 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||
 |
 |
||||||||||||||||||||||||||||
“Ch…Ch…Ch…Changes!” Turnover at the Top: Looking at FF Volatility Last year’s fantasy football season is over and most leagues are stored away in mothballs until next summer. But for us in the Zealots dynasty leagues, fantasy football is a year-long hobby. Here we’re already working towards next year, and are taking advantage of our “off-season” to tweak our rosters. Auctions are currently underway during our formal Free Agency period, and then next up is the all-important rookie draft in early May. As Zealots work on their off-season rosters, some lucky dynasty team owners are enjoying the luxury of having of the top fantasy names -- prized studs who led the leagues in scoring at their position last year. These workhorses either helped lead teams to a playoff run or form the solid nucleus of a rebuilding effort. Owners fortunate enough to have these elite players see them as flashes of gold that they can’t help but feel smug about. But should they? Should a fantasy football owner really breathe a confidant sigh, knowing that these “studs” will continue to perform at a high level? Just because they have some of the top players from this last year, how certain can they be that those same players will be top again next year? Is there any need to worry about turnover at the top, and if so, how big is that turnover? Does it vary depending on what position we’re talking about? What exactly are the odds that the studs from last year will actually be there again next year? That’s what I’d like to look deeper into. So in this article I’ll investigate the change among the top scorers at each position from year to year, using scoring results from four years of Zealots history. I’ll look at turnover and the range of scoring at “the top,” and talk about how we can use the numbers to create a “volatility score.” We’ll then see how this compares for each FF position. The Top, and QBs The first thing we need to do is to define “the top.” I’m going to call it the top twelve at each position, because there are 12 teams in each Zealots league. That means, on average, each Zealot classic-league team-owner might hope to have one of these top scorers on their roster at each position. I’m also going to use Zealots full-season regular-season scoring, as downloaded from MFL for the four 2003 – 2006 seasons, for all point values Next, we’ll start the scoring review with the Quarterbacks. Here are the top-scoring QBs in Zealots from the last four years and their associated scores:
These top-twelve QBs are ranked by their overall score for a given year. The column “Next” under each year is the change in ranking that the player had in the next year. For example Peyton Manning and Duante Culpepper were the #1 and #2 QB scorers in both 2003 and 2004, so there was zero change in their scoring positions from 2003 to 2004. Brett Favre was #6 in 2003 and then #4 in 2004, so he had a change of 2 spots upward in the next year, and it’s listed as a positive 2. On the other hand, Matt Hasselbeck went from #3 in 2003 to #12 in 2004, a change of nine spots downward, so his change in the next year is a negative nine. Now note that of the top twelve QBs in 2003, Steve McNair, Jon Kitna, Brad Johnson, and Jeff Garcia were all not in the top 12 again in 2004. So the “(out)” indicates they fell out of the top 12 in the next year. This is what we’re looking for – since these 4 QBs failed to make the list in 2004, there was a turnover between 2003 and 2004 of 4 out of the 12 QB slots, or 33.3%. These values are shown in the first two summary rows below the 2003 numbers. So looking at the 2004 list, we see that 5 QBs out of top 12 in 2004 failed to make the top 12 in 2005. This gives a turnover rate of 5 out of 12, or 41.7%. Between 2005 and 2006 there were 6 QBs that failed to stay in the top 12, or 50.0% -- this means fully half of the top 2005 QBs didn’t make the top 12 again in 2006. Taking the total numbers for the past three year turnovers at QB, we get an average top-12 turnover rate of 5.0, or 41.7% for QBs. So we can estimate that roughly 5 of the QBs in the top 12 in 2006 will no longer be in the top 12 in 2007. They may not do horribly, and they may make it back to the top 12 in a later season, but odds are that they won’t be one of the “top” QBs in 2007. Which ones specifically? Of course, we can’t say – just that we can reasonably expect a turnover of five. QB Scoring Impact OK, so now we have an estimate of five QBs that we can reasonably expect to fall out of the top 12 in scoring next year, but what else can we say about “changes at the top?” Well, a lot depends on the scoring range: If there is very little