#1 By: DanC, August 3rd, 2018 13:11
I remember there being a season-level similarity app last year. Does it still exist? Has it been incorporated into a different app?
#2 By: DaveCabanFF, August 3rd, 2018 14:52
I included historical projections (built in a very similar fashion to those in the sim app) in the Stat Explorer which is part of the FFDRAFTPREP package. I'd recommend checking it out!
#3 By: DanC, August 3rd, 2018 16:57
Checked out the stat explorer. Any way you could create a column for efficiency stats (Yds/carry/attempt/rec)? I was hoping to use the similarity app to generate efficiency rates for the projection machine that are more player specific than league averages, but less subject to the randomness that career averages might be.
Or this could all be moot, if you have a better way of assigning yds-per rates in the projection machine, in which case, please advise.
#4 By: DaveCabanFF, August 4th, 2018 22:04
Now that you mention it, I think that a fair amount of readers would find that interesting. I'll add doing a post that includes tables with that info to my list of to-dos. I will say that for certain measures you lose some context cause the stat lines formed by the average historical projections just about always end up hovering around league average. If I remember correctly from when I looked into this in the past, yards per carry is a good example of this.
Personally, I prefer to review a particular player's history, see how often they go above or below a particular percentile, review their history and then regress or progress them over the percentile that they hover around based on the recent seasons. I do take into account their historical projection when doing this, and if a player's comps killed it in the coming year, I'd take this into account. This surely isn't an exact science but it keeps me within the realm of possibilities without getting too carried away. For most players, tweaks to efficiency don't make major impacts (within one particular metric) as long as you're not going over the 75th or below the 25th percentile.