Thread by @DFBeanCounter

So here is a thread on testing the new #bulletproofprospect process for wide receivers.

I took the exact same procedure and applied it as though it was 2011 and then tested it on the classes from 2012-2017 to see how it worked.

Lets dive in.

First off, here were the players it identified as bulletproof+ from 2003-2011. Please remember I only use the bulletproof process on players drafted rounds 1-3.

12/18 hit = 67%

I guess I should mention how the tiers work.

Transcendent > Generational > Bulletproof > Cornerstone > Maybe > Bust.

The tier names are NOT to be taken literally. They are just fun, so relax.

Cornerstone had 28 players of which 15 hit. 54%

These are them

Maybe was 6/17 = 35%

Bust was 6/49 = 12%

This list is too long for one screen shot.

This is a more manageable way to look at it for only 1st and 2nd rounders on the bust list.

So then I took these same thresholds and same methodology and applied it to 2012-2017 classes to see if it could predict who would and wouldnt be gud at catching footballs.

Here are the Bulletproof+ WR's.

15/19 = 79%

Cornerstone had 8/19 = 42%

1/11 = 9%

Bust had 3/26 - 11.5%

Hilariously outperformed the Maybe tier, but these are extremely small samples sizes where one guy can alter things by quite a bit.

Anyway, the hit rates are fairly comparable on both sets so it seemed to do its job predicting the future to some degree. 🤷‍♂️

So then I went back to the full sample, re-ran the calculations and applied it to the full population and will just roll this forward each year.

BPP+ - 28/37 = 76%
Cornerstone - 23/50 = 46%
Maybe - 7/27 = 26%
Bust - 8/73 = 11%

Here are a few sample classes so you can see how each group fared.

This is what the 2017 class looks like:

This is the 2016 class.

This is the 2015 class.

Oof that Jaelen Strong miss.

That 2014 class was truly special.

2013 class.

2012 was a rough one. Oh-fer in the BPP tier.

2011 class.

To add a layer of context on top. I have dynasty adp data back to '07. Here are the WR hit rates by dynasty rookie draft round for WRs taken rounds 1-3 in the NFL draft from '07 to '17:

Round 1 - 28/57 = 49%
Round 2 - 15/40 = 37.5%
Round 3 - 5/26 = 19%
Round 4 - 2/9 = 22%

Here are the hit rates by #bulletproofprospect tier in round 1 of dynasty rookie drafts to contrast.

BPP+ - 18/21 = 86%
Cornerstone - 6/17 = 35%
Maybe - 2/5 = 40%
Bust - 2/14 = 14%

Lets look at BPP+ vs all other tiers in round 1 of dynasty rookie drafts.

BPP+ - 18/21 = 86%
All other tiers - 10/36 = 28%

Then from there we can look at how rookie year #adptrends affects each tier, or rookie year production, etc.

It gets pretty neat.

Anyway, this is the engine that powers my dynasty rankings on patreon.

The first tier is only $3 if you are interested!

patreon.com/DFBeanCounter

This thread goes through a breakdown on each tier and whats included along with a few samples.

https://twitter.com/DFBeanCounter/status/1303893793293303808?s=20

Here are a few of the 2020 class profiles based more or less on the new methods. The thresholds may have changed a bit by the time I was done, but the grades are accurate.

https://twitter.com/DFBeanCounter/status/1303196831338307590?s=20

https://twitter.com/DFBeanCounter/status/1301840332435275778?s=20

https://twitter.com/DFBeanCounter/status/1303893705057656832?s=20

https://twitter.com/DFBeanCounter/status/1304639487884976128?s=20

This was an article on Bryan Edwards using the old BPP methodology. The new one looks much less favorably on him. He is no longer Bulletproof. He was consistently inefficient in college

Though I still have him ranked above consensus

dynastyleaguefootball.com/2020/04/20/bul…

This one was on Denzel Mims using the old methodology.

The new methodology is still not much of a fan.

dynastyleaguefootball.com/2020/03/23/bul…

FWIW, if you want to play around with all these numbers etc. I use @pahowdy 's database which is incredible.

Check out his patreon, he has about 10,000 spreadsheets on there which are bound to make you a better fantasy football player.