difference in scoring between the #1 QB and the #12 QB, then falling out of the top 12 next year probably won’t matter that much. In that situation where there’s very little drop-off in scoring, your QB can probably still score nearly as many points at, say, #15 overall as he would if he were in still in the top 12, and there’s not a whole lot of importance to having a “top” QB. But what if the #12 QB scores less than half of what the #1 QB does? In this case, you can likely expect that drop-off rate to continue, and you may find yourself at a significant scoring disadvantage by not having one of the top QBs, and that turnover rate will have much more impact to your roster. Looking back at the data for the 2003 QBs, we see that perennial stud Peyton Manning scored 328.88 points as the #1 QB, and Jeff Garcia scored 261.26 points at #12. That means that there was a range of (328.88 – 261.26), or 67.62 points between #1 and #12. This 67.62 point range is 20.6% of Peyton’s top score, meaning that there was a drop-off in scoring of 20.6% between #1 and #12. This drop-off rate is shown in the lower summary row of the table. Note that we’ll record drop-off percentage instead of raw point totals since it will allow us to compare year-to-year and across the positions regardless of differing scoring rules or ranges. Now moving to 2004, we see that both Manning and Culpepper exploded for over 400 pts, so that there was a much larger scoring range between #1 and #12 – In 2004 there was a drop-off of 44.0% in scoring between #1 Peyton and #12 Hasselbeck. In 2005, the scoring drop was 25.8% between #1 and #12, and finally, it was 38.8% last year in ‘06. Averaging these four scoring ranges, we can estimate that the drop-off in scoring for QBs between #1 and #12 will be around 32.3% in any given year. So now we can put these conclusions together and estimate that 5 of the top QBs from 2006 will not be back in the top 12 in 2007, and that these QBs will likely score less than 67.7% (or down 32.3%) of whatever the #1 QB scores. This alone should be enough information to make anyone recognize that having a top-12 QB this year is nowhere near a guarantee that they’ll be in the top-12 next year, and that it may have a pretty significant impact on their scoring – they’ll get over 30% less points than the #1 QB if they aren’t in the top 12. QB V-Score But we can also use these numbers to create a “volatility” score that will allow us to compare top-12 turnover and its scoring impact across all positions. I’m considering volatility to be a combination of how much turnover there is, and how badly that turnover will affect a roster. In this way, both the turnover rate and the scoring range are important and related. That’s because even if there was total turnover in the top-12 from year to year (100% turnover), its impact would vary depending on the scoring range. If there wasn’t much difference in the scoring, then turnover by itself wouldn’t have much impact – a player out of the top-12 would then score similarly to a player in the top-12. But at the other extreme, if all the scoring dropped off in the top 12 slots for a position, then #12 would have 0.00 pts, and the drop off would be 100%. In this case it would be all-or-nothing to have a top player, and the impact of turnover would be huge, even if it was only a small turnover rate. So we can try to combine these two estimates. If we multiply the turnover percentage by the scoring drop-off percentage, we get a number we can use to measure “volatility.” If we take this number and base it against a mid-range value of 50% turnover * 50% scoring drop-off (which gives an overall result of 25%, or 0.25), we get a more user-friendly number that allows easier comparison of values when looking across the positions. So we can calculate a Volatility Score (“V-score”) for each position by multiplying the turnover rate by the scoring drop-off percentage and then dividing it by our mid-range base of 0.25: ( (% Turnover) * (% Scoring Drop-Off) / 0.25 ) For the QBs, our estimated V-score becomes: ((41.7% * 32.3%) / 0.25) = 53.8 (QB V-score). Now we’ve got a number for volatility. What does this value mean? If the end result was 100, then it would mean that there is around 50% turnover and a 50% scoring drop-off for the QBs, just like our mid-range base estimate. Since the QB V-Score is less than 100, it tells us that one or both of those percentages are less than 50% -- there’s not a full 50% turnover and/or the scoring drop-off is less than 50%. And in the case of the QBs, both are true. But this QB V-score of 53.8 becomes much more useful once we calculate similar scores for all the positions. Then we can compare the V-scores to determine which positions