PPR Articles (1 Viewer)

[stickboy]

Since I play in several PPR leagues, I was very happy to see that Jeff had posted a series of articles on PPR strategy that looked at 7 years of data to analyze trends. What I found was that the articles really are not about PPR, they are general trends that are occurring. For example, it cites the relative increase in value of the QB position as the offenses open up, but QB is not any different between PPR and non-PPR. These are valuable insights, but not very helpful with PPR strategy. Similarly, the RB position is discussed relative to stud RB vs. RBBC, but there is not much discussion around the consistency of points around pass catching RB's like an MJD. The analysis and discussion is mostly around why you need a stud RB, and not related to PPR.

Ok, so I should skip those parts and read the analysis of the WR position and its contribution in PPR (where there are additional points to be had). What I found was an assertion (with an example rather than data) that WR scoring is very hard to predict year to year, and that even with PPR the points are at best equivalent to a mid-range QB. The assertion has some merit as the top QB's and RB's are generally easier to predict than the top WR's, but is this true overall? If I create tiers rather than absolute numbers, is this assertion true? And why compare WR to QB, isn't the whole idea of VBD to win the position battles as we have required starting lineups? And if you look at the data, 2008 was a true outlier in number of WR in the top 50 compared to the previous few years.

Sorry if this sounds negative, I think there is is a lot of valuable information in looking at scoring trends for top 50 players in PPR leagues over 7 years. And I can see the line of thought where it evolved into an article on which positions are most valuable and the best contributors even with a point per reception to move folks away from stud RB. This is useful in draft planning depending on your slot, and especially useful for flex planning. My point is that it has little to do with PPR.

What I would be looking for is insights into whether the additional passing catching RB's reduce the scarcity factor for mid-range RB's and thus change draft strategy after the top RB studs. Or whether the top WR's become more valuable or less valuable with the PPR -- is it YAC/TD or receptions that make the top WR's unique and worth having, or do I target Welker?

Again, apologies for the tone because the content is reasonable, it is just mislabeled and not what I was looking for.

 

[steviey]

While we are on the subject, is there a good place for PPR ADP anywhere?

 

[Jeff Tefertiller]

I do appreciate you reading the articles and your response. You are correct that QB's in general are undervalued in most leagues. PPR leagues are becoming more and more prevalent so I thought I would use PPR-specific data. I agree that QBs are valuable, even in PPR leagues, and are taken for granted.

When I look at average non-PPR stats, and compare, there are differences. The top 10 players overall in 2008 were QBs. In fact, 15 of the top twenty players overall are QBs. The issue is that many in PPR leagues are forsaking QBs for WRs. Many in PPR leagues load up on WRs for the reasons mentioned in the article (career length, etc). In addition, the RBs and WRs that finish at the top are different players depending on the scoring.

With the trend of WR values increasing, many frequently pass on QB's early in PPR leagues more often than in non-PPR. Also, one point I would like to emphasize is the impact of "elite" players. For this reason, I included points per game averages.

 

[fridayfrenzy]

I found the articles interesting but it was frustrating to keep reading about the comparisons between different positions (ie. 7 QBs in top 10, and only 3 WRs, etc.). Maybe I just missed it, but I didn't see the relevance to how one position is more valuable than another by stating how many were in the top 10, top 30, etc.

7 QBs in the top 10 does not really signify anything in terms of valuing the positions. The Top 10 could all be QBs, and it still wouldn't show which position is more valuable.

 

[two_dollars]

I disagree that the WR's are harder to predict than the RB's and QB's for PPR. In fact, I believe it's the opposite and WR's are the better trend when you're looking at WR1's vs QB 1's. Look at last year's stats.

Top 10 WR: Andre Johnson, Larry Fitzgerald, Calvin Johnson, Roddy White, Greg Jennings, Brandon Marshall, Anquan Boldin, Steve Smith, Antonio Bryant, Wes Welker, and 11 and 12 were Terrell Owens and Randy Moss. Many of these guys went in the first 4 rounds.

