FBG Predictions Vs. Actuals (1 Viewer)

[bagger.]

If you're serious about winning... you should just concentrate on making your own correct and not worrying about FBG's projections.

<{POST_SNAPBACK}>

 

[LHUCKS]

FF Index does a quantified comparison every year of 10/12 sites' rankings...FBG is one of the parties, so your best bet is to look there. 

The value of FBG doesn't lie in the projections, it's the tools it provides.  Ask any of the sharks around here, almost all of them use their own projections, but pay for the subscription for the tools/information.

I suppose if you're a novice and you're drafting straight from projections that is where you could find value here at FBG, but if you're serious about winning wouldn't you want a second opinion from another source anyhow?

<{POST_SNAPBACK}>

If you're serious about winning... you should just concentrate on making your own correct and not worrying about FBG's projections.
<{POST_SNAPBACK}>

Agreed, I obviously do that, but was speaking to those that don't have enough time or interest to do their own.

 

[Joe T]

Thanks FFL.  Yeah if you look at the top 10 WR.  Ouch but there are reasons for that.  If we could have the Brass analyze it and make a feature of it, it would be awesome.

<{POST_SNAPBACK}>

I don't think Brass Sack wants to do all that work.
<{POST_SNAPBACK}>

GB #907

 

[tytyty]

Thanks for the amazing input Joe T. I am completely at the mercy of FBG. I don't have my own thought patterns and rely solely on them.

Anyway still would love to see it Dodds! Thanks

 

[Joe T]

The main problem with this site as I see it is too many people fall into the group think or herd mentality. Not enough people question the rankings... they all ### pat each other and talk about how good the rankings are.

Every year, I come on here and question why player X is ranked higher than player Y over and over. Does it cause some grief? Hell yes. And it should. We should be ripping the rankings and projections to shreds making sure that all the logic is in line and that every ranking can be defended... not just haphazardly assumed to be correct.

We should have been asking (and did ask in some cases):

Why is McGahee ranked so high?

Why is Burleson higher than Steve Smith when he has done far less?

Why is Burleson much higher than Donald Driver when he has done far less?

Why is Michael Clayton so high?

People should be forced to defend their rankings... and when they are wrong, they should change them. This part rarely happens around here, but it should.

You are more likely to prove someone wrong around here and then have them blindly defend it all year rather than admit they were wrong.

 

[bagger.]

The main problem with this site as I see it is too many people fall into the group think or herd mentality.  Not enough people question the rankings... they all ### pat each other and talk about how good the rankings are.

Every year, I come on here and question why player X is ranked higher than player Y over and over.  Does it cause some grief?  Hell yes.  And it should.  We should be ripping the rankings and projections to shreds making sure that all the logic is in line and that every ranking can be defended... not just haphazardly assumed to be correct.

We should have been asking (and did ask in some cases):

Why is McGahee ranked so high?

Why is Burleson higher than Steve Smith when he has done far less?

Why is Burleson much higher than Donald Driver when he has done far less?

Why is Michael Clayton so high?

People should be forced to defend their rankings... and when they are wrong, they should change them.  This part rarely happens around here, but it should.

You are more likely to prove someone wrong around here and then have them blindly defend it all year rather than admit they were wrong.

<{POST_SNAPBACK}>

After reading that God Bless America started playing in my head.

 

[lebowski]

The main problem with this site as I see it is too many people fall into the group think or herd mentality.  Not enough people question the rankings... they all ### pat each other and talk about how good the rankings are.

Every year, I come on here and question why player X is ranked higher than player Y over and over.  Does it cause some grief?  Hell yes.  And it should.  We should be ripping the rankings and projections to shreds making sure that all the logic is in line and that every ranking can be defended... not just haphazardly assumed to be correct.

We should have been asking (and did ask in some cases):

Why is McGahee ranked so high?

Why is Burleson higher than Steve Smith when he has done far less?

Why is Burleson much higher than Donald Driver when he has done far less?

Why is Michael Clayton so high?

People should be forced to defend their rankings... and when they are wrong, they should change them.  This part rarely happens around here, but it should.

You are more likely to prove someone wrong around here and then have them blindly defend it all year rather than admit they were wrong.

<{POST_SNAPBACK}>

I"m personally more interested in why players are ranked a certain way.

 

[Mark Wimer]

Let me jump in here, as I've been around the FBG's staff/approach for a long time and I have some familiarity with the official projections vs. staff rankings.

The official projections are handled by David Dodds, with input from Joe Bryant (and others) based on David's extensive NFL database and their many, many years of experience in so doing. Members of the staff have input into David's paradigm - it's not a monolithic thing, and David is willing to hear valid criticism/argument about his projections. Behind the scenes, all of the staff have an ongoing dialogue about NFL players/topics that we all use to inform our rankings/projections, via email, phone calls, etc. Just last week, I called up Clayton Gray to talk about Brett Favre, for example. We all network and seek each other's input.

The staff members maintain their own set of projections/rankings, totally independent from David's work. I myself (cannot speak for all the staff, but only speaking to my personal practice) maintain projections for all likely fantasy starters/top backups/benchwarmers in a standard (not IDP) league, for publication to the site. In fact, David and I have been known to disagree radically from time to time on particular players - often, I'm "wrong" and he's "right". Sometimes, though, I make a solid call/analysis that David didn't see coming down the pike.

Bottom line - Dodds and Bryant do not hesitate to call out a staff member if they think their projections/rankings are out of line, and they demand accountability. Trust me, it is no fun to be on the hotseat with David Dodds breathing down your neck.

The strength of Footballguys.com is that David and Joe do not insist on a "group think" structure where every staff member has to drink the Kool-Aid and toe the line when it comes to projections/rankings. We are invited, but not compelled, to participate in the various rankings/site features like faceoffs; spotlights; player rankings, etc.

So, ultimately the question comes down to - what projections do you want to compare? Shall we do an analysis of every staff member's final rankings vs. the NFL actuals, and then eliminate all the "injury issues" like Javon Walker 2005? We put out 162 customized cheat sheets for various leagues during the pre-season draft rush 162 cheatsheets - shall we do an analysis of how close all 162 cheatsheets came out vs. NFL actual, given all 162 scoring paradigms covered?