are more volatile. So we’ll go on and do these calculations for each position, but first, one last side note on the QBs: Look back at the QB data, and notice that this Peyton Manning guy seems to be pretty good. He’s been the #1 QB three of the last four years, and only dropped to third the year he wasn’t #1. Tom Brady is the only other QB who’s been in the top 12 since Zealots started, and he was #10 in both 2003 and 2004. We calculated a general V-score for the overall QB position, but obviously not ALL QBs are going to have an equal volatility. The V-score tells us about the general top-12 at the position of Quarterback, not about specific QBs. If we wanted to, we could create some type of volatility score for each individual QB based on their specific changes in scoring and position year-to-year, but that wouldn’t be very useful – it’s hard to predict trends from a single player. Instead, we want to make general roster predictions based on the overall position of QBs, and then compare that position with others on the rosters. So for our purposes, the general QB V-score of 53.8 will be more useful – we want to look at the forest, not individual trees. The Top RBs So now let’s look at the changes in the top-12 for the all-important Running Back position:
Looking at the RBs, we can see that there is an even higher average turnover in the top-12 than there was at QB. Here the number of RBs who fell out of the top-12 was 8 from ’03 to ’04, 6 from ’03 to ’04, and 7 from ’04 to ’05 – this gives a yearly average of 7.0 (or 58.3%). So we can estimate that for any given year, over half of the top-12 RBs will fail to return to the top-12 in the following year. So how is this turnover rate impacted by the scoring range? Here too it’s much worse for the Running Backs than it was for the Quarterbacks. The least amount of drop-off of 34.7% in 2004, when Portis’ 197.00 pts at #12 were 34.7% lower than Alexander’s 301.60 pts at #1, and the greatest drop-off was last year when Tomlinson broke 400.00 pts, giving a drop-off down to Ladell Betts’s #12 score of 187.78 of over half (56.2%). Overall, a yearly estimate for the scoring drop-off in the top 12 is the average of 46.3%. So it will be a pretty significant scoring impact to have a RB drop out of the top-12, and there’s a better-than half chance they will. Calculating our “V-score” for the RBs, we get: ((58.3% * 46.3%) / 0.25) = 108.0 (RB V-score). This is twice as high as the QB V-score of 53.8, and indicates much more volatility among RBs than QBs – not only are they more likely to fall out of the top scoring range, the average scoring impact will be significantly greater. Of course, the obvious exception to this rule is Tomlinson. He’s the only RB who’s stayed in the top-12 for all four years of Zealots leagues. So as with QBs, our RB V-score captures the overall trend of the top-12 RBs in general, but there will be variation in the volatility of individual RBs in any given year. The Top WRs Moving on to the Wide Receivers, we see:
Now we see a ton of turnover. At least 8 WRs have fallen out of the top 12 every one of the last three years. Looking at the table, we see that the average turnover rate in the top 12 for WRs is 8.3, or 69.4%. But again, there are exceptions – Marvin Harrison, Chad Johnson, and Torry Holt have all managed to stay in the top 12 for each of the last four years. Although there is generally a lot of turnover at the Wide Receiver position, one saving grace is that the scoring impact is less than we’ve seen at RB, and even less than at QB. The average drop-off in scoring from #1 to #12 for the last four years is only 30.4%. This depth of scoring helps offset the negative effects of the huge turnover rate, as we can see in the V-score for the WRs: ((69.4% * 30.5%) / 0.025) = 84.8 (WR V-score). This is significantly higher than the 53.8 V-Score for QBs due to the high WR turnover rate. But the slow drop-off in scoring helps the value stay lower than the V-score of the RBs (108.0), where the scoring drop-off is more severe. The Top TEs Here are the numbers for the top-12 Tight Ends for each of the last four years:
This is the reverse situation of the Wide Receivers. Here we see much less turnover (an average of 5.0 or 41.7% -- identical to the QBs), but the drop-off in scoring is the highest yet. The average drop-off between the #1 scoring TE (Gates in three of the years) to the #12 TE is 53.5%. This means the #12 TE usually scores less than half of the #1 TE. So if a TE does fall out of the top 12, they’re also going to score pretty poorly given this high drop-off rate. In terms of the volatility V-score: ((41.7% * 53.5%) / 0.25) = 89.2 (TE V-score). A little higher than for the WRs, even though there is significantly less turnover at TE than at WR because of the huge drop-off in scoring rates. The Top PKs The common FF thought is “a kicker is a kicker.” Let’s see if the numbers bear that out:
Well, the turnover rate is still high in many ways – almost half of the Kickers in the top 12 each year fail to make the top 12 in the next year (an average of 5.7 or 47.2%). But the turnover rate is not much more than at TE or QB. And the drop-off in scoring between the #1 PK and the #12 PK is the lowest we’ve seen yet: Only 27.8% on average for each of the last four years. So if a Kicker does drop out of the top 12, they’re still probably going to score decently in most cases. This shows up in the PK V-Score, too: ((47.2% * 27.8%) / 0.25) = 52.5 (PK V-score). This means that Kickers have the lowest V-score we’ve seen so far, just under QBs. In many ways, this low level of volatility backs up the common thought that “a kicker is a kicker” – they’re comparatively stable from year-to-year, and there is less variation in their scoring across the rank within a given year. The Top DLs So now let’s check out the defensive side of the ball, starting with the Defensive Linemen:
Here we see the same high level of turnover as the WRs – an average of 8.3 Defensive Linemen (69.4%) falling out of the top-12 on average each year. But as with the offensive positions, there are exceptions: J. Taylor has stayed in the top 12 each of the last four years, finishing at #1 last year. But for the DLs, the scoring drop-off is a little worse than it is for WRs. There has been an average scoring drop-off of 35.3% from the #1 DL to the #12 DL each year. Calculating the V-score for the D-Linemen, we get: ((69.4% * 35.3%) / 0.25) = 98.0 (DL V-score). This is the second-highest V-score we’ve seen so far. Although many FF fans think that there’s a ton of depth at all the defensive positions, there is a significant amount of volatility among the top DLs. This suggests that the scoring advantage from having some of the top DLs is worthwhile, and that these top DLs change quite a bit from year to year – this defensive position is worth tracking closely to try and gain an edge at. The Top LBs Moving on to the Linebackers, we see:
High turnover again, but a little less than for WRs and DLs: An average turnover of 8.0 each year (66.7%). But check out the very low drop-off in scoring. There’s only an average 22.1% drop-off in scoring between the #1 LB and the #12 LB – It’s true that linebacker is a deep position. Here’s the V-score for the Linebackers: ((66.7% * 22.1%) / 0.25) = 59.0 (LB V-score). This is a V-score down in the range of the QBs and PKs, even though there’s a 2 out of 3 chance that a given top-12 LB will not return to the top-12 in the next year. The slow drop-off in scoring range really helps settle the volatility. At LB, the low V-score seems to bear out the common thought about there being plenty of defensive depth to fill your roster. The Top DBs Lastly, let’s look at the numbers for the Defensive Backs:
Now we’re talking turnover! An average of 9.0 of the top 12 DBs will fail to stay in the top 12 in the next year – this is a huge average turnover rate of 75%. But again, just like the LBs, there is a very low drop-off in scoring at the DB position. This is another very deep position, and the average drop-off from #1 to #12 is again only 22.1%. This helps give a lower volatility rating than this position might otherwise have: ((75.0% * 22.1%) / 0.25) = 66.2 (DB V-score). So even if a given top-12 DB falls out of the “top” in the next year, odds are they will still have decent scoring, as can be seen in the V-Score. Again, the “plenty of defensive depth” theory holds for DBs. Comparing V-Scores and Summarizing OK -- Let’s summarize and compare across the positions:
As we saw earlier, the most volatile position is RB, as shown by their high V-Score. The RB position has both a turnover rate of over 50% and a relatively high drop-off rate in scoring. The least volatile is Kicker, where the depth of scoring (and low scoring drop-off rate) and a lower turnover rate combine to give the most stability. The other positions range in between. So what does this all mean? Well, in a lot of ways, we just added hard numbers to what many successful FF owners already knew at a gut level. But we can still take several key points away from all of this:
Comments, questions, and feedback welcome at WritingStaff@zealotsfield.com. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|||||||||||||||||||||||||||||||||||||||||||||||||||||
|
©2010 Zealots Field. All Rights Reserved. Zealots Field is not associated in anyway, affiliated with, endorsed, or licensed by the National Football League, any NFL team, or the NFLPA.
|