Top 10 QB: Drew Brees, Aaron Rodgers, Jay Cutler, Philip Rivers, Kurt Warner, Payton Manning, Donovan McNabb, Matt Cassel, David Garrard, Chad Pennington, and 11 and 12 were Tony Romo and Tyler Thigpen. You could have had half these guy round 10 or later last year.

Top 10 RB: DeAngello Williams, Matt Forte, Michael Turner, Maurice Jones Drew, Ladanian Tomlinson, Thomas Jones, Steve Slaton, Brian Westbrook, Adrian Peterson, and Chris Johnson. Several of these were grabbed in round 6 or later.

Point being, you have 2-3 spots to fill with WR's. You have a pretty good idea who the guys are that get the targets. Even if one slips a bit, you likely have a good WR2 that you will start every week anyway. With QB's, you only need to fill one spot, and drafting almost any 3 starters is going to end up giving you a good chance at a top 10 QB. I'll bet if you grab stud WR/RB early, or even 2 RB and 2 WR, you can wait and grab a combo of these guys: Garrard, Cassel, Orton, Palmer, Schaub, Edwards.. and get a top 10 QB.

 

[ookook]

I found the articles interesting but it was frustrating to keep reading about the comparisons between different positions (ie. 7 QBs in top 10, and only 3 WRs, etc.). Maybe I just missed it, but I didn't see the relevance to how one position is more valuable than another by stating how many were in the top 10, top 30, etc. 7 QBs in the top 10 does not really signify anything in terms of valuing the positions. The Top 10 could all be QBs, and it still wouldn't show which position is more valuable.

 

[stickboy]

When I look at average non-PPR stats, and compare, there are differences. The top 10 players overall in 2008 were QBs. In fact, 15 of the top twenty players overall are QBs. The issue is that many in PPR leagues are forsaking QBs for WRs. Many in PPR leagues load up on WRs for the reasons mentioned in the article (career length, etc). In addition, the RBs and WRs that finish at the top are different players depending on the scoring. With the trend of WR values increasing, many frequently pass on QB's early in PPR leagues more often than in non-PPR. Also, one point I would like to emphasize is the impact of "elite" players. For this reason, I included points per game averages.

If I understand your points above and the articles, you are saying that based on total points scored in a PPR scoring system among the top players, especially the "elite" players at the very top, QB is a more valuable position to target than WR, even with the added reception points. The data supports this, looking at top 10 (4 QB, 2 WR) and top 30 (12 QB, 8 WR) positional averages over 7 years. The same is true of RB over WR using this criteria of total points. The question I would ask is whether this is the right criteria to guide the decision on which positions to target in which rounds (or how much to budget for each position in an auction). Would it be better to look at the slope of the total points curve at each position and determine where I can optimally gain the most advantage at each position? Do the elite players really differentiate themselves more the next few tiers at that position, especially with the PPR scoring adding points? Where does PPR flatten the curves for RB and WR (QB is not different from non-PPR)?I have done this analysis with my PPR scoring systems (higher yardage and TD points than yours) and have found that QB and WR are equally valuable above RB. Given the crazy fixation so many folks have with stud RB's, this becomes an easy call for me as I can get two mid-range RB's (easier in PPR) and compete reasonably because I kick butt at QB and WR and Defense. And as others in this thread have pointed out, your comments about projection predictability being harder at WR are incorrect. When I looked at FBG's predicted vs. actuals from last year (using my own criteria on what was a successful prediction, YMMV), I arrived at about a 75% success rate for QB and WR, and a 62% success rate at RB. I spend a lot more time tuning the RB predictions from FBG than other positions accordingly, but this also factors into my strategy to focus on QB and WR. This, of course, is scoring system dependent, and may not translate for other PPR leagues.Thanks for listening and discussing, sb

 

[Jeff Tefertiller]

I found the articles interesting but it was frustrating to keep reading about the comparisons between different positions (ie. 7 QBs in top 10, and only 3 WRs, etc.). Maybe I just missed it, but I didn't see the relevance to how one position is more valuable than another by stating how many were in the top 10, top 30, etc. 7 QBs in the top 10 does not really signify anything in terms of valuing the positions. The Top 10 could all be QBs, and it still wouldn't show which position is more valuable.