I think the question being posed in this thread is much more complex than it appears on first glance, due to the broad-based, wide-ranging approach here at Footballguys.

My .02.

 

[David Yudkin]

The main problem with this site as I see it is too many people fall into the group think or herd mentality.  Not enough people question the rankings... they all ### pat each other and talk about how good the rankings are.

Every year, I come on here and question why player X is ranked higher than player Y over and over.  Does it cause some grief?  Hell yes.  And it should.  We should be ripping the rankings and projections to shreds making sure that all the logic is in line and that every ranking can be defended... not just haphazardly assumed to be correct.

We should have been asking (and did ask in some cases):

Why is McGahee ranked so high?

Why is Burleson higher than Steve Smith when he has done far less?

Why is Burleson much higher than Donald Driver when he has done far less?

Why is Michael Clayton so high?

People should be forced to defend their rankings... and when they are wrong, they should change them.  This part rarely happens around here, but it should.

You are more likely to prove someone wrong around here and then have them blindly defend it all year rather than admit they were wrong.

<{POST_SNAPBACK}>

Here is a major issue that I have and one that there is no right or wrong answer to. I generally rank guys a fair amount different than "the herd." The problem is, if I post my rankings (which I really have never done but will consider this year) is that when I have revealed some of them, I have been hit with a barrage of PMs, posts, and emails demanding an explanation.While I understand the inquisitiveness of people, to have to justify EVERY ranking and provide an analysis of each guy, while potentially valuable to the masses, is effectively busy work for staff.

I have had threads addressed specifically AT ME over the years that went on FOR PAGES AND PAGES on my perception of players, situations, predictions, projections, etc. Many times, the time required to appease the masses took 10 times as long as the time involved in putting together the initial information.

As an example, it may take you 10 minutes to drive to work. How long would it take to walk instead? That's essentially what people ask sometimes, and to be available to 65,000 people is not practical most times.

Since so many people are barking to "defend rankings," I for one do not want to embark on the journey to have to justify each and every pick. If you guys want to debate them, I say more power to you, but to expect staff to have to provide a thorough explanation IMO is asking for too much.

I agree that if some staff want to answer an occassinal question or chime in on a thread once in a while, great, but we really cannot be at the beck and call of everyone that needs us.

 

[bagger.]

So, ultimately the question comes down to - what projections do you want to compare? Shall we do an analysis of every staff member's final rankings vs. the NFL actuals, and then eliminate all the "injury issues" like Javon Walker 2005? We put out 162 customized cheat sheets for various leagues during the pre-season draft rush 162 cheatsheets  - shall we do an analysis of how close all 162 cheatsheets came out vs. NFL actual, given all 162 scoring paradigms covered?

I think the question being posed in this thread is much more complex than it appears on first glance, due to the broad-based, wide-ranging approach here at Footballguys.

My .02.

<{POST_SNAPBACK}>

I'd love to see the variance between any staff member's projections and actuals as well as the consensus version. Maybe Chase projects TEs better than others. Maybe David's overall projections are better.I understand it isn't easy, but I would rather see resources put in there than some other fluff that the site has. But again that comes down to personal preference of what one finds value in and another does not.

I no longer subscribe to FBG but donate to Drinen's site (profootball-reference) as I find I use that infinitely more useful than FBG content. Others would be in the opposite camp.

Maybe you can't do this before this year, but there are many technical people here who are not looking for a scape goat or to point fingers but to see how projections differ and how everyone can improve.

Projections will always be off. It is just a matter of scale.

If Staff member X continuously projects better for RBs than the rest I will look at that staff's rankings closer in the next year.

 

[LHUCKS]

It should be clarified as to what projections/rankings we're talking about.

Here's the following different groups/projections rankings.

A) Dodds/Bryant FBG Projections

B) staff consensus rankings, consisting of individual staff rankings

C) messageboard rankings

I think we're talking about A, but the "herd mentality" comment seems to apply to B and/or C.

Just an observation.

 

[Iron_Far]

The main problem with this site as I see it is too many people fall into the group think or herd mentality.  Not enough people question the rankings... they all ### pat each other and talk about how good the rankings are.

Every year, I come on here and question why player X is ranked higher than player Y over and over.  Does it cause some grief?  Hell yes.  And it should.  We should be ripping the rankings and projections to shreds making sure that all the logic is in line and that every ranking can be defended... not just haphazardly assumed to be correct.

We should have been asking (and did ask in some cases):

Why is McGahee ranked so high?

Why is Burleson higher than Steve Smith when he has done far less?

Why is Burleson much higher than Donald Driver when he has done far less?

Why is Michael Clayton so high?

People should be forced to defend their rankings... and when they are wrong, they should change them.  This part rarely happens around here, but it should.

You are more likely to prove someone wrong around here and then have them blindly defend it all year rather than admit they were wrong.

<{POST_SNAPBACK}>

 

[Kleck]

Instead of a general call for something from no one in particular, you'll do much better contacting David Dodds specifically. He is ultimately responsible for the successes and failure of the projections I think you're speaking of.

There are no "group think" projections that represent the entire site. We're all encouraged to post our rankings and be prepared to defend them. I always look forward to my "Face Off" assignments after the rankings are first posted. I was very wrong about Andre Johnson last season (highest ranking of the staff), but very right about Steve Smith and Joey Galloway last season (highest ranking of the staff).

 

[tytyty]

Thanks for coming in. And thanks for the insight. The only thing I am looking for is a comparison of the raw numbers. How a person tweaks it to any of 162 scoring variations is up to them. I just want to see a basic look at it using any one normal fantasy scoring system that FBG guys use when projecting FP. If manning is projected 4000 yds and 48 TD's and he actually does 3300 and 39, no matter what scoring system he will have done worse then the projections. So for sake of argument, maybe the rankings before and after are based on FP's which FBG uses. So if Manning in that system would of had, based on projections, 380 FP, and actual would have been 340, that would show a -40. Thats all

 

[David Yudkin]

I was very wrong about Andre Johnson last season (highest ranking of the staff), but very right about Steve Smith and Joey Galloway last season (highest ranking of the POSTED BY staff).

