.3

Matthew Stafford
262.9 20.8

Ben Roethlisberger
250.6 15.3

Philip Rivers
249.7 9.4

Derek Carr
248.8 9.9

Cam Newton
247.3 6.7

Kirk Cousins
243.8 6.7

Andrew Luck
236.6 4.5

Tom Brady
232.6 4.0

Jameis Winston
229.4 3.2

Andy Dalton
225.5 3.2

Marcus Mariota
202.9 2.5

Sam Bradford
202.0 2.0

Devonta Freeman
198.0 61.0

Leonard Fournette
195.7 54.9

Carson Wentz
193.7 1.8

Carson Palmer
190.1 1.4

Alex Smith
186.8 1.4

David Johnson
183.3 31.5

Joe Flacco
183.0 1.6

Dak Prescott
182.8 1.8

Tyrod Taylor
182.5 1.8

Ryan Tannehill
180.8 1.7

Antonio Brown
180.1 34.9

T.Y. Hilton
180.0 20.7

Full 2017 Rankings and Auction Values

How to Value Backups in Fantasy Football

It’s a tricky calculation to figure out how to value backups. During an auction, how much of your budget do you allocate to backups? When drafting, which is more important: a backup quarterback, or a fourth WR? First, let’s take a step back. We need backups for four scenarios:

  1. To fill in during bye weeks
  2. As replacement if a starter is injured
  3. As replacement if a starter underperforms
  4. As replacement if the backup’s performance exceeds the starters

The first scenario is straightforward to calculate: if the backup has the same bye week as your starter it has 0 value to you. If the backup has a different bye week, it’s value is about 1/16 of his projected value. The other three scenarios require some probabilistic work, because we do not know if our starter will be out-performed by the backup or injured during the season.

 

Above is a graph of expected points for Andrew Luck

 

Marcus Mariotta’s graph is similar. Luck is projected to score about half a point per game more. There’s a lot of uncertainty in the projection, and to see that a heap of ‘idunno’ is probably justified for future-looking predictions, try predicting the future yourself at quncertain.com.

In the system, Luck is projected as the 8th best QB and Mariota is projected as around 14th best, but it’s conceivable for either to be vastly better than this ranking at the end of the year. There is also injury concern with Luck. So, some teams will no doubt have Luck as a starter and Mariota as a backup with breakout potential. Also, both players have different bye weeks.

So how do you value Mariota? He could be your starter by the end of the season. Or, Luck could reel off an MVP year, with Mariota riding the bench and contributing very little to your score. It depends on who plays better. Because you are able to adjust your rosters week to week, you’re able to swap in the player with the better chance of a big game. Looking at the expected points for a weekly max-type function with Luck+Mariota gives a graph like this:

 

Expected points from a Luck, Mariota combination.

As you can see, backups are useful because they:

  1. Mitigate downside risk
  2. Increase the median of expected points

Adding a third backup is an improvement again, just not as dramatic. Adding Trevor Siemian as a third QB gives a graph like this:

Expected points for Luck, Mariota, Siemian combination, truncating the left tail, elongating the right tail, and shifting the median a little to the right.

 

With data like this we can calculate the value of backups, directly relating price to the amount they increase a team’s chance of winning a league. Calculations for positions with multiple starters, and flex options are more complicated because of the many combinations to consider, but the results are approximately the same. I hope to have my auction valuations on the site soon.