creating player projections

[sublimeandy]

how do you guys create your player stat projections? i just did my projections for the top 25 rb's and they seem a little unrealistic after comparing them to past years.

[sixer03fan]

I usually just go by the expert projections on different sites, and then avg them together.. This year to make my prerankings I averaged the cheat sheets of KFFL, Yahoo, ESPN, and Fantasy Guru. The mock drafts on this site is also very helpful.

help me out?
viewtopic.php?t=333163

[bobbing_headz]

I guess I'll go through my whole process.

First I usually start in about end of June where I just browse around taking in everything on the internet, TV, etc. I'll probably spend a good month just reading articles, scanning the Cafe, etc., where I begin to formulate ideas and predictions. No hard numbers yet though. I find that this way when you start making numbers it's easier as I'll already have a general idea of how I want my cheatsheet to look. Now I know it may be a bit late for this but I recommend taking at least a couple days, maybe a week, to just try and take in as much as possible.

After that I'll get started on projections. First thing to do is to grab up to date depth charts. I used the ones from KFFL, they seem pretty good. Next thing to do is to decide how many players to rank. I generally just assume that in a pretty standard league teams draft about 2 QBs, 5 RBs, 5 WRs, 1 TE, K and D. I don't really go in depth too much with this so just multiply each by 12 (or how ever many teams) and thats your number. If you're lazy you can just rank half that (so 12 QBs, 30 RBs, etc.) and get the remaining numbers from an outside source.

When I get to creating the numbers the first thing I'll do is usually make some sort of 3 Year average as a starting point. Then I'll change that based on what I've learned/deducted. Is the player consistant? Injury Prone? New HC? Loss of Surrounding Players? etc. etc. By this point I'll usually have pretty good numbers. A good thing to do is after you're finished making these numbers run them through team numbers over the past couple years. You'll know something is wrong if you projected the Lions to have 500 rush attempts, or Oakland WRs to catch 30 TDs. That kinda stuff. Also, make sure your totals add up, i.e. don't have Fitz and Boldin catching 12 TDs each and Leinart throwing only 16 or 17.

A couple things left to do. If it's your first time then after you've got the numbers check over a couple other projections from various sources to make sure yours aren't completely the opposite. It's fine to have guys who are different from what most project, I'd actually hope you do, just don't have LT rushing for 40+ TDs or stuff like that.

The last thing I'll do is turn my numbers in fantasy points and create cheat sheets for each position. Almost every year I'll take a look and see that a guy is maybe too high or too low and I'll tweak the numbers a bit to fit my tastes.

This is by no means the end. I constantly update my projections throughout training camp/preseason. Injuries, depth changes and other factors affect projections quite a lot.

Hope that helps

(Long)

[Crimedogg32]

Best way seems to do it is sort it about by teams then just trickle down. First off Id project numbers of rushes and pass attempts then project yards per pass and yards per rush to make sure everything is inline and you dont have a RB say with 8 YPR which is is unrealistic. Next step then is to divide up the pass attempts between the QBs rushes between the RBs and Receptions between the RBs, WRs, and TEs. Bascially just start at the top (a team) then work your way down until you have all projections for everyone on the team and thus everyone in the NFL (you can just say rank top 3 WRs or whatever then just group other WRs into category so you arent projecting 5 catches for every teams 5th WR). I feel this is the best way to do projections so you dont have the QB throwing 1000 more yards then a WR or the other way around and can help you find sleepers when you realize someone has to catch and throw all those yards you predicting.

[Atog]

For me, I do a moderate amount of research on the player's tendencies and such, and cap it all off by taking three bowls of dog food and labeling them "8 TDs", "9 TDs", and "10 TDs" to see how my dog feels on the matter.

[FantasyMan13]

and cap it all off by taking three bowls of dog food and labeling them "8 TDs", "9 TDs", and "10 TDs" to see how my dog feels on the matter.

I didn't know Funston was a cafe user...

[mattb47]

Best way seems to do it is sort it about by teams then just trickle down. First off Id project numbers of rushes and pass attempts then project yards per pass and yards per rush to make sure everything is inline and you dont have a RB say with 8 YPR which is is unrealistic. Next step then is to divide up the pass attempts between the QBs rushes between the RBs and Receptions between the RBs, WRs, and TEs. Bascially just start at the top (a team) then work your way down until you have all projections for everyone on the team and thus everyone in the NFL (you can just say rank top 3 WRs or whatever then just group other WRs into category so you arent projecting 5 catches for every teams 5th WR). I feel this is the best way to do projections so you dont have the QB throwing 1000 more yards then a WR or the other way around and can help you find sleepers when you realize someone has to catch and throw all those yards you predicting.

Yea, I think this is really the best way to go about it. Normally, barring some pretty hefty coaching changes, the amount of times a team will run and pass each year will be pretty similar to years past and so starting from there is a good way to begin and just working out what you think the division of labor will be for each player is where it starts to get tricky. I think this is likely the most accurate way to go about it though, even though nearly all predictions will be very inaccurate when the season starts anyway.