<{POST_SNAPBACK}>

Fixed. I had Smith at WR 5 by the time the season was starting and IIRC had Galloway in the Top 15 (along with Glenn in the Top 20). I also had AJ as the #30 WR. See my comments above for why I don't post rankings. How many people would have called me out if I posted that in the staff rankings?

 

[LHUCKS]

I agree that if some staff want to answer an occassinal question or chime in on a thread once in a while, great, but we really cannot be at the beck and call of everyone that needs us.

<{POST_SNAPBACK}>

And I might add that every year I put out a consensus rankings critique and I get comments defending or arguing against positions concerning specific players from over half of the staff.

 

[BroadwayG]

This annual thread gets more ridiculous each year.

 

[David Yudkin]

I agree that if some staff want to answer an occassinal question or chime in on a thread once in a while, great, but we really cannot be at the beck and call of everyone that needs us.

<{POST_SNAPBACK}>

And I might add that every year I put out a consensus rankings critique and I get comments defending or arguing positions concerning specific players from over half of the staff.
<{POST_SNAPBACK}>

To be clear, I think we all (regular fols and staff) alike appreciate what you post and the rankings and commentary you provide to the site. It's a valuable asset and one that often goes unheralded (and comes free).But we staff also have a ton of other things to do for the site and still need to complete those assignments in addition to (most of us) having other jobs to do and families to tend to.

If staff guys have 14 articles to get out by Tuesday, getting embroiled in a 3 hour debate on McGahee with Joe T is counter productive (even though I had Willis ranked at 15 last year).

 

[Maurile Tremblay]

I was very wrong about Andre Johnson last season (highest ranking of the staff), but very right about Steve Smith and Joey Galloway last season (highest ranking of the POSTED BY staff).

<{POST_SNAPBACK}>

Fixed. I had Smith at WR 5 by the time the season was starting and IIRC had Galloway in the Top 15 (along with Glenn in the Top 20). I also had AJ as the #30 WR. See my comments above for why I don't post rankings. How many people would have called me out if I posted that in the staff rankings?
<{POST_SNAPBACK}>

Doesn't count if you don't post them.

 

[Wade]

The strength of Footballguys.com is that David and Joe do not insist on a "group think" structure where every staff member has to drink the Kool-Aid and toe the line when it comes to projections/rankings.

<{POST_SNAPBACK}>

I think by that he means they don't have to cram into the low-rise jeans....

 

[Maurile Tremblay]

I agree that if some staff want to answer an occassinal question or chime in on a thread once in a while, great, but we really cannot be at the beck and call of everyone that needs us.

<{POST_SNAPBACK}>

And I might add that every year I put out a consensus rankings critique and I get comments defending or arguing against positions concerning specific players from over half of the staff.
<{POST_SNAPBACK}>

These have proven to be helpful in the past. I remember, for example, some great discussion about Muhsin Muhammad before the 2004 season in that thread.

 

[LHUCKS]

Thanks for coming in.  And thanks for the insight.  The only thing I am looking for is a comparison of the raw numbers.  How a person tweaks it to any of 162 scoring variations is up to them.  I just want to see a basic look at it using any one normal fantasy scoring system that FBG guys use when projecting FP.  If manning is projected 4000 yds and 48 TD's and he actually does 3300 and 39, no matter what scoring system he will have done worse then the projections.  So for sake of argument, maybe the rankings before and after are based on FP's which FBG uses.  So if Manning in that system would of had, based on projections, 380 FP, and actual would have been 340, that would show a -40.  Thats all

<{POST_SNAPBACK}>

what good does this do, unless you have projections for the exact same format from another source?Would you then go, "oh Site A is better than Site B?"...even then you would need several years of data to support such a claim.

The point is this, the cost/benefit analysis for conducting such a study just doesn't support that it's time well spent.

 

[David Yudkin]

I was very wrong about Andre Johnson last season (highest ranking of the staff), but very right about Steve Smith and Joey Galloway last season (highest ranking of the POSTED BY staff).

<{POST_SNAPBACK}>

Fixed. I had Smith at WR 5 by the time the season was starting and IIRC had Galloway in the Top 15 (along with Glenn in the Top 20). I also had AJ as the #30 WR. See my comments above for why I don't post rankings. How many people would have called me out if I posted that in the staff rankings?
<{POST_SNAPBACK}>

Doesn't count if you don't post them.
<{POST_SNAPBACK}>

I will be happy to post rankings this year if no one gets annoyed (or offers up death threats) if I completely ignore ALL emails, PMs, and threads asking me to explain them.And while I never posted the complete rankings I had, I did offer up where I ranked many guys in threads for months and months leading up to the season, so it's not like I am taking credit after the fact.

In actuality, I am not even looking for any credit at all. I put together what I did for my own use. If others want to look at that info, fine, but I never felt I was "competing" with other people's rankings.

 

[Joe T]

The main problem with this site as I see it is too many people fall into the group think or herd mentality.  Not enough people question the rankings... they all ### pat each other and talk about how good the rankings are.

Every year, I come on here and question why player X is ranked higher than player Y over and over.  Does it cause some grief?  Hell yes.  And it should.  We should be ripping the rankings and projections to shreds making sure that all the logic is in line and that every ranking can be defended... not just haphazardly assumed to be correct.

We should have been asking (and did ask in some cases):

Why is McGahee ranked so high?

Why is Burleson higher than Steve Smith when he has done far less?

Why is Burleson much higher than Donald Driver when he has done far less?

Why is Michael Clayton so high?

People should be forced to defend their rankings... and when they are wrong, they should change them.  This part rarely happens around here, but it should.

You are more likely to prove someone wrong around here and then have them blindly defend it all year rather than admit they were wrong.

<{POST_SNAPBACK}>

Here is a major issue that I have and one that there is no right or wrong answer to. I generally rank guys a fair amount different than "the herd." The problem is, if I post my rankings (which I really have never done but will consider this year) is that when I have revealed some of them, I have been hit with a barrage of PMs, posts, and emails demanding an explanation.While I understand the inquisitiveness of people, to have to justify EVERY ranking and provide an analysis of each guy, while potentially valuable to the masses, is effectively busy work for staff.

