For those of you that do your own projections (1 Viewer)

[chris1969]

I'm working on revising my projections for next year and I'm trying to make it more of a formula to eliminate as much of the subjectivity as I can.

Things I currently take into account:

1) 3 year performance

2) Coach/Philosophy

3) Surrounding Players

4) Stadium ie.. turf receiver, dome kicker, cold weather RB

5) Injury history

6) Momentum ie.. on the rise or decline

7) Age

8) Schedule

9) Touches

Anything else that you would recommend?

Also I'm looking for a good resource for individual player scouting reports and a resource with detailed injury data to work into my formula if anyone has any suggestions.

 

[chris1969]

I'm projecting not alot of responses to this thread.

 

[MLBrandow]

Everything is a case by case basis. I think you pretty much hit all of the things I would consider other than one key thing:

Work Ethic.

That's a big deal, and is the reason guys like Braylon Edwards and Deuce McAllister are making more noise than most expected.

 

[Drop]

You should consider the guys that will be matched up against them. WR/CB matchups are simple, but you should try to find out if a guy will be shadowed or if he will be hidden in different sets at times to escape the #1 CB. Check out ProFootballWeekly.com or anything written by Pro Football Weekly, you will usually find little tidbits and comments from scouts.

 

[chris1969]

You should consider the guys that will be matched up against them. WR/CB matchups are simple, but you should try to find out if a guy will be shadowed or if he will be hidden in different sets at times to escape the #1 CB. Check out ProFootballWeekly.com or anything written by Pro Football Weekly, you will usually find little tidbits and comments from scouts.

Thanks for the suggestion, but I should have explained a little better. These are my pre-draft projections.Thanks for the work-ethic thing MLBrandow, I do konda take it into account, but would have forgot to put in in the formula.

 

[MLBrandow]

One other good one is total team rushing attempts or passing attempts, and see if you find a trend in that... some teams typically run the ball more than others, and yes this is obvious, but the actual numbers can be key in preventing some bonehead predictions, like Bush getting 250 attempts.

 

[chris1969]

One other good one is total team rushing attempts or passing attempts, and see if you find a trend in that... some teams typically run the ball more than others, and yes this is obvious, but the actual numbers can be key in preventing some bonehead predictions, like Bush getting 250 attempts.

I automatically account for that with coaching philosophy and touches.

 

[Mr. Mojo]

I spend a lot of time on the schedule difficulty, even evaluating the difficulty vs. the run and pass compared to the previous year. I also factor in players added and lost.

 

[turnit643]

For one more column to add to your pile ... look around and find a handful (5 maybe) of other projections from the various internet sites. Combine them into one big list of average projections for each player (I realize the point of this is to make your own, but why not capitalize on the work that many other people are doing) and run them through your scoring sytem to establish a starting point.

I generally begin with this and the 3-year past performance numbers and then adjust it all with the considerations you've listed. From there, it's not too hard to come up with drafting tiers and/or even a VORP-like inter-positional ranking.

 

[Biabreakable]

I see the words "factors and formulas" being used to alter the 3 year average model.

What weight do you give each of these considerations listed in the formula then?

Or how much of a factor is each category?

1) 3 year performance

2) Coach/Philosophy

3) Surrounding Players

4) Stadium ie.. turf receiver, dome kicker, cold weather RB

5) Injury history

6) Momentum ie.. on the rise or decline

7) Age

8) Schedule

9) Touches

10)Work Ethic

Are you adding or subtracting from the 3 year average by a certain %?

Are you shifting the distribution of the pie for one player to another based off this? And if so by how much?

I think the method of weighting based off these categories has more impact that the categories themselves.

 

[John Mamula]

I'm working on revising my projections for next year and I'm trying to make it more of a formula to eliminate as much of the subjectivity as I can.

Things I currently take into account:

1) 3 year performance

2) Coach/Philosophy

3) Surrounding Players

4) Stadium ie.. turf receiver, dome kicker, cold weather RB

5) Injury history

6) Momentum ie.. on the rise or decline

7) Age

8) Schedule

9) Touches

Anything else that you would recommend?

Also I'm looking for a good resource for individual player scouting reports and a resource with detailed injury data to work into my formula if anyone has any suggestions.

You want to start with each teams' projections and then distill it down to the players. You don't need to make this overly complicated, just look at a 3 year weighted average, analyze what has changed in the coming year and then make adjustments accordingly.However, your 3 year weighted average gives you a very solid baseline as to where your players will be. This values past performance and helps you shy away from hype.

With the exception of defense, schedule is overrated.

Use Drinen's site to load past stats into Excel. Once you set up some formulas there should be no work in updating stats from year to year and all of your work can be in projecting.

I would project for 2 QBs, 3 RBs, 4 WRs, 2 TEs, 1 PK, and 1 DT per team. Anything more is a waste and you can typically leave a few hundred yards out from your team projections compared to the sum of your player projections to account for the players who will never see a fantasy roster. No use in wasting time projecting whether RB4 on a team has 20 or 30 carries.

 

[turnit643]

I see the words "factors and formulas" being used to alter the 3 year average model.What weight do you give each of these considerations listed in the formula then?Or how much of a factor is each category?1) 3 year performance2) Coach/Philosophy3) Surrounding Players4) Stadium ie.. turf receiver, dome kicker, cold weather RB5) Injury history6) Momentum ie.. on the rise or decline7) Age8) Schedule9) Touches10)Work EthicAre you adding or subtracting from the 3 year average by a certain %? Are you shifting the distribution of the pie for one player to another based off this? And if so by how much?I think the method of weighting based off these categories has more impact that the categories themselves.

