The thread for people who do their own projections (1 Viewer)

[Scoresman]

Let's have a thread on making our own projections.

How many of you do this? What stats/information do you use?

I like to make my own projections, not because I think I'm better at projecting than the experts, but because it really helps me get to know every player and not form biases, and all that. Plus, I'm a huge numbers/spreadsheet nerd so I love doing it.

Myself, I project players by projecting team offenses. I basically have a giant spreadsheet that has tabs for each team. I look at last year's stats, coaching tendencies/changes, OL stats, team defense changes and vegas props, all to project each team's passing/running offenses. I look at each team's total projected offense compared to last year and last three year average and make sure what I have makes sense.

Then I apply that to individual players based on things like career YPA/YPR/ADOT, etc. These player projections then are used to populate VBD cheat sheets.

I also compare my projections for each player to fantasypros expert consensus projections, not because I'm trying to match them but to see how much I differ from them. If I differ greatly, I need to have a good reason for it. Just a good check of everything.

 

[BobbyLayne]

Good idea for a thread.

Agree wholeheartedly the baseline for the model starts with determining team projections. I like to begin with analysis ToP, pace and number of plays and of run/pass ratios. Where there is continuity (say Ben Johnson in Detroit) I'll look at two years data. Other situations are a little more educated guesswork - 5 new HCs and 11 new OC/playcallers.

But my first task is always trying to determine how many plays I have to work with for each team, and of those how many will be passes and how many will be runs. You're going to be all over the map (IMO) if try to build it in reverse and do the player projections first.

Not sure how many truly stat out every player a la Mike Clay. Obvious Q: How many would cover most fantasy relevant players?

• QB - every team's starter plus partial season starters like Colt McCoy, Taylor Heinicke, Kyle Trask, Jacoby Brissett. I'm not overly concerned about the 28 or so guys who will likely see 75 or less PA.
• RB - last there were 57 RBs with 300 or more rushing yards, 60 with 20 or more targets, and 15 with <20 targets but more than 300 yards rushing. Thus 57 + 15 = 72
• WR - last year there were 100 players with 40 or more WR targets.
• TE - last year there were 40 players with 40 or more TE targets.
• That's 248 offensive skill position players + 32 Ks and 32 DTs.

Roughly 300 projections.

Be back later to add more.

 

[Scoresman]

I don’t bother stating out every player. I do 3 RBs, 3 WR, and 2 TE for every team. T Then I have a row for RB4+, WR 4+ etc and just lump stats for other players there to meet team totals.

 

[Harry Frogfish]

Mad respect for those that do. It's become too easy to rely on other's work, and then the intangibles involved. I liked it way better before.

Best of luck to those that are that committed, they deserve it.

 

[Gally]

I don't have the time to do that much work but I will alter known projections for the purposes of making my own draft lists. I don't necessarily go in full detail for those players but look and decide if I am higher or lower on a player based on the projections I have found. Kind of a modified model. The raw numbers don't matter to me as much as the relative numbers player to player.

 

[Joe Bryant]

Love this idea.

 

[yankeerebel]

I've been playing since 1996 and the whole reason I pay for FootballGuys is so I no longer have to do all that research myself. I may choose one guy over another on my pick based on my own opinions but I pretty closely stick to the opinion of the guys who work full time compiling all of the data. Thank you, FootballGuys!