I apologize if it was frustrating to you. The articles were written less about which position is more valuable than the others, but more of a way to look at trends among the positions. I used points per game as the measuring stick because I was wanted to see what helped teams win. There were some things that emerged (at least for me) like how there has not been a true stud back lately. This has had a trickle down effect on the positional distributions. I started looking at this because many on this board have been touting the hoarding of WRs in PPR leagues, especially in dynasty ... and I wanted to see if it was a worthwhile investment in the short-term (with the agreement that WRs have several advantages over RBs).After writing these, I came away less thinking some positions are most valuable, but a look at value and the effect some players have. The top QBs give their fantasy owners a great advantage (like with the Brees, Peyton, and Campbell example in the articles). In addition, the top passers were not the ones drafted first in a given season. Lastly, I appreciate the effect of the studs more than before. For instance, the difference between WR1 (Boldin) and WR12 (Owens) is the same difference as WR13 (Moss) to WR62 (Hixon). The difference was 7.3 points/game in each case. This tells me that there is little difference in WRs after the first few. I hope this helps and thanks for reading the articles.

 

[Jeff Tefertiller]

Here are the top ten of RB and WR according to MFL PPR drafts in 2008 with the draft pick listed to the left and the finish after the player name. This was on a points per game basis. What I take from this is that the RBs taken as RB1, most finished as RB1 or RB2 while a few of the WRs dropped way off.

1 LT 8 Moss 132 ADP 11 TO 123 Westy 1 Wayne 144 Addai 25 Braylon 405 SJax 4 Fitz 36 MB3 15 AJ 27 Gore 13 Colston 188 Portis 16 Housh 209 Lynch 17 SSmith 410 LJ 27 Chad 4411 MJD 5 Welker 1112 Grant 30 Plax 38And, if we want to go one more spot:13 Bush 10 Holt 43I agree on the QBs you have listed. In the articles, I encouraged owners to wait until a few are taken since the top passers drafted usually do not finish in the top three for one reason or another. While many leagues start more WRs than RBs, the smallish points/game difference after the first few makes it as though an owner can draft value WRs a few rounds later and not give up too much. Thank you for reading and discussing.

I disagree that the WR's are harder to predict than the RB's and QB's for PPR. In fact, I believe it's the opposite and WR's are the better trend when you're looking at WR1's vs QB 1's. Look at last year's stats.Top 10 WR: Andre Johnson, Larry Fitzgerald, Calvin Johnson, Roddy White, Greg Jennings, Brandon Marshall, Anquan Boldin, Steve Smith, Antonio Bryant, Wes Welker, and 11 and 12 were Terrell Owens and Randy Moss. Many of these guys went in the first 4 rounds.Top 10 QB: Drew Brees, Aaron Rodgers, Jay Cutler, Philip Rivers, Kurt Warner, Payton Manning, Donovan McNabb, Matt Cassel, David Garrard, Chad Pennington, and 11 and 12 were Tony Romo and Tyler Thigpen. You could have had half these guy round 10 or later last year.Top 10 RB: DeAngello Williams, Matt Forte, Michael Turner, Maurice Jones Drew, Ladanian Tomlinson, Thomas Jones, Steve Slaton, Brian Westbrook, Adrian Peterson, and Chris Johnson. Several of these were grabbed in round 6 or later.Point being, you have 2-3 spots to fill with WR's. You have a pretty good idea who the guys are that get the targets. Even if one slips a bit, you likely have a good WR2 that you will start every week anyway. With QB's, you only need to fill one spot, and drafting almost any 3 starters is going to end up giving you a good chance at a top 10 QB. I'll bet if you grab stud WR/RB early, or even 2 RB and 2 WR, you can wait and grab a combo of these guys: Garrard, Cassel, Orton, Palmer, Schaub, Edwards.. and get a top 10 QB.