I have had threads addressed specifically AT ME over the years that went on FOR PAGES AND PAGES on my perception of players, situations, predictions, projections, etc. Many times, the time required to appease the masses took 10 times as long as the time involved in putting together the initial information.

As an example, it may take you 10 minutes to drive to work. How long would it take to walk instead? That's essentially what people ask sometimes, and to be available to 65,000 people is not practical most times.

Since so many people are barking to "defend rankings," I for one do not want to embark on the journey to have to justify each and every pick. If you guys want to debate them, I say more power to you, but to expect staff to have to provide a thorough explanation IMO is asking for too much.

I agree that if some staff want to answer an occassinal question or chime in on a thread once in a while, great, but we really cannot be at the beck and call of everyone that needs us.

<{POST_SNAPBACK}>

I probably wasn't very clear. I'm not suggesting you or any staff member defend your individual rankings.I'm suggesting that the board should always be questioning the logic of the staff rankings with parts of the board defending the rankings they see fit to defend and parts of the board ripping the logic to shreds.

I understand that in no way could any staff member be expected to defend every ranking.

My suggestion is that 98% of the people on here accept the rankings/projections without questioning them when ALL of those people should be ripping/defending/arguing what goes into the logic behind them.

 

[LHUCKS]

To be clear, I think we all (regular fols and staff) alike appreciate what you post and the rankings and commentary you provide to the site.  It's a valuable asset and one that often goes unheralded (and comes free).

But we staff also have a ton of other things to do for the site and still need to complete those assignments in addition to (most of us) having other jobs to do and families to tend to.

If staff guys have 14 articles to get out by Tuesday, getting embroiled in a 3 hour debate on McGahee with Joe T is counter productive (even though I had Willis ranked at 15 last year).

<{POST_SNAPBACK}>

I think you misunderstood my intent David. I was merely pointing out that if you want the staff to back up their rankings with discussion, the staff has shown in the past that it is willing to do so. That was my only point. I don't expect any interaction from the staff myself, I do it more for my own analysis and to create conversation on the messageboard...which I enjoy obviously.

 

[Joe T]

If staff guys have 14 articles to get out by Tuesday, getting embroiled in a 3 hour debate on McGahee with Joe T is counter productive (even though I had Willis ranked at 15 last year).

<{POST_SNAPBACK}>

This is where we disagree.I'm not looking for a 3 hour debate with you or anyone specifically. I'm looking for a rationalization of the logic of his overly inflated ranking with the board in general. I'm trying to open the eyes of the masses that sometimes the experts don't spend enough time on every ranking.

I think a lot of the staff gets offended when their logic is questioned. They shouldn't be. They should take pride in the fact that someone is taking the time to analyze their rankings good or bad.

When the board hashes out the "bad" or "illogical" rankings, everyone wins. Even the staff... although I think most times they fail to see this.

 

[Kleck]

I think a lot of the staff gets offended when their logic is questioned.

<{POST_SNAPBACK}>

You should be specific here. Who and how many are "a lot"?

 

[fatness]

If the rankings are available from last year and people on the message board want to analyze them, they should. I don't see how it's helpful for staff to take time away from what they're doing for this year to analyze last year. I do see how it's helpful for interested people on the message board to do it.

The best use of time seems to be:

Staff: keep the new stuff coming

Message Board: analyze the previous year's stuff

Both: see where that analysis leads

A cooperative effort like that would work better than "You should do it" and "We don't have time".

 

[JayMan]

Thank you David for posting up your comments!

I just think that it is interesting to look at last year's projections (from FBG or your own) in order to see trends that we might be missing...

1. Are we overevaluating the dropoff of RBs in their early 30s?... probably projecting lower values for Tiki... what does that mean for this year, CMart, Dillon?...

2. Are we underestimating the value of good 'ol vet WRs and reaching for top young guns?... projecting better numbers for Michael Clayton than Derrick Mason for example... in 2006, does this mean that we'll overvalue Reggie Brown versus Keyshan?...

I think this kind of analysis are interesting... and I don't think we want to look at last year's FBG projections to say... "wow, they really missed on Javon and Willis"... everybody had them in their respective top10-position wise last year!!

I want to look at past projections to be able to gauge where certains players might go... If I want Mason (the good 'ol vet WR) on my team - I want to know if I'll need to get him with my 7th, 9th or 11th pick... and projections/results help in that manner...

My 2 cents...

 

[Joe T]

I think a lot of the staff gets offended when their logic is questioned.

<{POST_SNAPBACK}>

You should be specific here. Who and how many are "a lot"?
<{POST_SNAPBACK}>

2?

 

[Family Matters]

Thank you David for posting up your comments!

I just think that it is interesting to look at last year's projections (from FBG or your own) in order to see trends that we might be missing...

It'll be difficult to detect trends and carry them over form 1 year to another. You will do more harm than good if you trry that.

1. Are we overevaluating the dropoff of RBs in their early 30s?... probably projecting lower values for Tiki... what does that mean for this year, CMart, Dillon?...

Again, what happens one year versus another is like trying to see if the same wheather patterns will prevail from 1 year to the next. Sure we know it's going to be warm in the summer but trying to figure which day because of what happened last year is not realistic.

2. Are we underestimating the value of good 'ol vet WRs and reaching for top young guns?... projecting better numbers for Michael Clayton than Derrick Mason for example... in 2006, does this mean that we'll overvalue Reggie Brown versus Keyshan?...

Same as above. But let me add that also depends on who you draft against, scoring rules and so on. Some leagues value positions and player types differently than others. How you interprelate that knowledge is where the difference comes in.

I think this kind of analysis are interesting... and I don't think we want to look at last year's FBG projections to say... "wow, they really missed on Javon and Willis"... everybody had them in their respective top10-position wise last year!!

It's interesting but it can be very misleading if you use old information, trends and analysis from 1 year to help predict what will happen the next.

I want to look at past projections to be able to gauge where certains players might go... If I want Mason (the good 'ol vet WR) on my team - I want to know if I'll need to get him with my 7th, 9th or 11th pick... and projections/results help in that manner...

Again, what happened last year with player x might be totally different than this year. E.g., Dillon will go much later in most leagues than he did last year. If you use last year's data you will likely overpay for him.

