This is a sample guest message. Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!
That's my thinking, diesel. I saw a list once, but can't find it now. How many yards did he have taken away because of penalties? Sure, he may not have broken it long without the hold (valid or not), but I still see him gettin' a few more of those this year. Add in your point about the injury, and I see is YPC closer to 4.8 than 4.1
I disagree there. Alexander was held back quite a bit last year because he always had to stay in to pass-block. (How do you think the second-worst pass-blocking OL in the league gave up the fourth-fewest [or whatever] sacks?) Alexander will surely improve on last year's numbers if he can stay healthy, if only because of the Solomon Page signing and Parker/McIntosh's return to health.
Yes. He was the #4 WR in the league last year in terms of PPG. (He played only 11 games, not the 13 that are listed.) He is good.
I agree with that one. IMO, Coles is clearly a better WR than Gardner (and he fits into the fun'n'gun better), and Gardner will probably lose targets as he goes from the #1 WR to the #2 WR.
If you don't like the Word of God, see Chase's face-off article. It's better.
Yes, tremendously. And if you don't think so, you've never seen Kelvin Garmon try to pass-block.
Don't back away now, 'cuz I think you're right. When looking at Conway, I see the FBG numbers as a 16 game "minimum". The dude was a monster last year (pre-injury) and is taking over a position that produced monster numbers last year. I feel much better about 80-1100-8 than 70-<1000-6
Since 1980 (which is roughly when I would say the "modern era" of football began, but that's entirely debatable), there have only been 3 WRs who have changed teams after scoring at least 160 fantasy points the previous season:Player...........Year1...Pts....PPG..Year2..PPG...%decline
Just for the sake of comparison, I've got:Curtis Conway: 70 rec, 1,011 recYDs, 5.3 recTDs (40 rushYDs, 0.5 rushTDs)Santana Moss: 50 rec, 697 recYDs, 4.7 recTDs (25 rushYDs)Wayne Chrebet: 54 rec, 729 recYDs, 5.7 recTDs (5 rushYDs)Chad Pennington: 3,857 passYDs, 21.5 passTDs, 14.3 INTs (62 rushYDs, 1 rushTD) I could easily see Moss doing better than that, though. I agree that the T.O.D. threads will be a lot of fun.
Dave - I'm usually right in line with your thinking but Wade has you on this one. You must take historical norms and incorporate them into your projections. You do a great job of 'mass balancing' the league-wide scoring & yardage. But, you have to apply what you know and take your best shot with respect to the norms.Back to the wager - can I get in on that action?
Dave - FBG will have 20,000+ pages this season prepared by many individual contributors. Most of us understand that FBG allows the individual contributors to present their own views & opinions. Generally, that's a good thing. My only request would be that each feature is clearly attributed to specific contributors (for the most part that's being done). If the projections are yours - great! If the depth charts are Shawn's - fine! If the rankings are Joe's - super! But, please clealry attribute them. Let us sort out who we like & agree with and who we don't like & disagree with.Last season I took issue with Chris Smith doing exactly what you did with respect to projecting & historical norms (like 20+ TDs for passers). If I know where you and Chris are coming from I can better digest your foresight.Keep up the good work!
Here's why I don't like that aproach. I'll take things to an extreme to illustrate a point, but I think the point still applies in less extreme cases that we would be working with.Say there's a RB out there (back A) who I expect to score 50% more than the next best RB (back B) scores, but he only has a 50% chance of playing this year. By your calculation, A's "projection" value is 75% of the B's, assuming B is a 100% lock to play. On the face of things, that might seem OK.BUT, say we have 5 other backs (C-G) that are going to be VERY close to B in terms of projected scoring, so that basically B-G are all about the same.If we are suing these projections to assign DRAFT VALUE at some point (which is the only reason I can think of to do the calculation you are talking about), B-G are all going to end up with a considerably higher value than A. But I'm not sure I wouldn't rather have A. Sure, it's only a 50% shot, but and IF player A "hits" the team that has him is going to have a SIGNIFICANT advantage in the race to be the ONE team out of 12 or whatver that wins the crown/money if applicable. Isn't A worth that shot?, He may not be worth MORE than B-G, but isn't he worth at least the same, not significantly less?BETTER STILL, what if the 50% chance of playing we assigned A applied on a game-by-game basis rather than a season basis which is typically the case in reality? In that case, even if he only plays half of his games, but is 50% better than all other backs while he's playing, isn't he worth it if you have at least somewhat reasonable knowledge which games he will play in?OK, that's the extreme one.Now on to a "real" one. If I knew in advance what Priest would only play in 14 games again this season, but he would play as well as he did last season, what would his "projections" look like? Sould I discount them by multiplying by 14/16? What is he worth relative to a guy who scores 14/16 as many points per game, but plays in all 16 games. Which would you rather have? I'd rather have Priest, BECAUSE WHEN HE IS OUT, HIS BACKUP IS PLAYING (fantasy BU, not necessarily NFL BU) and scoring points. The other guy's backup never plays obviously, so you are AHEAD in net points by a significant margin.Now if we are talking about Terrell Davis circa 2001 where you never know if the guy is going to play or not, that's a different story. But in general, I'd much rather have the guy who will play in a couple fewer games but will score more points when he DOES play. This makes the straight percentage reduction of his projections a little sketchy for me if you plan to send those values into a system (whatever system that is) to determine value.
