Dodds Projections (1 Viewer)

[Bankerguy]

Without giving away all the details of your projections for the QB's (As it is now $$$ material). One huge question for me:

In 2005, there were a total of 11 QB's with 20 passing TD's or more.

This year you are projecting 19 QB's to throw 20 or more.

I'm trying to understand the rationale behind this. Can you share your thoughts on this upswing in TD's?

Mods.....If this post is off-side or against the rules... (My apologies, delete if not appropriate)

 

[joffer]

every year i look forward to this thread

hopefully Maurile stops by with his individual vs. group projections argument.

 

[Chase Stuart]

every year i look forward to this thread

hopefully Maurile stops by with his individual vs. group projections argument.

LOL. I can't figure out if it's good or bad that I could guess the content of this post just by reading the title.In all seriousness, Bankerguy, I'll look to see if we've got anything from MT stored on this. It's doubtlessly better worded than whatever I'll come up with.

 

[Sabertooth]

Injuries. Projections don't take them into account by nature. Go by TD's per game and project it out over 16 games and see how many are over 16.

 

[Doug Drinen]

To add to what others have said...

I'll bet that if you check the number of teams that he has projected to throw X touchdowns it would macth up with the historical data pretty well. Likewise, I'd guess that his projected league TD pass total is in line with historical norms.

He just has a different distribution of TD passes. That's because, as others have said, he doesn't want to do too much guessing as to who is going to get injured.

 

[Chase Stuart]

Without giving away all the details of your projections for the QB's (As it is now $$$ material). One huge question for me:

In 2005, there were a total of 11 QB's with 20 passing TD's or more.

This year you are projecting 19 QB's to throw 20 or more.

I'm trying to understand the rationale behind this. Can you share your thoughts on this upswing in TD's?

Mods.....If this post is off-side or against the rules... (My apologies, delete if not appropriate)

Bankerguy,Here's the basic version. Injuries are pretty random, and it's impossible to factor them into your QB projections with any sort of accuracy (outside of maybe Bulger, Warner, Pennington. More on them at the end). So FBGs has two choices: predict that 10 QBs will get hurt each year (or whatever the exact number is) based on basically no empirical evidence, OR predict that each QB will remain mostly healthy. We're very confident that the most accurate set of projections (in terms of usefulness for your drafts) are done by choosing the latter method.

You'll notice that team TDs and leaguewide passing TDs are always within historical norms. It's just that starting QBs are generally "overprojected" while backup QBs are generally "underprojected." It's the basic problem that would come from something like this.

Go through each of the top 19 QBs...and ask do you think XXX will throw for 20 TDs? Dodds' answer will be 'yes'. If you then ask Dodds do you think 15 or more QBs will throw for 20 TDs or more, he'll probably tell you 'no'. Understanding that those two answers aren't inconsistent is the key to your question here.

But I feel extremely confident that Dodds' system of projecting QBs (and all other positions) is as good as anyone else's I've seen.

 

[Maurile Tremblay]

I'll bet that if you check the number of teams that he has projected to throw X touchdowns it would macth up with the historical data pretty well. Likewise, I'd guess that his projected league TD pass total is in line with historical norms.

Yep.Total TDs are in line with previous years. Total TDs by starters (versus backups) may not be. But even if they were, you'd still see less variance in starter-backup TD ratios in the projections than you would in end-of-year results; so you'd still see more 20 TD QBs in the projections than in end-of-year results.

And that is the way it should be.

 

[Musesboy]

Go through each of the top 19 QBs...and ask do you think XXX will throw for 20 TDs? Dodds' answer will be 'yes'. If you then ask Dodds do you think 15 or more QBs will throw for 20 TDs or more, he'll probably tell you 'no'. Understanding that those two answers aren't inconsistent is the key to your question here.

But I feel extremely confident that Dodds' system of projecting QBs (and all other positions) is as good as anyone else's I've seen.

Fairly similar to the top three RB rankings. We know what the general consensus is on the top three backs. But the top three have only repeated once in the last eleven years. I would expect a similar answer if you asked whether the top three would actually finish 1, 2, 3.