My 2 cents...

<{POST_SNAPBACK}>

JayMan-I don't mean to come across as being against what you say but it's likely going to sound that way. So my apologies. Just look at it for the info it's intended.My comments are in blue. It just seems clear to me that this type of info while interesting is mostly useless for the next FF year.

 

[Fullback Fro]

Drew Bledsoe will finish top 5.

 

[Kleck]

I think a lot of the staff gets offended when their logic is questioned.

<{POST_SNAPBACK}>

You should be specific here. Who and how many are "a lot"?
<{POST_SNAPBACK}>

2?
<{POST_SNAPBACK}>

Okey Dokey. I'm not disputing your claim. I've seen staff members handle criticism poorly. Just don't want you to blanket the majority with a negative comment when it should be directed at a specific person.

 

[Aaron Rudnicki]

any suggestions on how one would go about making this comparison? in other words, how would you evaluate the accuracy of the projections?

I don't have a clear idea of the best way to do this. If you're just looking for someone to give you last year's projections so you can compare them against the actual numbers, that's a much easier request than actually asking for someone to do the analysis/comparison.

 

[JayMan]

Thank you David for posting up your comments!

I just think that it is interesting to look at last year's projections (from FBG or your own) in order to see trends that we might be missing...

It'll be difficult to detect trends and carry them over form 1 year to another. You will do more harm than good if you trry that.

1. Are we overevaluating the dropoff of RBs in their early 30s?... probably projecting lower values for Tiki... what does that mean for this year, CMart, Dillon?...

Again, what happens one year versus another is like trying to see if the same wheather patterns will prevail from 1 year to the next. Sure we know it's going to be warm in the summer but trying to figure which day because of what happened last year is not realistic.

2. Are we underestimating the value of good 'ol vet WRs and reaching for top young guns?... projecting better numbers for Michael Clayton than Derrick Mason for example... in 2006, does this mean that we'll overvalue Reggie Brown versus Keyshan?...

Same as above. But let me add that also depends on who you draft against, scoring rules and so on. Some leagues value positions and player types differently than others. How you interprelate that knowledge is where the difference comes in.

I think this kind of analysis are interesting... and I don't think we want to look at last year's FBG projections to say... "wow, they really missed on Javon and Willis"... everybody had them in their respective top10-position wise last year!!

It's interesting but it can be very misleading if you use old information, trends and analysis from 1 year to help predict what will happen the next.

I want to look at past projections to be able to gauge where certains players might go... If I want Mason (the good 'ol vet WR) on my team - I want to know if I'll need to get him with my 7th, 9th or 11th pick... and projections/results help in that manner...

Again, what happened last year with player x might be totally different than this year. E.g., Dillon will go much later in most leagues than he did last year. If you use last year's data you will likely overpay for him.

My 2 cents...

<{POST_SNAPBACK}>

JayMan-I don't mean to come across as being against what you say but it's likely going to sound that way. So my apologies. Just look at it for the info it's intended.My comments are in blue. It just seems clear to me that this type of info while interesting is mostly useless for the next FF year.

<{POST_SNAPBACK}>

Point taken - and I can see where you are coming from... but I try to look at it in a different perspective...I'm not a big fan of "what happen last(s) year(s) will surely repeat itself this year" kind of thinking... I'm just interested in trends that are consistently being used in FF world...

I'm guessing that it's professional deviation! since I'm a stock analyst and use technical analysis, along with trend analysis and fundementals in order to pick sectors/stocks that we want to invest in daily...

I'm taking that same approach with FF - knowing that the "3rd year WR" theory is way overvalued and some guys like Brandon Lloyd were projected that way last year...

I'm just using those trends, along with the expert projections, to try and gauge when I should be targeting this or that player...

I might be putting too much effort into this... and I can value your point of view...

 

[glock]

any suggestions on how one would go about making this comparison? in other words, how would you evaluate the accuracy of the projections?

I don't have a clear idea of the best way to do this. If you're just looking for someone to give you last year's projections so you can compare them against the actual numbers, that's a much easier request than actually asking for someone to do the analysis/comparison.

<{POST_SNAPBACK}>

How 'bout projected vs actual year end, leave the analysis to the individual?

 

[Aaron Rudnicki]

any suggestions on how one would go about making this comparison? in other words, how would you evaluate the accuracy of the projections?

I don't have a clear idea of the best way to do this. If you're just looking for someone to give you last year's projections so you can compare them against the actual numbers, that's a much easier request than actually asking for someone to do the analysis/comparison.

<{POST_SNAPBACK}>

How 'bout projected vs actual year end, leave the analysis to the individual?

<{POST_SNAPBACK}>

seems reasonable. It's not like the projections are a secret. I sent Dodds an email to check this thread, so I'm sure he'll check in and reply soon.

 

[cashman88]

This thread has gotten away from the relatively simple request made at the beginning of the thread. That is, to "have an Article . . . that takes the Projections from last year and compares it to the End of Year Actual Stats". I would like to see FBG do this for Dodds projections, as well as the projections posted by staff members -- it can be a competition between them each year.

I, too am a FBG subscriber and think it would be interesting to read.

 

[chedha]

Thank you David for posting up your comments!

I just think that it is interesting to look at last year's projections (from FBG or your own) in order to see trends that we might be missing...

It'll be difficult to detect trends and carry them over form 1 year to another. You will do more harm than good if you trry that.

1. Are we overevaluating the dropoff of RBs in their early 30s?... probably projecting lower values for Tiki... what does that mean for this year, CMart, Dillon?...

Again, what happens one year versus another is like trying to see if the same wheather patterns will prevail from 1 year to the next. Sure we know it's going to be warm in the summer but trying to figure which day because of what happened last year is not realistic.

2. Are we underestimating the value of good 'ol vet WRs and reaching for top young guns?... projecting better numbers for Michael Clayton than Derrick Mason for example... in 2006, does this mean that we'll overvalue Reggie Brown versus Keyshan?...

Same as above. But let me add that also depends on who you draft against, scoring rules and so on. Some leagues value positions and player types differently than others. How you interprelate that knowledge is where the difference comes in.

