Fantasy Football Projections - And How I Create Them

[madd futher mucker 36]

As discussed in Part 1, it is NOT necessary to use projections at all, much less create your own. I know many on this board have developed such good powers of football observation combined with a well developed fantasy football "gut feel" that they can consistently be winning fantasy owners without them. Some of you at least partially rely on the projections of others and also do quite well.

 

So the second part of this series is only for the very few of you who would like an insight into how to build your own projections, or at least are curious to learn how I go about building mine. I'd venture to say that no two people create their own projections in exactly the same way. In describing my process in detail, I must tell you some of the (free) sources that I rely on to to construct my projections. At least one of my sources is competition to my favorite fantasy football website, FFToday. So close your eyes, Mike, and pretend that you didn‘t read some of this, OK?

 

The process really starts with daily reading of the most relevant articles from four main news indexes. The four must-read indexes are Rotoworld (for the shortest and most incomplete versions, although I love their in-depth articles, especially those by Evan Silva); The Fantasy Football Librarian (Sara Holladay is fantasy football smart and also quite the looker); theredzone.org/ ; and NFL Gateway - NFL News. By keeping up daily, I always maintain a current knowledge of the latest information relating to players from all 32 teams.

 

The projections process officially starts by reading and taking notes on Team Reviews found on any number of websites. Some very good ones I look at are Early Observations by our own Doug Orth; FBG’s Team Reports and CBSSports.com Team Reports. I also review the 2012 Coaching and Management Carousel which is pinned on the FBGs Shark Pool forum and I try to analyze the impact on a team's overall offensive performance.

 

Next I go to FBGs Player Projections and I review the four independent projections of players by position. The reason I use FBGs is because the four sets of projections are done independently and are all contained in the same place, so I can easily get and record four separate numeric ‘storylines’ for every player. (BTW, all content from FBGs is free until about mid July, and I have my base projections done way before then.) From these four sets of numbers for each player, along with my general understanding of the overall team dynamics, I create my own numeric storyline of what kind of season I think each fantasy relevant player is going to have. For each RB for example, I make a rough projection of rushing yards, receptions, receiving yardage and TDs and PPR fpts. For each category other than QB, I take my projections down to about 115 PPR fpts. Beyond that, individual projections are just not relevant anyway. Next, I list the players by position in order of total fantasy points to create preliminary rankings. I then study these rankings to see where my preliminary projections just don‘t make sense and make some reasonable adjustments. I also note those players who carry particularly high risk factors (ex: Jahvid Best), or particularly high ceiling break-out potential beyond what is reflected in my numbers.

 

Then I go to FBGs Offensive Projections By Team. There are several reasons why I also use FBGs for this. This is the source I first used when I started doing my own projections. Beyond that, I love the fact that they show three prior years of actual team history for all offensive categories. I plug into a Team Projections spreadsheet my estimation of the overall team QB Completions, Interceptions, Passing Yards, Rushing Yards, and Passing and Rushing TDs for the QB position. For RBs, I plug in my overall projection of Rushing Yardage, Rushing Tds, Receptions, and Receiving Yardage. I treat WRs and TEs as a single group and plug in the overall Recepts and Receiving yardage. These are my Positional Totals.

 

From there I go back to my rough projections for individual players on that team by position, and see if there is a reasonable ‘fit’ with the Team Total Projections. The Total columns should always be higher than the sum of my projected individuals, because I do not project the ‘bit’ players who are not fantasy relevant. I can look at any of the four independent FBGs Team Projections and decide what are reasonable numbers in total to plug in for these minor players. Then, by adding all the numbers, ideally I have good fits with individual players roughly adding to positional totals, and having Total Completions by QBs equaling Total Receptions by receivers and running backs, and Total Passing Yards roughly equaling Total Receiving yards etc.

