Dynasty League Football

Uncategorized

Maximizing Your Top Draft Picks By Minimizing Churn

It’s the age old question that all of us face when our turn to pick comes up in a start-up dynasty draft – which of these players holds the most long term value for my team?

It’s no secret that sustained high-level production is the life-blood of any successful dynasty team. One year wonders are fun to ride when you can, but depending on those players to keep you in continued contention for your league title is best left for those players who focus solely on re-draft leagues. None of what I’ve said so far in this article is earth-shattering or ground-breaking research for dynasty owners – we all look for players who produce year in and year out.

What I’m about to say next, however, just might be a bit of a revelation to many of you dynasty owners, probably even you reading this routinely sabotage your teams by selecting players, or rather players from positions, that have the lowest longevity with your highest picks. If a dynasty owner is in the dynasty experience for the long-term, one would expect that the players from the positions with the longest longevity would be selected with the highest picks. The problem usually occurs when owners plan for the dynasty experience with their brain, but only after drafting with their heart.

“Whoa, whoa,” you’re saying, “I put a lot of research into my picks, I take players who are strong and look great for the long term.”

I have no doubt that many of us share that sediment, myself included. However, some new research I’ve recently conducted has led me to dramatically reassess both how I value players and which positions I value the most.

Would you draft a player who was a top ten performer for his position last year with a first round pick? Depending on the player you would certainly consider it, who wouldn’t right?

What if you were presented a choice between two players from two different positions, both of whom were top ten players in their position, both the same age and each score roughly the same amount? What you may not realize is that one player’s position, on average, drops out of the top ten a full half year earlier than the other player’s position – that difference becomes even more exaggerated if the players in question have been in the top ten for their positions more than once. If the players in question have been in the top ten more than once, one player’s position appears in their top ten almost a full year longer than the other position. The decision seems elementary at this point – you select the player who will appear in their top ten an extra year, it just makes sense from a value standpoint, right?

 

Why then do nearly all of us covet running backs with first round picks and wide receivers with second or third round picks, but almost totally ignore quarterbacks until much later?

A simple glance at the MyFantasyLeague.com ADP for 2012 shows only five quarterbacks in the top 30 picks. Running backs, however, dominate the rankings with 13 entries and wide receivers account for nine of the top 30. Now, granted, this ADP combines re-draft and dynasty ADP rankings, but look at any site that focuses on dynasty football and the trend of valuing running backs over quarterbacks is prevalent everywhere – DLF included.  However, when you compare the average length of time any one player appears on their position’s top ten list running backs and wide receivers rank at the bottom, and not by a little, by a lot as you can see below:

I’ve tried to cram a lot of information important to this article into the chart you see above. First, let’s concentrate on the bars themselves; the blue bars represent all of the players who appeared in the top ten per position from 2002 through 2011, while the red bars represent players who appeared more than once between 2002 through 2011. The numbers above each bar represent how many times on average a player appears in each list. The numbers in the parentheses under each position are the number of players who appear on that position’s list between 2002 and 2011. The numbers in the brackets under each position are the number of players that appear more than once on that position’s list between 2002 and 2011.

As I pointed out earlier, you can clearly see that running backs and wide receivers, on average, only stay in the top ten for 2.29 years and 2.21 years, respectively, while quarterbacks, the position with the most top ten longevity, stay in the top ten nearly half a year longer on average.

If we focus solely on players who repeat their top ten performances for two or more years then the difference is startling. Your average running back that repeats a top ten performance doesn’t stay in the top ten much longer than the rest of his top ten running back compatriots, only staying there 3.07 years, a difference of 0.78 years. Wide receivers fare slightly better. A wide receiver who appears more than once in the top ten averages 3.45 years in the top ten, an average of 1.24 years, nearly a full year and a quarter longer than the average wide receiver appearance.

Reappearing quarterbacks stay longer than both running backs and wide receivers, appearing 3.78 years on the top ten, but the increase from the overall quarterback appearance average to the reoccurring appearance quarterback is exactly one year longer. Tight ends reign supreme over all positions, however, boasting players that reappear on the top ten for 3.82 years on average, an increase of 1.38 years over the average for all of the top ten tight ends from 2002 through 2011.

The case seems pretty cut and dry at this point, however, we need to look a little further to get a full picture of what’s going on here.  The numbers on the above graph under each position play an important part in interrupting this data. This can be viewed as the churn rate for the top ten for each position. Since we are looking at data that encompasses ten years’ time with ten top players at each position per year, we can obviously expect 100 potentially different players if there was a 100% churn rate. The first number under each position is the number of different players who appear on that position’s top ten during the ten year period in question. This can essentially be viewed as a percentage, so wide receivers have the largest churn rate at 47%, followed by running backs that had a 45% churn rate, then tight ends at 43% and a nice drop off to quarterbacks who only churned at 36%.