I think this kind of analysis are interesting... and I don't think we want to look at last year's FBG projections to say... "wow, they really missed on Javon and Willis"... everybody had them in their respective top10-position wise last year!!

It's interesting but it can be very misleading if you use old information, trends and analysis from 1 year to help predict what will happen the next.

I want to look at past projections to be able to gauge where certains players might go... If I want Mason (the good 'ol vet WR) on my team - I want to know if I'll need to get him with my 7th, 9th or 11th pick... and projections/results help in that manner...

Again, what happened last year with player x might be totally different than this year. E.g., Dillon will go much later in most leagues than he did last year. If you use last year's data you will likely overpay for him.

My 2 cents...

<{POST_SNAPBACK}>

JayMan-I don't mean to come across as being against what you say but it's likely going to sound that way. So my apologies. Just look at it for the info it's intended.My comments are in blue. It just seems clear to me that this type of info while interesting is mostly useless for the next FF year.

<{POST_SNAPBACK}>

Family Matters - I think you've got a point when saying the type of information can be useless for the next FF year. But as you just stated, it would be someting interesting to look at, so my point is, why not publish it? It's just information that can be referenced while doing research/analysis if you want to... I don't see what the harm would be in releasing just a simple comparison on what the original FBG projection was before the start of the season, and what the NFL actuals were at the end of the season, with a section to comment on why x player did or did not perform as expected. I understand there are many variables involved in such a comparison, but that can be noted in the comments section..

 

[Family Matters]

Thank you David for posting up your comments!

I just think that it is interesting to look at last year's projections (from FBG or your own) in order to see trends that we might be missing...

It'll be difficult to detect trends and carry them over form 1 year to another. You will do more harm than good if you trry that.

1. Are we overevaluating the dropoff of RBs in their early 30s?... probably projecting lower values for Tiki... what does that mean for this year, CMart, Dillon?...

Again, what happens one year versus another is like trying to see if the same wheather patterns will prevail from 1 year to the next. Sure we know it's going to be warm in the summer but trying to figure which day because of what happened last year is not realistic.

2. Are we underestimating the value of good 'ol vet WRs and reaching for top young guns?... projecting better numbers for Michael Clayton than Derrick Mason for example... in 2006, does this mean that we'll overvalue Reggie Brown versus Keyshan?...

Same as above. But let me add that also depends on who you draft against, scoring rules and so on. Some leagues value positions and player types differently than others. How you interprelate that knowledge is where the difference comes in.

I think this kind of analysis are interesting... and I don't think we want to look at last year's FBG projections to say... "wow, they really missed on Javon and Willis"... everybody had them in their respective top10-position wise last year!!

It's interesting but it can be very misleading if you use old information, trends and analysis from 1 year to help predict what will happen the next.

I want to look at past projections to be able to gauge where certains players might go... If I want Mason (the good 'ol vet WR) on my team - I want to know if I'll need to get him with my 7th, 9th or 11th pick... and projections/results help in that manner...

Again, what happened last year with player x might be totally different than this year. E.g., Dillon will go much later in most leagues than he did last year. If you use last year's data you will likely overpay for him.

My 2 cents...

<{POST_SNAPBACK}>

JayMan-I don't mean to come across as being against what you say but it's likely going to sound that way. So my apologies. Just look at it for the info it's intended.My comments are in blue. It just seems clear to me that this type of info while interesting is mostly useless for the next FF year.

<{POST_SNAPBACK}>

Family Matters - I think you've got a point when saying the type of information can be useless for the next FF year. But as you just stated, it would be someting interesting to look at, so my point is, why not publish it? It's just information that can be referenced while doing research/analysis if you want to... I don't see what the harm would be in releasing just a simple comparison on what the original FBG projection was before the start of the season, and what the NFL actuals were at the end of the season, with a section to comment on why x player did or did not perform as expected. I understand there are many variables involved in such a comparison, but that can be noted in the comments section..
<{POST_SNAPBACK}>

I understand and there wouldn't be any harm at all but I doubt it would be simple to produce. It would take some serious man hours to put it together and would that be the best use of resources? I would rather see the investment go towards something tangiable that we can benefit from.OTH-if it's easy then do it.

 

[Joe Bryant]

5 Year subscriber here.  I love the Articles, Stats, Advice, Forums, Tools, etc.  It seems though about every year at this time I make a request and it never gets fulfilled.

The Request -

Can we have an Article or whatever you want to call it that takes the Projections from last year and compares it to the End of Year Actual Stats?  So for example if FBG had Javon Walker the 5th WR going into last year and obviously his Actual Stats were way down compared to his FBG Projection, you can show the differential and have the remark - Injured Most The Year.  The purpose of this is not for injured players but the a direct comparison of the Players in their Positions for Before and After the Season.  Then let us analyze why the 2nd rated QB Projected finished 13th for example. 

Projecting without looking back to see the mistakes or the successes doesn't make too much sense in my opinion.

Now mind you I would do this on my own but I can barely use a computer.  Also, last year there was a lot of agreement on this task but it never came to fruition.

And for the funny guys (and you're really not)  I know Projections are guesses.  There is no science and I do not need FBG to give me the Powerball numbers for tomorrow night.  Just a little analysis on why things went off the track or why they were right on. 

Humbly yours,

A long time FBG guy.  Thanks

<{POST_SNAPBACK}>

Hi tyty,Thanks for the good words and thanks for hanging with us this long. Our projections for several years back are available and this might be a fun thing (and surely humbling) excercise. The tough thing is agreeing on the best way to measure. And then also what you're measuring against. Other sites? How they would compare to a perfect projection? That kind of thing.

We always get a ton of feedback on stuff we do and intelligent discussion is always welcome. People are 100% wrong there if they think we don't recognize that helps. What we will not do is get into the "what were you smoking to make a call that stupd?" type stuff. Those are a waste of time. But many of the guys here are fantastic for having good discussion and backing it up with a good argument. We're always anxious to hear that kind of thing.

Thanks again for being a part of what we're doing.

J

 

[Unlucky]

My 2005 projections are linked to in my sig.

I'm currently working on 2006, and in the process, comparing my 2005 projections to the 2005 actual, to hopefully learn from any procedural mistakes.