 

If the fit is poor between my individual player projections and the Total Position projections, I obviously have to re-examine my thought processes for both the composite numbers and my projections for the individual players. This is another check on my thought process which will often lead to me revising a player’s individual projection. When I started doing this I had some pretty gross ‘misfits’ between individual and composite numbers. These days, I usually get things to add up pretty closely. Understanding the process I've described so far would be much easier if I showed just one team as an example. When (if) my computer guy can restore my lost team projections, I'll edit in one projected team so you can see it.

 

I want you to understand that I use the FBGs projections merely to serve as a template and as a rough guide to help clarify my own thinking; since they are four independent sets of projections it is easy not to let any one of them influence me very much, even though I do interpret the different numeric storylines they represent. So the projections truly represent my thoughts and my storylines.

 

When I am satisfied with the initial finished projections, I have a starting point for any subsequent revisions based on any fantasy football news that might influence them. For instance, if a fantasy relevant player gets injured prior to my draft and will be out for any significant period of time, I must immediately go back to my total Team Projections and re-allocate the rushing yardage, receiving yardage receptions and TDs among the players remaining to pick up the slack. This re-allocation must also be consistent with how I now view that the injury will affect the team offense as a whole. Similarly, if I see that a certain wide receiver has moved up the depth chart in a major way, I will have to re-examine not only my projections of that player's fantasy performance, but also identify and modify those receivers most likely to lose some touches due to the change in the pecking order.

 

For the finishing touch I always look at specific position strength of schedules to determine how much impact they will have on the Projections, and again I might make a minor tweak to my numbers. While there are other sources for this such as FFToolbox, there is no better tool for this Doug Orth’s Big Boards of Schedule Analysis. Since he also includes his projections for fpts and fpts/game, I can sneek a peek over his shoulder to see how closely my stars allign with his. The second last step is to separate my projections by tier breaks where it is logical to do so. Finally, a day or two before my actual draft, I add in very current and reliable ADPs along side of the projections to determine several value targets for each round of the draft.

 

EDIT: My computer guy was not able to tell me what corrupted my fantasy football files, but he WAS able to restore them.

 

Here is an example of what the (very easy team to project) BALT Team Projections look like with my individual Player Projections incorporated in them:

 

Balt.... Pass/Rec Yd.............Ints....Comb TDs.....Compl/Recpts......Rush Yds

Flacco............. 3400............. 14............ 22...............300...................80

other.................. 150............. 0.................8.................0

QB Tot.............3550............. 14.............22...............308...................80

Rice...................630...............................13................70................1310

other.....................60...............................2.................14.................490

RB Tot................690..............................15................84................1800

T Smith...............980...............................7.................62...................20

Boldin..................830..............................5.................58...................20

other....................234..............................1.................23

WR Tot..............2044.............................13...............143...................40

Dickson...............444..............................4..................44

Pitta....................370..............................3..................37

TE Tot.................814..............................7..................81

 

Tot.....................3550.............................35...............308................1920

Rushing Yds.......1920

Tot Offense.........5470

 

The process is not all that easy to describe in words, but it is actually not that difficult to do with a little practice. In fact, by doing them in this way, I actually take several shortcuts from the way I've seen others do theirs, but IMO the shortcuts do not detract from the finished product. Please feel free to chime in with criticisms, comments, questions, or suggestions.

[The Football Guru 211]

Seriously, I may need to hire you as my PR guy Now if I could only get Sara to mention me a bit more on her website and posts

 

Quick thought about the post. Just like MFM suggests, projections are constantly evolving throughout the preseason. In fact, I'll be starting mine some time next week in all likelihood and expect to tweak them continuously until the beginning of the season. One reason I use my the method I do is to keep my projections in check. For example, it is easy to say that Trent Richardson could be a 1400-1500 yard rusher this season, but unless I can justify giving him 1-2 100-yard games over the course of the five division games he'll play during the fantasy season, then that number is probably too optimistic.