 

[Chase Stuart]

Go through each of the top 19 QBs...and ask do you think XXX will throw for 20 TDs? Dodds' answer will be 'yes'. If you then ask Dodds do you think 15 or more QBs will throw for 20 TDs or more, he'll probably tell you 'no'. Understanding that those two answers aren't inconsistent is the key to your question here.

But I feel extremely confident that Dodds' system of projecting QBs (and all other positions) is as good as anyone else's I've seen.

Fairly similar to the top three RB rankings. We know what the general consensus is on the top three backs. But the top three have only repeated once on the last eleven years. I would expect a similar answer if you asked whether the top three would actually finish 1, 2, 3.

Exactly Museboy. It's certainly unlikely that all three finish in the top 3, and the odds of them finishing in any specific order is about 6 times as unlikely as that.

 

[Maurile Tremblay]

Go through each of the top 19 QBs...and ask do you think XXX will throw for 20 TDs? Dodds' answer will be 'yes'. If you then ask Dodds do you think 15 or more QBs will throw for 20 TDs or more, he'll probably tell you 'no'. Understanding that those two answers aren't inconsistent is the key to your question here.

 

[Mark Wimer]

I'd add in to say that the projections at the top of the by-position rankings are going to look right in line with historical norms, in general - but as you get lower down, the projections will tend to skew higher in terms of FP because as other posters have pointed out, injuries are for the most part random occurances (there is no way to know that relatively young, healthy guys like Carson Palmer or Daunte Culpepper were going to blow their knees out last year - however, we did know that Marshall Faulk's knees were getting balky and that he had a lot of issues with needing them drained due to swelling in the joints caused by ongoing issues. Once an injury becomes "chronic", then it can be factored in to some extent or other, but that is always a case-by-case thing and not something you can generally apply across a position).

There are also other random events that occur each year which cannot be predicted before the fact for the majority of guys - players getting 4 game drug/steroid suspensions, for example (although Dolphin Ricky Williams is perhaps one counterexample of a predictable drug suspension); family tragedies that lead to a game or perhaps games being missed (death of a family member; serious illness of a family member); and accidents like Ben Roethlisberger's crash this past off-season.

So, anytime a forward looking set of projections is compared to real-world actuals from previous years, there is a certain amount of "hind-sight is 20/20" in saying "previous years actuals say the number Xth player should be right around X TDs, but your projections say X+5" (or whatever the difference works out to be).

My .02.

 

[ookook]

This actually illustrates one of my biggest pet peeves with all the FBG data I get: I really wish everything woudl be based on per week statistics not season totals.

I know, one can convert and this one does.

But it makes a big difference in so many discussions on these boards...where rankings are biased because of a couple missed games...sometimes at the very end of the season.

All of us know (okay most) that the value of missed games is not zero. That is, we are losing the difference between our starter and backup at that position.

Total season data counts these as zero.

Is frusterating sometimes.

 

[Chase Stuart]

This actually illustrates one of my biggest pet peeves with all the FBG data I get: I really wish everything woudl be based on per week statistics not season totals.

I know, one can convert and this one does.

But it makes a big difference in so many discussions on these boards...where rankings are biased because of a couple missed games...sometimes at the very end of the season.

All of us know (okay most) that the value of missed games is not zero. That is, we are losing the difference between our starter and backup at that position.

Total season data counts these as zero.

Is frusterating sometimes.

That's a fair point. I know for our rankings in 2003, we were supposed to project Jimmy Smith and Mike Vick (and maybe one other guy) a little differently than everyone else. Smith had a four game suspension on his plate, so we were (wisely) told to project Smith's numbers over the remaining 12 games, and then add in the average backup (or something like that) for the first four games when we ranked Smith. That way our "rankings" wouldn't reflect EOY fantasy point totals, but rather at what point we should draft him. The same went for Vick who was schedule to miss the first six games or so, but obviously ended up missing quite a few more.For what it's worth, I think that's a very smart way to do the rankings. We should always be more concerned with getting our drafts right than getting our pre-season projections to look like the EOY results.