 

[Jeff Tefertiller]

When I look at average non-PPR stats, and compare, there are differences. The top 10 players overall in 2008 were QBs. In fact, 15 of the top twenty players overall are QBs. The issue is that many in PPR leagues are forsaking QBs for WRs. Many in PPR leagues load up on WRs for the reasons mentioned in the article (career length, etc). In addition, the RBs and WRs that finish at the top are different players depending on the scoring. With the trend of WR values increasing, many frequently pass on QB's early in PPR leagues more often than in non-PPR. Also, one point I would like to emphasize is the impact of "elite" players. For this reason, I included points per game averages.

If I understand your points above and the articles, you are saying that based on total points scored in a PPR scoring system among the top players, especially the "elite" players at the very top, QB is a more valuable position to target than WR, even with the added reception points. The data supports this, looking at top 10 (4 QB, 2 WR) and top 30 (12 QB, 8 WR) positional averages over 7 years. The same is true of RB over WR using this criteria of total points. The question I would ask is whether this is the right criteria to guide the decision on which positions to target in which rounds (or how much to budget for each position in an auction). Would it be better to look at the slope of the total points curve at each position and determine where I can optimally gain the most advantage at each position? Do the elite players really differentiate themselves more the next few tiers at that position, especially with the PPR scoring adding points? Where does PPR flatten the curves for RB and WR (QB is not different from non-PPR)?I have done this analysis with my PPR scoring systems (higher yardage and TD points than yours) and have found that QB and WR are equally valuable above RB. Given the crazy fixation so many folks have with stud RB's, this becomes an easy call for me as I can get two mid-range RB's (easier in PPR) and compete reasonably because I kick butt at QB and WR and Defense. And as others in this thread have pointed out, your comments about projection predictability being harder at WR are incorrect. When I looked at FBG's predicted vs. actuals from last year (using my own criteria on what was a successful prediction, YMMV), I arrived at about a 75% success rate for QB and WR, and a 62% success rate at RB. I spend a lot more time tuning the RB predictions from FBG than other positions accordingly, but this also factors into my strategy to focus on QB and WR. This, of course, is scoring system dependent, and may not translate for other PPR leagues.Thanks for listening and discussing, sb

I appreciate the discussion. Basically, this was a look at points per game and not a total points. To sum it up, the elite RBs gave their owners a big advantage most seasons. Yes, the top WRs did well, too, but many seasons, the elite RBs were gold. Given how many WRs produced similarly (on a per game basis), an owner could make up for quality with quantity. Your scoring does differ than what the articles were based on, WCOFF scoring. On the predictability of WRs vs RBs, the WRs taken the highest did not finish high. Out of the top five of each drafted, the WRs only had one finish in the top ten at the position, while the RBs had three and another at eleven.

 

[stickboy]

...I appreciate the discussion. Basically, this was a look at points per game and not a total points. To sum it up, the elite RBs gave their owners a big advantage most seasons. Yes, the top WRs did well, too, but many seasons, the elite RBs were gold. Given how many WRs produced similarly (on a per game basis), an owner could make up for quality with quantity. Your scoring does differ than what the articles were based on, WCOFF scoring. On the predictability of WRs vs RBs, the WRs taken the highest did not finish high. Out of the top five of each drafted, the WRs only had one finish in the top ten at the position, while the RBs had three and another at eleven.