 

I think one key (there are many) to good projecting is not looking at or comparing your final numbers for one player to another player until you have finished your projections. I find this helps me remove bias and give some good thought about whether Player A really is a "lock" to outperform Player B. This isn't to say I don't tweak the numbers a bit, but if the matchups steer me in the direction of a stud struggling and an under-the-radar player excelling, I don't fight it.

[Frozenbeernuts 1,652]

I found an easier way to project points for players. I am more of a gut feel guy but do like to incorporate stats into my preperation. I am much better at projecting the order in which players will finish than projecting their actual stat line.

 

I take the past 3 years of fantasy production from each position. I add up each players stats and convert them to my leagues scoring rules. When i add them up though i do assign names just numbers. I compile the list going down until a player reaches about 120 points. Then i assemble them in numerical order since scoring changes change the order.

 

After I have my list of of the past 3 years for every position I separate each position and write in a row the past 3 years top performer, then second, third, so on.... 375 345 349. I then add them up and average them. The average is what I will assign to the player in the correaponding spot on my rankings. It gives a pretty good idea of what your player will score if they end up in the spot you think they will.

 

After all that is done I do some tinkering to stats that do not seem to match the player they were assigned to. It really gave me some perspective on how many points the 10th best rb will score compared to the tenth best wr. Of course the system will not account for those dream seasons from players like moss, brady, LT, but no one projects a receiver to score 24 td anyway.

 

The one hiccup i had though was with Qbs this year. They had such a spike in numbers that I just took last years stats and rolled with that. I may be wrong because it was more a product of the lockout, but we all know how the league is much more pass oriented. It kind of reminded me that yeah there are a bunch of rb this year that could be studs, but history says there will only be a few, no matter how many you think there will be.

[madd futher mucker 36]

I found an easier way to project points for players. I am more of a gut feel guy but do like to incorporate stats into my preperation. I am much better at projecting the order in which players will finish than projecting their actual stat line.

 

I take the past 3 years of fantasy production from each position. I add up each players stats and convert them to my leagues scoring rules. When i add them up though i do assign names just numbers. I compile the list going down until a player reaches about 120 points. Then i assemble them in numerical order since scoring changes change the order.

 

After I have my list of of the past 3 years for every position I separate each position and write in a row the past 3 years top performer, then second, third, so on.... 375 345 349. I then add them up and average them. The average is what I will assign to the player in the correaponding spot on my rankings. It gives a pretty good idea of what your player will score if they end up in the spot you think they will.

 

After all that is done I do some tinkering to stats that do not seem to match the player they were assigned to. It really gave me some perspective on how many points the 10th best rb will score compared to the tenth best wr. Of course the system will not account for those dream seasons from players like moss, brady, LT, but no one projects a receiver to score 24 td anyway.

 

The one hiccup i had though was with Qbs this year. They had such a spike in numbers that I just took last years stats and rolled with that. I may be wrong because it was more a product of the lockout, but we all know how the league is much more pass oriented. It kind of reminded me that yeah there are a bunch of rb this year that could be studs, but history says there will only be a few, no matter how many you think there will be.

 

Sorry, maybe I'm missing something but I just can't see much value to what you are doing. IMO, merely averaging historical production will not lead to a data base that will yield decent results come your draft day. You are probably better off just ranking players according to your 'gut' unless there is more to your 'projections' than what you described.

[Frozenbeernuts 1,652]

Sorry, maybe I'm missing something but I just can't see much value to what you are doing. IMO, merely averaging historical production will not lead to a data base that will yield decent results come your draft day. You are probably better off just ranking players according to your 'gut' unless there is more to your 'projections' than what you described.

I rank guys and historical averages will give me a good idea how much they will score. Like i am high on a bunch of receivers but all of them cant score 350 points. So my 12th ranked receiver is going to score around 235 points. I dont have my rankings infront of me but i think that is somewhere close. I corelate the historical 12th wr with my 12th wr