Chris - that's fine - use a risk factor. But, I think its folly not to compare your league-wide totals to historical norms to base your projections on fact rather than fiction. If, after applying your risk factor, you end up with 25 QBs throwing for 25+ TDs, you've missed the mark. If, after applying your risk factor, you see 17 RB/WR/TE with 10+ TDs, you've hit the historical target.
It looks like there is a difference of opinion on this. I strongly disagree with your approach and strongly support David's, for the reasons stated in the current thread and in this one. (You're basically endorsing the AVT approach of attaching ex post historical data to ex ante projections, which is the wrong way to do it, IMO.)In any event, I don't think you're going to get David to change his mind on this. However, you don't need to. If you want to try to predict which 7 QBs will miss at least 4 games (which is basically what you'd be trying to do in using "historical norms" to reduce the number of QBs projected to get 20 TDs), you will be able to do this when the VBD App comes out by reducing a bunch of the starting QBs' projections and, I would guess, increasing a bunch of the backup QBs' projections to make up for it.
I do however have one question to post: I noticed that you guys have no QBs throwing for 30+ TDs. I can't believe that you really think none will, as 2002 was the only year of the last 9 where none did (Brady falling 1 short with 29). My real question for you guys is who do you think has the best chance in 03? Because after looking more at the QBs I'm drawing a blank. As of now I'm leaning towards Manning, Warner (if he can stay on the field), Brooks (love his weapons), and Hassy (Hawks O is primed for a big yr.)
This is kind of the same principle. Suppose there are 10 QBs who each have a 15% chance of throwing 30+ TDs. On average, then, there will be 1.5 QBs each year who will throw 30+ TDs. However, projecting any particular QB to throw 30+ TDs would be incorrect: the odds are against each and every one of them as individuals. So your individual projections will show zero QBs throwing for 30+ TDs.Edit: But to answer your question, I think Warner and Manning are the most likely.
Convenient analogy but try this one (on the fly so work with me on this one). You have a retirement plan with 20 stocks to invest in. On average 5 stocks outperform the major indexes annually, 10 hit the indexes on the nose and 5 underperform. Assuming that you want to diversify and assuming that you can pick as many stocks as you like but you must distribute contributions equally to however many you select, how many stocks will you invest in? Surely not all 20 - you'd guarantee yourself some losers. You'd probably do your best to select the 5 or so that you feel have the best opportunity to outperform the indexes.I know - buy an index fund and forgettaboutit ......
We're all doing the same thing - taking our crystal balls for a ride. Some of us predict, others forecast, still others use pirated rankings. For FBG to purvey forecasts that buck historical trends we'd have to accept the premise that its up to us to add a 'risk factor' anyway. Why? Tell us who you think will perform well within the limits of historical norms and allow the subscriber to accept or reject the information.
If each stock has the same expected return to the best of your knowledge, and transaction costs were not an issue, you would minimize variance (I hate the word "risk" in that context) by buying all of them. If they don't have the same expected return to the best of your knowledge (i.e., you like some of them better than others), you will buy the ones you like best. (The exact number depends on your taste for variance.) But what you would not do in this situation is say: "I think Stock A is the best stock of all of them. I calculate its expected return to be 15% per annum, which is higher than my calculations for any of the others. However, since at least one stock each year returns at least 30% per annum, I should max out all my credit cards at 20% to buy as much of Stock A as I can. After all, I'm projecting it to be the best stock, and the best stock always returns at least 30%."