 

[Bankerguy]

Thanks for the quick replies.

I was just curious.

 

[tytyty]

This actually illustrates one of my biggest pet peeves with all the FBG data I get: I really wish everything woudl be based on per week statistics not season totals.

I know, one can convert and this one does.

But it makes a big difference in so many discussions on these boards...where rankings are biased because of a couple missed games...sometimes at the very end of the season. 

All of us know (okay most) that the value of missed games is not zero. That is, we are losing the difference between our starter and backup at that position.

Total season data counts these as zero.

Is frusterating sometimes.

That's a fair point. I know for our rankings in 2003, we were supposed to project Jimmy Smith and Mike Vick (and maybe one other guy) a little differently than everyone else. Smith had a four game suspension on his plate, so we were (wisely) told to project Smith's numbers over the remaining 12 games, and then add in the average backup (or something like that) for the first four games when we ranked Smith. That way our "rankings" wouldn't reflect EOY fantasy point totals, but rather at what point we should draft him. The same went for Vick who was schedule to miss the first six games or so, but obviously ended up missing quite a few more.For what it's worth, I think that's a very smart way to do the rankings. We should always be more concerned with getting our drafts right than getting our pre-season projections to look like the EOY results.

Ok, so what FBG'er , uhh hum, Chase, will put together a "Draft Special" taking in to consideration all the non "End Of Season" but rather use the improbable, impossible, nay no-way-jose---- injury, suspension, avg per game super duper draft guide. Seriously though, I WANT IT!!!

 

[Chase Stuart]

This actually illustrates one of my biggest pet peeves with all the FBG data I get: I really wish everything woudl be based on per week statistics not season totals.

I know, one can convert and this one does.

But it makes a big difference in so many discussions on these boards...where rankings are biased because of a couple missed games...sometimes at the very end of the season.

All of us know (okay most) that the value of missed games is not zero. That is, we are losing the difference between our starter and backup at that position.

Total season data counts these as zero.

Is frusterating sometimes.

That's a fair point. I know for our rankings in 2003, we were supposed to project Jimmy Smith and Mike Vick (and maybe one other guy) a little differently than everyone else. Smith had a four game suspension on his plate, so we were (wisely) told to project Smith's numbers over the remaining 12 games, and then add in the average backup (or something like that) for the first four games when we ranked Smith. That way our "rankings" wouldn't reflect EOY fantasy point totals, but rather at what point we should draft him. The same went for Vick who was schedule to miss the first six games or so, but obviously ended up missing quite a few more.For what it's worth, I think that's a very smart way to do the rankings. We should always be more concerned with getting our drafts right than getting our pre-season projections to look like the EOY results.

Ok, so what FBG'er , uhh hum, Chase, will put together a "Draft Special" taking in to consideration all the non "End Of Season" but rather use the improbable, impossible, nay no-way-jose---- injury, suspension, avg per game super duper draft guide. Seriously though, I WANT IT!!!

As far as I can tell, Carson Palmer is the only guy this would be applicable for so far, and it's unclear whether or not he'll miss any time. Same for Daunte Culpepper if he regresses.But let's say Palmer is expected to be ready in week 6. That would mean he would miss four games (Cincinnati has a bye week 5.) Dodds has Palmer projected at 277 FPs, which doesn't appear to include any special downgrade for possible missed games. (Palmer is being credited with 93.2% of all Bengals QB pass attempts; Hasselbeck has 95% of all Seahawks QB pass attempts as a comparison). So we'd take 12/16 of 277 FPs and get 208 FPs. Then I'd add to that 25% of the total FPs for QB18 -- Bryon Leftwich (243.2) -- and add those 61 points. Now I've got Palmer's "value" equalling 269 FPs. That moves Palmer from 4th to 10th in the QB projections, as opposed to 30th (if we gave him 208 points). So Palmer would be worth taking as the 10th QB in your draft if you agreed with all of Dodds' projections and we knew that Palmer would miss the first four games.

Once again, I expect the staff rankings to reflect this methodology once we get news on this about Palmer or anyone else. But as of right now it sounds like Palmer's going to play week one.