That's where the tinkering comes into play and that's what'll really separate your list from everyone else's. The reason I suggested collecting projections from other places and combining them into an average projection to start with is that they've all incorporated much of this same info into their work already, which means less tweaking is necessary. As for specifically how much weight for each category, you just have to play around with it and find what makes sense to you ... afterall it's just attempted fortune-telling to identify ballpark numbers and not an exact science.

 

[John Mamula]

The reason I suggested collecting projections from other places and combining them into an average projection to start with is that they've all incorporated much of this same info into their work already, which means less tweaking is necessary.

that actually just gives you an average of a bunch of bad projections.you will perform much better doing your own as you can capitalize on the sheep jumping off the cliff with poor player projections done by fantasy sites.

 

[Ministry of Pain]

I'm working on revising my projections for next year and I'm trying to make it more of a formula to eliminate as much of the subjectivity as I can.Things I currently take into account:1) 3 year performance2) Coach/Philosophy3) Surrounding Players4) Stadium ie.. turf receiver, dome kicker, cold weather RB5) Injury history6) Momentum ie.. on the rise or decline7) Age8) Schedule9) TouchesAnything else that you would recommend?Also I'm looking for a good resource for individual player scouting reports and a resource with detailed injury data to work into my formula if anyone has any suggestions.

Stop projecting. It's a waste of time. I don't think anyone "projected" McNabb, Bernard Berrian, or Frank Gore to do anything close to what they are doing...you would be time better spent if you got to know the OL inside and out, including the back ups...then knowing the defenses inside and out throughout the league so when schedules come out you can make an assessment of the schedule. You also need to understand coaching philosophy...some teams it wouldn't matter who is palying for them, the players never seem to do as well as we hope.

 

[turnit643]

The reason I suggested collecting projections from other places and combining them into an average projection to start with is that they've all incorporated much of this same info into their work already, which means less tweaking is necessary.

that actually just gives you an average of a bunch of bad projections.you will perform much better doing your own as you can capitalize on the sheep jumping off the cliff with poor player projections done by fantasy sites.

Do you really think your projections are going to be that different (and/or that much better) from what every other person that spends time and effort doing it is going to come up with? Everyone has access to the same info and will likely come to similar conclusions about the peripheral factors. Averaging those other projections I spoke of provides a decent baseline to launch from. That was my point.

 

[John Mamula]

Do you really think your projections are going to be that different (and/or that much better) from what every other person that spends time and effort doing it is going to come up with?

Yes.

 

[turnit643]

Do you really think your projections are going to be that different (and/or that much better) from what every other person that spends time and effort doing it is going to come up with?

Yes.

Right on. You rule.

 

[chris1969]

My actual plan is to use a little math to quantify what I already do. I'm gonna try my own version of DVOA and I'll try to mathmatically come up with results I'll compare to past years actual nfl numbers. If my results beat DVOA then I'll go with that. As far as what i'm building on I'll go with the charted plays for DVOA and go from there.

For instance I might say right now I'm projecting RB#1 to get 1200 yards, but I know he has had an ACL injury. What are the odds he might get another one. He stayed healthy all last year, but had 370 carries. What percentage of other RB's that had at least 350 carries stayed healthy the next year? He was a free agent this year and went from a west coast offense to a ball control offense. How will that effect his touches? He's also moved to a cold weather outdoor stadium, how will that effect his touches in my leagues playoffs?

You get the picture.

 

[Rogi]

I know it's tempting to dive into the numbers and estimate projections, but I think it is as accurate as palm reading. Even though it's only week 6, who predicted the number #1 players by position would be:

QB - McNabb

RB - Westbrook

WR - Berrian

TE - Winslow

K - Gould

DST - Bears (this isn't a surprise)

I don't project stats before the draft, but I do rank by position. I take into account:

- offense (run or pass oriented)

- schedule

- past injuries

- age

- surrounding players

- team off-season changes (o-line additions or losses, coaching changes)

- for kickers, I prefer one who plays in a dome or warm climate - I draft kickers last

- contract year?

And for some reason, I still always lose to the guy who does no research and picks strictly by name recognition.

Do you plan on tracking your prediction accuracy? I only ask because I never see any accuracy rates associated with website predictions.

 

[chris1969]

I know it's tempting to dive into the numbers and estimate projections, but I think it is as accurate as palm reading. Even though it's only week 6, who predicted the number #1 players by position would be:

QB - McNabb

RB - Westbrook

WR - Berrian

TE - Winslow

K - Gould

DST - Bears (this isn't a surprise)

I don't project stats before the draft, but I do rank by position. I take into account:

- offense (run or pass oriented)

- schedule

- past injuries

- age

- surrounding players

- team off-season changes (o-line additions or losses, coaching changes)

- for kickers, I prefer one who plays in a dome or warm climate - I draft kickers last

- contract year?

And for some reason, I still always lose to the guy who does no research and picks strictly by name recognition.

Do you plan on tracking your prediction accuracy? I only ask because I never see any accuracy rates associated with website predictions.

I had Mcnabb QB#3, Winslow TE#5 and Chicago Def #1I might check out the contract year thing and add it to the formula. I do plan on doing test runs and compare them to DVOA for previous years to see if it looks viable and then I'll be tracking the results every year and tweaking as I go.