I believe you are saying that stud RB is still the way to go because the scoring curve (pts/game or total pts, doesn't matter) is still steep at the top and flattens in the middle. This is probably true with the WCOFF scoring as the 3rd down backs don't play enough to compete with the top runners who occasionally catch passes for long yards also. So Stud RB is still the way to go in PPR.I believe you are also saying in other parts of the thread that QB is still more important than WR which is surprising to many coaches given the additional points added per reception. If I understand correctly, you are suggesting that the 2nd tier of QB's (maybe QB4-QB9?) are better than a top 5 WR in PPR. I'm not sure of the logic or data behind this, except that there are more WR's available in the middle tier of WR's that have upside. My guess is that the top WR's are not affected much by PPR (most pts come from yards and TDs), and QB's are not affected at all by PPR, so this question is independent of PPR. I don't follow the logic that there are many more WR's to choose from in PPR, as the question is differentiation against your opponent, not pts per game. Example drop-off data within the tiers would be helpful, I don't have WCOFF scoring available to me easily.There also appears to be mixing of the questions of points differentiation and predictability. You provide a number of examples that attempt to show that WR is harder to predict, but the example of 3 out of 10 RB vs. 1 out of 10 WR says that both are crap shoots, not that one is better. My opinion is that we need to try to solve each one independently -- where are the optimal positions to draft to differentiate your team via VBD, and then how to best predict the players that will hit those spots. In my analysis, FBG is weaker at predicting WR than QB or RB but better than ADP. So I look for ways to tune the FBG projections for the parts of the WR curve that I want to target the most. Factors such as injury risk, consistency, and YAC upside potential may require higher weighting when looking at WR projections.Again, thanks for the discussion,sb

 

[greyone]

I enjoyed the ppr articles. Yes it did become a bit redundant. But when you study numbers things can become repetitious. Its the nature of the charts. Any trend off by 1, 2 or 3 is considered big difference by trend analysis. BTW i love trend analysis.

Maybe the disconnect comes from the lack of incorporating players names. People want to see the article and charts at work. There were a few name drops but not enough to see the big picture put together.

Some owners beginner, mid level or expert understands the articles but wants more. I know I did. The application of the articles and which players they apply to would really brighten the pages.

 

[Jeff Tefertiller]

I enjoyed the ppr articles. Yes it did become a bit redundant. But when you study numbers things can become repetitious. Its the nature of the charts. Any trend off by 1, 2 or 3 is considered big difference by trend analysis. BTW i love trend analysis. Maybe the disconnect comes from the lack of incorporating players names. People want to see the article and charts at work. There were a few name drops but not enough to see the big picture put together. Some owners beginner, mid level or expert understands the articles but wants more. I know I did. The application of the articles and which players they apply to would really brighten the pages.

Thank you for the feedback. This is my next step as I find time. I really want to see the turnover year after year for fantasy starters at QB, RB, and WR.

 

[Jeff Tefertiller]

...

I appreciate the discussion. Basically, this was a look at points per game and not a total points. To sum it up, the elite RBs gave their owners a big advantage most seasons. Yes, the top WRs did well, too, but many seasons, the elite RBs were gold. Given how many WRs produced similarly (on a per game basis), an owner could make up for quality with quantity.

Your scoring does differ than what the articles were based on, WCOFF scoring. On the predictability of WRs vs RBs, the WRs taken the highest did not finish high. Out of the top five of each drafted, the WRs only had one finish in the top ten at the position, while the RBs had three and another at eleven.

I believe you are saying that stud RB is still the way to go because the scoring curve (pts/game or total pts, doesn't matter) is still steep at the top and flattens in the middle. This is probably true with the WCOFF scoring as the 3rd down backs don't play enough to compete with the top runners who occasionally catch passes for long yards also. So Stud RB is still the way to go in PPR.I believe you are also saying in other parts of the thread that QB is still more important than WR which is surprising to many coaches given the additional points added per reception. If I understand correctly, you are suggesting that the 2nd tier of QB's (maybe QB4-QB9?) are better than a top 5 WR in PPR. I'm not sure of the logic or data behind this, except that there are more WR's available in the middle tier of WR's that have upside. My guess is that the top WR's are not affected much by PPR (most pts come from yards and TDs), and QB's are not affected at all by PPR, so this question is independent of PPR. I don't follow the logic that there are many more WR's to choose from in PPR, as the question is differentiation against your opponent, not pts per game. Example drop-off data within the tiers would be helpful, I don't have WCOFF scoring available to me easily.



There also appears to be mixing of the questions of points differentiation and predictability. You provide a number of examples that attempt to show that WR is harder to predict, but the example of 3 out of 10 RB vs. 1 out of 10 WR says that both are crap shoots, not that one is better. My opinion is that we need to try to solve each one independently -- where are the optimal positions to draft to differentiate your team via VBD, and then how to best predict the players that will hit those spots. In my analysis, FBG is weaker at predicting WR than QB or RB but better than ADP. So I look for ways to tune the FBG projections for the parts of the WR curve that I want to target the most. Factors such as injury risk, consistency, and YAC upside potential may require higher weighting when looking at WR projections.