Scratching beneath the surface, a couple of things should immediately stand out to you. First, the difference in churn from the more highly valued wide receivers to the often ignored quarterbacks is an amazing 23%, not simply an 11% difference. No, my math isn’t wrong, you don’t compare the totals to the original 100%, since we are comparing the positions we compare the numbers from each position. Therefore, the difference from 36% to 47% is a full 23% increase (100 – [(36/47)*100]). In more understandable terms, this means that if you select a wide receiver who has appeared in the top ten just once, over a quarterback who has ranked in the top ten at his position just once, you’re 23% more likely to see that wide receiver fall out of the top ten the next year  (Jordy Nelson, I’m looking squarely at you)! You may be thinking this same examination needs to be done by the serious dynasty owner right away!

Well, not really, here’s a handy chart to refer to:

What we have to do at this point is divide the number of re-appearing top ten players by the number of all the top ten players. Doing so yields us the following – clearly, as a dynasty fantasy football owner drafting a player who could fizzle out one times out of four over a similarly talented player is just not acceptable! What happens if we have to make a decision between two highly talented players who have appeared in the top ten of their position more than once? Here’s where things get a bit complicated, but also where some interesting facts are unveiled.

Quarterbacks: 23/36 = 0.6389 then 0.6389 * 100 = 63.89%

Running Backs: 28/45 = 0.6222 then 0.6222 * 100 = 62.22%

Wide Receivers: 23/47 = 0.4894 then 0.4894 * 100 = 48.94%

Tight Ends: 22/43 = 0.5116 then 0.5116 * 100 = 51.16%

Great, lots of numbers, but what in the world does it mean? I’ll walk you through it step by step.

We are taking the number of players who appear in the top ten multiple times and dividing that number by the number of players in the ten year period from 2002-2011 who have appeared in the top ten for that position. We are then given a number between zero and one. We then take this number and multiply that by 100 to get a percentage. This is the last number in each of the lines above.

Uhhh, OK, so what? What does that number mean?

It’s simple – that number, in conjunction with the average number of years that re-occurring players who appear in the top ten, shows us the true dynasty worth of top ten players who aren’t flukes. If we encounter a higher percentage here with a higher average number of years, this tells us how highly we should value a given position, quarterback for instance.

  • More top ten quarterbacks who appear in the top ten for one year tend to reappear in the top ten more than one year.
  • Quarterbacks also tend to stay in the top ten longer than running backs or wide receivers.
  • Fewer tight ends re-appear in the top ten multiple years, but the ones who do reappear tend to be staples.
  • Running backs tend to reappear in the top ten more than one year, but they stay in the top ten far shorter than any other position.
  • The wide receiver position tends to be the worst long term investment with fewer wide receivers appearing in the top ten than any other position and those wide receivers who do reappear in the top ten have the second shortest stay in there out of the four offensive positions.

Dynasty owners always aim for players who reside in that “sweet spot” where longevity intersects with top notch production. The numbers from the past ten years however seem to indicate the players most dynasty owners target early in initial drafts actually hamper the long term viability of their team’s success. Those owners looking to be successful longer than their league mates would be strongly advised to target young quarterbacks who have appeared in the top ten two or more times over a similarly producing player at any other position. Additionally, targeting wide receivers with high draft picks appears to be a poor investment long term.

Clearly there are exceptions to every rule, Calvin Johnson and Larry Fitzgerald come to mind immediately, however, chasing exceptions is a losing man’s game. The number I’ve laid out should serve any dynasty owner with a great roadmap from which to value players who post similar fantasy points, but play at two different positions. Doing so has always been a hard proposition for most of us, yet now with firm numbers in hand, the decision should be much easier – you pick the player from the position with the lowest churn and longest longevity.

It should be very interesting to see how drafting strategies could be modified based on this research. One strategy perhaps would be that you pass on running backs and wide receivers with your picks in the top two rounds. Instead, you might opt for a high end young quarterback and tight end in rounds one and two. From the third round on you hold your initial draft as most others would grabbing the best young running backs and wide receivers available.

Based on research still in its infancy, this strategy is, admittedly, risky. However, the numbers are solid and point to this strategy being very sound. Grabbing a combination such as Aaron Rodgers, Cam Newton or Matthew Stafford at quarterback and Rob Gronkowski or Jimmy Graham at tight end is very feasible in this drafting scenario and many young, high profile running backs are available in the third and beyond. It’s hard to argue that a team built around those studs could fail long-term, which seems to lend a lot of credence to this research.

Now get out there and truly build your team for the true long-term!

 

31 Comments
0 0 votes
Article Rating
Subscribe
Notify of
31 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Richard Smith
10 years ago

Very interesting.

Chris Russell
10 years ago

This is exactly what I did drafting from the 8 spot in a 12 team dynasty PPR startup. It is not about reaching but taking the positional longevity advantage as the coin flip in who to draft.

My first four picks were Stafford, Gronkowski, Harvin and Jonathan Stewart.

Risky of course but longevity buys more time for success from the picks as the rest of your team develops. Otherwise you end up chasing your tail to replace the early RBs that are falling off as your team is ready to compete.

Though if your goal is to win early in the leagues existence then go for the best player regardless and build again sooner.

Josh Gans
10 years ago

Very interesting article, thanks for sharing and the research!