5. Lower the player's projections to the amount found in 4) by distributing some of his touches to the backups on the real NFL team.This isn't perfect in that it doesn't include the lack of depth during the weeks that my RB3 is starting for me. That is a risk that I will by manage by hand during the draft, knowing if I take Pittman, I probably want to take a RB4 earlier than normal as well.
You guys are right - the projections ARE in line with his past production. What got me was the ranking - I just couldn't imagine Booker being a top-10 WR, or drafting him before guys like K Rob, C. Johnson, or Toomer. But that's not really an issue with the projection, which is rather hard to argue with. Shows what goes when you post on reaction rather than research!
Its not a mistake - its real life. Why buck historical norms? Think of it this way - there was about 116,000 yards passed/received in 2002 in the NFL. That's within 7% of the past 6 years rolling average AND it includes the impact of the Texan's expansion. Why would I forecast much more than 116,000 yards for all the passers in the NFL for 2003? It wouldn't make sense. The group total is relevant. Much like Unlucky, I project for each team. Then I 'mass balance' the totals to make sure they fit into the historical data. You're a card player - think of it as dealing from a deck. There are 52 to go around. Period.So, we're back to crystal balls again. He who guesses right wins. My biggest point though is 'Please don't purvey forecasts that don't reflect real life'. All you're be doing is tossing out numbers that we'll have to 'adjust' anyway.
For the exact same reason you should "buck historical norms" in blackjack by refusing to take insurance 30% of the time even though 30% of insurance bets will be winners.
You wouldn't. David didn't. His total passing numbers for the league are in line with historical norms. They are just distributed more to the starting QBs and less to the backup QBs because he's not singling out particular starters to miss half the season with injuries.
I'm pretty sure Unlucky does it the Dodds way (i.e., ending up with around 18-20 QBs with 20+ TDs rather than 10-12 QBs with 20+ TDs).Unlucky?
While we're on the subject of Curtis Conway: Last year he had 57 catches and only one dropped ball. That's the most impressive "drops" stat I've seen so far from last year aside from Keenan McCardell's 61 catches and zero drops.STATS, Inc.
starting with 'team stats' is the way to do projections. I look at the past two seasons for each team and then compile projected team statistics first before going onto the individual players. I don't just copy last year's stats though.I look at teams that have improved aspects of their offense or gotten worse this year (such as offensive line play- the Texans have to be stronger and receiver ability - the Redskins are stronger than a season ago) and adjust the numbers based on those observations.
Hello, first of all, yes I incorporate the Risk % to set up my draft sheet which is the reason for using projections in the first place.In regards to your example, (really is extreme) I myself would rather have one of the more likely players to contribute if it is in the first few rounds of a draft but could see myself rolling the dice on the player with the high risk later on. My projected numbers have Priest Holmes in the top five and Tim Couch in the top twelve. Would I be comfortable taking Holmes in the first five picks with his injury concerns at this point in the pre-season. No way! Assigning him a risk factor doesn't try to assume, he'll miss 1 game, 2 games or 12 games. The risk factor simply states that I feel player X is a bigger risk this season to hit the numbers projected than player Y.Tim Couch is a really good example of this. I feel that whoever wins the starting job in Cleveland is capable of putting up top ten numbers in 2003. However if I do a draft tomorrow and don't assign a risk %, I would have Couch too high for a player that may not even start this season. He has a very real RISK FACTOR and it is Holcomb possibly winning the job outright leaving Couch as the backup. So while I like the possibility of Couch having a big season, my Risk % drops him down to around 25 on my draft sheet.The nice thing about assigning the Risk % is that it is easy to change on the fly. In a month from now, if Couch looks like the job is officially his, I can change the risk % and have him climb my rankings.Another player example is Fred Taylor. He is a fellow very capable of top five numbers if he can stay healthy. Assigning him a risk % helps me to curb my enthusiasm in regards to his potential.I just think it is a solid way to project every player being healthy through the entire season yet also take into account a player being (a) injury prone or (B) in serious competition for the job.
