at

image credit: the maneater

Bottom Line up Front (BLUF)
1) PreSeason draft rankings are correct less than half of the time
2) Many leagues start only one at some positions. For those, the accuracy was: QB (50%), TE (60%), K (40%) and DST (20%)
3) Get used to waiting for a DST, along with that kicker, in the draft
4) High variances (projected – actual) show how nobody saw these coming: Smith, Wentz (QB), Kamara (RB), Thielen (WR) & Engram (TE)

All projects in the real world should include a review upon completion to understand what went well…and what didn’t. The idea behind this is to recognize how one, or the organization, can improve on the next iteration. Well, I am going to apply this idea to fantasy football. Specifically, I am going to look at how analysts predicted players at each position against how those players actually turned out.

Well? How did analysts do you might ask? The answer depends on what is thought of as a good result. When you choose to use your first two picks on a starter at WR and RB, you expect them to each be top 10 talent at their position. Unfortunately, you’ll likely only get this out of one of the two. If you’re in a PPR league you’re actually getting worse than that. 40% of the projected top 10 ended up in the top 10. RBs in standard scoring were the “best”, along with TEs, at 60% ending up in the top ten.

It’s a good thing many wait until the last pick to choose a kicker. They’re at 40%- so at least you didn’t spend big to make little. DSTs win the chase for last place. Of the projected top 10 teams…wait for it…TWO…two of them actually ended up in the top 10. I, along with many others, started picking up a DST in the mid-rounds. I will tell you this…I will not make that mistake next year. 20% tells me one thing: pick a position player at that spot next year and wait on a DST until all of the other starting skill positions are filled.

So where did they go? This part is likely obvious: they ended up lower in the ranks. The good news is that they likely didn’t fall too far. That is, unless you took the overall #1 in fantasy: David Johnson. If you “lucked out” and snagged him…you know that he ended up on the training table instead of your lineup. Rewinding back to not falling far…a bunch ended up in the #2 tier. You can see this in the fact that the top 20 ratio of those staying in the top 20 was much better than that of the top 10. Analysts were about 20 points (that’s 20 points, NOT 20% better for you stat geeks out there) better when the group is expanded to 20.

Another rewind to K and DST: It looks much better for these two positions when you look at the top 20 plot. This is really important for these two positions, as you only play one of each in most leagues. The only tier that matters is #1…so the fact that it’s better for both in the top 20 view does not matter.

The view below was used for the previous discussion.

This next part will focus on individuals. Which of them, by position, did really well? Alternatively, which ones didn’t? We know from the previously mentioned details of poor projections that roughly half are under-performing. With this in mind, we have a lot of honorable, as well as a lot of dishonorable, mentions. Let’s get started!

QBs
– Alex Smith stole the show. You will notice that he is at the top of the plot below. This plot is showing the difference (or variance) of the projection vs. the actual. In his case, he was projected to be ranked 22nd at QB and he finished at 4th. So, positive is good…and he was VERY positive
– Many won’t be surprised to see Carson Wentz right beneath Smith
– Those in the middle without a bar (or a small bar) landed roughly where they were expected to be
– Someone had to lose the 4 spot to Alex Smith. The winner (or loser) of that position was Matt Ryan. He is at the bottom of the list as a result of tumbling from 4th down to 15th

RBs
– Alvin Kamara is truly one of the biggest surprises in the last (make up a number) years. He ended up at 3rd overall in RB, despite being ranked 53rd on draft day
– Dion Lewis followed Kamara on the good side of the plot. However, this highlights one of the reasons this plot isn’t quite enough to tell the story. He moved from 51 to 15. This is still great, but we’re talking about moving to a #2 RB, vs. moving to one of the best #1s. This highlights the main reason I would opt for the lower plot (with the circles) in this article when looking at top movers
– We talked a bit about top 10 RBs moving down into the teens. There are four of them right there at the bottom of this plot. Howard, Freeman, Elliott and Murray all moved from the top 10 into the next set of 10 in final rankings

WRs
– Jones and Anderson went from expected bye week fill in up to tier #2 territory.
– Adam Thielen was a big surprise and was even better, as he finished as a tier #1 (#8 WR)
– Allen, Hill, Landry, Hopkins and Fitzgerald snagged top 10 spots from others after significantly outperforming what was expected of them
– Mike Evans is the main drop to notice, as he dropped from #4 draft day to #17 final ranking

TEs
– Don’t forget to look only at data that matters in the plots. It is best to adjust this view to a smaller segment of players. Chop down to those finishing in the top 10 by moving the default 20 at the top range of the “final position rank” to 10. This will show only the tier 1 TEs, as that is all that typically finds a roster spot (yes…there is a little higher in cases where someone stashes extra high ranked TEs or there are flex positions with TE as an option. In these cases, you can likely move the 10 up a little- perhaps to 15)
– Evan Engram was the primary good news at this position, moving from 23 to 5
– Zach Ertz was called out in the draft guide last year as being under-rated. I’m glad to see him living up to that, moving from #7 to #3

Ks
– I’ll mention two things: 1) keep it in the top 10 on the plot and 2) Greg Zuerlein was #31 on draft day and ended up as the #1 kicker. Yeah

DST
– This is a big surprise to me to see how bad the projections were in DST. Note that those first EIGHT were all projected to be out of the top 10. Draft day rankings couldn’t have been much worse if you had a blind-folded monkey throwing darts

For the dashboard: It is set on just QB for the default view (this is to help load times). Just select a different position with the upper-left filter. You can contract/expand the number showing (based on ending season ranking) with the center filter. Again, it’s best on a desktop…but “OK” if you hold a mobile device sideways.

Well, this next view is just a different way to view what was just previously discussed. The cool thing about this view is that you can see how many within each tier moved up (or down) to get there. In a perfect world, where analysts predict every ranking perfectly, every single indicator would rest equally along the centerline. This essentially would mean that they got exactly what was expected and that the variance (predicted finish – actual finish) was zero for everyone. We know that this world isn’t perfect.

I really like the ability to quickly see what positions have the most movers, along with how much they moved. For instance, I can see that nearly all of the players to move into the top 30 got to where they ended up by moving up over TEN SPOTS. This is incredible. How do we get better? That’s tough and if I knew a definitive answer…I’d give it. But the short answer is that this is why it’s best to take an aggregate view of as many analyst projections as possible. This is what I do on this site. I will continue to gather as many opinions as possible, which should push the percentage correct on those predictions as high as possible. You should also understand that this means there is A LOT of top tier gold in those middle rounds of the draft.

Knowing how bad these projections are should do one of two things for you:
1) Create less stress because you know ahead of time that only about half of your picks will actually meet expectations
2) Create more stress because you know ahead of time that only about half of your picks will actually meet expectations

Good luck with that!

Comments

comments

Footnotes

  1. Smith also describes himself as AccuWeather’s vice president of international strategy on his LinkedIn page.

  2. My husband, Christopher Baker, is a project executive at the Weidt Group, a Minnesota-based company that offers some similar services to EnergyCap.

Filed under

Comments