Perhaps I'll post the results on my site and pass it along to you guys.

 

[Joe T]

I think a lot of the staff gets offended when their logic is questioned.

<{POST_SNAPBACK}>

You should be specific here. Who and how many are "a lot"?
<{POST_SNAPBACK}>

2?
<{POST_SNAPBACK}>

Okey Dokey. I'm not disputing your claim. I've seen staff members handle criticism poorly. Just don't want you to blanket the majority with a negative comment when it should be directed at a specific person.
<{POST_SNAPBACK}>

Oh. I thought you were asking tongue and cheek.Didn't think you really wanted me to name names.

You are right, it is not all and probably not a majority.

Still don't want to call out people. But they know who they are.

 

[David Dodds]

Here is where I am at with all of this. The article would take a long time. Right now, I am way more concerned about getting fresh quality content to the website and magazine as opposed to going backwards and analyzing things. The fact my hard drive crashed and I actually do not even have the raw data anymore also makes this exercise tougher than most.

I can assure you that I take my projections very seriously. My first cut projections take nearly 40 hours and match to NFL norms in virtually every category (ie add up all the RB yards and YPC is with 99% of league wide numbers, etc).

Here is my process:

- From the team view, run the last three years of data as a baseline.

- Analyze differences in coaching philosophy, personnel, off-season moves, ages of players, etc to determine run/pass team prediction.

- Make sure all teams added up fit with in NFL norms. Normalize the data to make this happen if needed.

- Begin team by team to fit numbers into this team view keeping historical ypc, catches per year, role in offense, etc consistent with team numbers generated above.

- When all teams are done, I then convert to rankings.

- I also do checks for 1,000 yard rushers, 1,000 yard passers, and comparison to last three year's AVT numbers to ensure the data matches up from a historical perspective.

I am about 8 hours into this process for this year. I feel it is way more important to work on this year's numbers than spend a lot of time on last year's numbers. I am not trying to duck anything, I just don't see the point. I do not believe it will have me change my process as I have outlined above.

The magazine and site will keep me swamped until around May 20th. Perhaps I will revisit this then. To me the more important comparion would actually be to examine where I thought each team would perform at and how they did. That is the biggest basis for error in my opinion. Kevin Jones is off because I projected Detroit's running game too high for example.

I do think the three year look at the past always tries to keep these errors from being too big though. In the end, people break out and people underperform. That is the nature of the beast.

 

[JayMan]

Here is where I am at with all of this. The article would take a long time.  Right now, I am way more concerned about getting fresh quality content to the website and magazine as opposed to going backwards and analyzing things.  The fact my hard drive crashed and I actually do not even have the raw data anymore also makes this exercise tougher than most. 

I can assure you that I take my projections very seriously. My first cut projections take nearly 40 hours and match to NFL norms in virtually every category (ie add up all the RB yards and YPC is with 99% of league wide numbers, etc).

Here is my process:

- From the team view, run the last three years of data as a baseline.

- Analyze differences in coaching philosophy, personnel, off-season moves, ages of players, etc to determine run/pass team prediction.

- Make sure all teams added up fit with in NFL norms.  Normalize the data to make this happen if needed.

- Begin team by team to fit numbers into this team view keeping historical ypc, catches per year, role in offense, etc consistent with team numbers generated above.

- When all teams are done, I think convert to rankings.

- I also do checks for 1,000 yard rushers, 1,000 yard passers, and comparison to last three year's AVT numbers to ensure the data matches up from a historical perspective.

I am about 8 hours into this process for this year.  I feel it is way more important to work on this year's numbers than spend a lot of time on last year's numbers.  I am not trying to duck anything, I just don't see the point. I do not believe it will have me change my process as I have outlined above. 

The magazine and site will keep me swamped until around May 20th.  Perhaps I will revisit this then.  To me the more important comparion would actually be to examine where I thought each team would perform at and how they did.  That is the biggest basis for error in my opinion.  Kevin Jones is off because I projected Detroit's running game too high for example.

I do think the three year look at the past always tries to keep these errors from being too big though.  In the end, people break out and people underperform.  That is the nature of the beast.

<{POST_SNAPBACK}>

This is an excellent post... since it looks at the fundemental question here... would looking back at past years projections change the way FBG project this years numbers?...From David's answer above - it wouldn't... it seems to me that it would be irrevelant to post them then... since effort should be placed on this year rather the past ones...

Another excellent point - the second highlighted part... the projections are/were way off if one thought that the Bengals (taking into account all variables: coaches, opponents, roster depth, etc.) would pass for 20TDs instead of, let's say, 45TDs... The underlying players projections are just then screwed by that figure... CJ going from 5 to 12 TDs... Housh from 4 to 9... Henry from 3 to 8... etc.

This is the real issue here, in my opinion... why should the Bengals toss 20TDs instead of 45?...

My 2 cents...

 

[Mark Wimer]

And I might add that every year I put out a consensus rankings critique and I get comments defending or arguing against positions concerning specific players from over half of the staff.

<{POST_SNAPBACK}>

True.LHUCKS and I have clashed many times, over many players, to the benefit of my rankings and adding to the general discourse on the board. Pony Boy, Bass N Brew, CalBear and many others have challenged me to improve my work.

I appreciate that sincerely. There are very few fields of endeavor where one recieves instant feedback - fantasy football happens to be one of those few.

As long as no ad hominems are tossed about, I am more than willing to hear criticisms regarding my perceptions of the current state of the NFL. I think that is universally true of the FBG's staff.

All we ask in return is a modicum of civility.

I'm excited to hear the commentary on our first cut of 2006 rankings!

MW

 

[MelvinTScupper]

I think the only valuable element to such an exercise would be to stack up our projections vs several others.  i believe there is some sort of comparison done amoungt different sites, but I don't think it's as comprehensive as some may be yearning for here.

If we were the only data point, would being "fairly close" 47% of the time really mean anything?  I would suppose that if other sites were "fairly close" 32% of the time than we could have done "better."

<{POST_SNAPBACK}>

We did this measuring deviation from the ranking between my rankings last year and David and Joe's rankings last year.As a result, MTScup now owes me $20.

<{POST_SNAPBACK}>

link?I know we made this bet, but I'd like to see your analysis, and how it was determined.