Again, thanks for the discussion,

sb

Thanks. In your first paragraph, my point is that there are very few "stud RBs". The points per game bear that out. You can see where the points/game averages for the top RBs shows that there is a shortage of top tier RBs. After the elite RBs, then it becomes a matter of VBD. If you plug in the scoring into a VBD app, you notice that it weighs the RBs at the top due to a shortage of top tier RBs AND a. the balance at WR position and b. hard to determine top wr (IMO). Using points/game averages versus total points does make a difference. It changes the results.My example was not how many of Top 10 ADP finish in Top 10, rather how many that were drafted as Top 5 at position finished in Top 10. This example was used to illustrate the difficulty of finding the top performers at each position. So, yes, three RBs drafted in Top 5 RBs finishing in Top 10 (and one at 11) shows predictability when compared to the WRs. In addition, I know few that were gutsy enough to draft Addai at 1.04.

Hope this answers your questions.

 

[Sweet Love]

I won a PPR league I was in last year for the first time. I certainly did not have the strongest team, but ended up winning it with Steve Smith (traded Cutler to get him), Colston (who was hurt a majority of the year), Walter (late pick) and Breaston (pick up). I had Gore as my RB and did piece meal the rest of the way. I did stumple upon Rivers in a later round, and that solidified it for me. Oh yeah I got Gonzo as TE4 in our draft but ended up paying off in spades. I have had good WRs in this league before (I had Driver Chambers and Tana Moss the year they all went off...about 3 yrs ago) and had better RBs at the time. But I never had a killer QB in this league until last year. It may have been a coincidence, but I am inclined to take QBs 6 and 8 off the board and solidify my chances of getting a top 5 QB, versus my typical waiting on QBs and getting QB12 and 13 and hope someone does OK.

 

[finnref]

I was very interested in this data since i have done similar for my legue,ppr one pt for all positions. QB leads in top 5 with Rb 5 pts behind. WR trails by 25 pts. The tenth player at each position is essentially the same. At position 15 WR takes the lead and subsequently outscores the other positions at all comparable levels. I prefer weekly scoring rather than per game scoring. I disagree that it is easy to replace a top scorer who is out for 5 weeks or so with a 3 or lower ranked player. So I post this for any comparisons with other available data and comments. In our league we play Two QB and Two D so the drafting emphasis changes markedly.

Comparison of Positions by Total FP for Season

Rank QB RB WR TE D

5 306 < 5 > 298 < 25 > 274 173 < 10 > 181

10 239 < 0 > 237 < 0 > 243 127 < 35 > 161

15 204 < 5 > 209 < 15 > 226 101 < 45 > 145

20 166 < 20 > 184 < 20 > 204 83 < 45 > 131

25 133 < 25 > 161 < 25 > 187 73 < 45 > 117

30 102 < 30 > 140 < 30 > 173 61 < 35 > 97

35 na ---- 123 < 35 > 159

40 na ---- 113 < 40 > 150

45 na ---- 101 < 40 > 141

50 na ---- 91 < 40 > 130

Top 5 QB and RB outscore WR. Top 10 QB, RB, WR score the same. Top 15,20,25,30 WR outscore RB and QB of same rank. Top 15,20,25,30 RB outscore QB of same rank. WR 20 and below match or outscore QB and RB of higher rank. Top 5 TE equals Top 30 WR. Top 10 TE equals Top 35 RB or Top 50 WR

Comparison of Positions by Points per week.

Rank QB RB WR TE D

5 20 20 18 11.5 12

10 16 16 16 8.5 11

15 14 14 15 7 10

20 11 12 14 5.5 9

25 9 11 12.5 5 8

30 7 9 11.5 4 6.5

35 na 8 11

40 na 7.5 10

45 na 6.5 9

50 na 6 8.5