 

[tytyty]

Here is where I am at with all of this. The article would take a long time.  Right now, I am way more concerned about getting fresh quality content to the website and magazine as opposed to going backwards and analyzing things.  The fact my hard drive crashed and I actually do not even have the raw data anymore also makes this exercise tougher than most. 

I can assure you that I take my projections very seriously. My first cut projections take nearly 40 hours and match to NFL norms in virtually every category (ie add up all the RB yards and YPC is with 99% of league wide numbers, etc).

Here is my process:

- From the team view, run the last three years of data as a baseline.

- Analyze differences in coaching philosophy, personnel, off-season moves, ages of players, etc to determine run/pass team prediction.

- Make sure all teams added up fit with in NFL norms.  Normalize the data to make this happen if needed.

- Begin team by team to fit numbers into this team view keeping historical ypc, catches per year, role in offense, etc consistent with team numbers generated above.

- When all teams are done, I then convert to rankings.

- I also do checks for 1,000 yard rushers, 1,000 yard passers, and comparison to last three year's AVT numbers to ensure the data matches up from a historical perspective.

I am about 8 hours into this process for this year.  I feel it is way more important to work on this year's numbers than spend a lot of time on last year's numbers.  I am not trying to duck anything, I just don't see the point. I do not believe it will have me change my process as I have outlined above. 

The magazine and site will keep me swamped until around May 20th.  Perhaps I will revisit this then.  To me the more important comparion would actually be to examine where I thought each team would perform at and how they did.  That is the biggest basis for error in my opinion.  Kevin Jones is off because I projected Detroit's running game too high for example.

I do think the three year look at the past always tries to keep these errors from being too big though.  In the end, people break out and people underperform.  That is the nature of the beast.

<{POST_SNAPBACK}>

Ok, since I am the one who started this let me put the original request to bed. If I want to do my own comparisons I will just do it. Thanks for the insight on how you come to your conclusions. Being that I don't have the time or patience to employ your method I always use your projections as a starting point and put my own twist on my projections from there. I feel this is fair since I pay for the site.Can I/We ask this.... Doing this as long as you have, there must be a point in the season or afterwards where you look back and go , hmm I was off in pre season on this guy or that guy. The thought must go into your head on why or how this happened.

Going by the method you use, do you find some of the ways you project to trend as a staple of projecting correctly and some that might be more iffy. Example, on a guy that you projected off quite a bit good or bad... did you put too much into his age, or did the coaching philosophy prove to have a different affect on the projection that you thought?

Just thought through the years of experience that you might see some tendencies that work or do not on a consistent basis. That kind of insight would be great to hear and help on our own projections. Thanks

 

[culdeus]

So, ultimately the question comes down to - what projections do you want to compare? Shall we do an analysis of every staff member's final rankings vs. the NFL actuals, and then eliminate all the "injury issues" like Javon Walker 2005? We put out 162 customized cheat sheets for various leagues during the pre-season draft rush 162 cheatsheets  - shall we do an analysis of how close all 162 cheatsheets came out vs. NFL actual, given all 162 scoring paradigms covered?

I think the question being posed in this thread is much more complex than it appears on first glance, due to the broad-based, wide-ranging approach here at Footballguys.

My .02.

<{POST_SNAPBACK}>

I'd love to see the variance between any staff member's projections and actuals as well as the consensus version. Maybe Chase projects TEs better than others. Maybe David's overall projections are better.I understand it isn't easy, but I would rather see resources put in there than some other fluff that the site has. But again that comes down to personal preference of what one finds value in and another does not.

I no longer subscribe to FBG but donate to Drinen's site (profootball-reference) as I find I use that infinitely more useful than FBG content. Others would be in the opposite camp.

Maybe you can't do this before this year, but there are many technical people here who are not looking for a scape goat or to point fingers but to see how projections differ and how everyone can improve.

Projections will always be off. It is just a matter of scale.

If Staff member X continuously projects better for RBs than the rest I will look at that staff's rankings closer in the next year.

<{POST_SNAPBACK}>

From a stats perspective you are gonna run into major sample size issues there.

 

[culdeus]

Here is where I am at with all of this. The article would take a long time.  Right now, I am way more concerned about getting fresh quality content to the website and magazine as opposed to going backwards and analyzing things.  The fact my hard drive crashed and I actually do not even have the raw data anymore also makes this exercise tougher than most. 

I can assure you that I take my projections very seriously. My first cut projections take nearly 40 hours and match to NFL norms in virtually every category (ie add up all the RB yards and YPC is with 99% of league wide numbers, etc).

Here is my process:

- From the team view, run the last three years of data as a baseline.

- Analyze differences in coaching philosophy, personnel, off-season moves, ages of players, etc to determine run/pass team prediction.

- Make sure all teams added up fit with in NFL norms.  Normalize the data to make this happen if needed.

- Begin team by team to fit numbers into this team view keeping historical ypc, catches per year, role in offense, etc consistent with team numbers generated above.

- When all teams are done, I then convert to rankings.

- I also do checks for 1,000 yard rushers, 1,000 yard passers, and comparison to last three year's AVT numbers to ensure the data matches up from a historical perspective.

I am about 8 hours into this process for this year.  I feel it is way more important to work on this year's numbers than spend a lot of time on last year's numbers.  I am not trying to duck anything, I just don't see the point. I do not believe it will have me change my process as I have outlined above. 

The magazine and site will keep me swamped until around May 20th.  Perhaps I will revisit this then.  To me the more important comparion would actually be to examine where I thought each team would perform at and how they did.  That is the biggest basis for error in my opinion.  Kevin Jones is off because I projected Detroit's running game too high for example.

I do think the three year look at the past always tries to keep these errors from being too big though.  In the end, people break out and people underperform.  That is the nature of the beast.

<{POST_SNAPBACK}>

In that case wouldn't it be relatively easy to see where the error is. Compare total team stats to actual stats and figure out the variance. Then figure the error within the team player-to-player. i.e. this player accounted for 15% of the FP production compared to projected 25% etc. etc.

It would probably be a worthless effort and show that the team projections are at the heart of the error (taking injuiries out of the mix).