Eric MacKenzie
10 years ago

I agree with the premise here, and my team reflects that, but the article doesn’t seem to account for the fact that most leagues are starting multiple running backs and wide receivers and only one QB & TE. That increases the value of RBs and WRs significantly in my view. I need a top ten QB, but if my top ten RB becomes a top twenty RB, he’s still a starter.

Cyrus Miller
10 years ago

Hate to be the grammar police, but you wrote: “share that sediment”

It should be “share that sentiment.”

Minor, but thought I would bring it to your attention. I’ll comment more when I finish the article.

Cyrus Miller
10 years ago

Okay, first, to disagree with some things:

1- It seems odd to compare the top ten QB’s (in a start 1 QB league) to the top 10 RB and WR (in a start 2+ RB and WR league). Also, the distinction of top ten is kind of silly (if someone is #11, do they not count as top ten?). What I would much prefer is statistical analysis comparing:

a) Elite levels of play– for example, QB and TE should only include the top 6, whereas RB and WR can include the top 12. Why? Because having the #10 QB isn’t an advantage, it means you are at a disadvantage. Getting an adequate QB is not going to help you win. You want an elite QB. Whereas landing the #10 RB might be an advantage (if it is your RB2) or at least keeps you at average.

b) Effective levels of play– This is where a top 10 QB comes in. They might be effective. For example, if Brady is top 3 most years, but dips to top 10 one year, he is still effective and you can start him. That means you don’t need to trade for a replacement. For QB and TE, I think this means top 12, whereas for RB and WR, this means top 24. Having a RB who is top 5 one year and #18 the next is still going to help your team.

Also, I don’t think we need to adhere strictly to the top 12 or top 24… for example, if the top 10 QB’s are clearly better and the #11-16 are clearly worse, don’t reward the #11 and #12– don’t include them. However, if the top 15 QB’s are clearly better and the #16 QB is clearly worse, then include all top 15.

Finally, some statistical analysis is needed to account for injury (if they are injured during a game, don’t count it). Maybe some other things could be included, I am not sure. (Inconsistent players who score half their points in two games shouldn’t count as elite, but would be effective, for example)

2- To agree with some things.
While I don’t think the statistics shown are that relevant, I have to agree with the general theory that elite QB’s and elite TE’s are a great way to build a team. I also include elite WR in this, as I think the statistics above are flawed. For example, a player in a great offensive scheme or used in a peticular way (Laurent Robinson, for example) might show up as a top 10 player, but then disappear. That leads the WR position to seem like it isn’t consistent. However, an elite WR can be very consistent, so all this statistic shows is that WR represents opportunity as many non-elite players can end up in the top 10.

In case I didn’t explain that well, here it goes again: If you are drafting a player in the 2nd or 3rd round of a startup, you are getting a stud WR like Julio Jones, Hakeem Nicks or Percy Harvin. Those players have a very high chance of scoring in the top 12 unless injured. However, the top 12 will also include several randoms like Laurent Robinson that won’t repeat. Your analysis indicates to avoid the WR position because they have a low consistency rate, yet if you get a top WR, the consistency rate is fine.

Anyway, I agree with the premise that elite QB, WR and TE are a great way to build a team. That is why I took a trade that had me giving up the 1.01 pick to get Jimmy Graham + a lot extra. I feel like Graham will be effective for the entire duration of Richardson’s career and then some, and the extra helps me improve at other positions.

What I don’t agree with is undervaluing RB’s because of their short shelf life– while in their prime, a top RB has much higher trade value and is much harder to replace. Therefore, the best strategy is to take a sure thing at RB and ignore the question mark RB’s until after you fill out your roster.

Cyrus Miller
Reply to  TheFFGhost
10 years ago

I’ll try to reply in order of yours:

Preamble– While flex positions and 2 QB leagues mix it up a bit, I think everyone can get behind a baseline of 1 QB, 2 RB, 2 WR and 1 TE for this sort of analysis. I just think ignoring the RB’s and WR’s from 13-24 can lead to inaccurate stats.

1a) While me saying this might prove your point, I think it makes the statistics you use to support your point lose value. For an extreme example, if you are including the top 10 at the position, there could be a QB8, QB9 and QB10 that are always the #8, #9 and #10 respectively. That leads your statistics to show that QB is consistent and always performs. But as I said and you agreed, being #8-10 is actually a disadvantage.

1b) You are again relying on your statistics when I am questioning them. That three year window you speak of is only top 10– but if you include top 24, it could be a 5 year window. We don’t know, as I haven’t run the numbers. So I am not persuaded by losing the expected production.

As for injury, I only meant the individual games that an injury occurs. For example, a RB might be amazing for 5 games, and then in week 6 gets hurt before doing anything– that counts as a 0. Then he might be out for 3 games, and come back and be amazing for 7 more. Total, he was amazing for 12 games, but he had one 0 (due to injury) and then missed three games. If you only use cumulative stats, he doesn’t show up. If you use average ppg (which is always my preference) he shows up as above average. If you use average ppg ignoring the 0 due to injury, he shows up as amazing. See my point? I prefer someone who has elite upside when healthy then someone who plods along but remains healthy. The elite upside is either your starter or he is injured and you start someone else.

2) If that is your message, to avoid one year wonders but go after the stud WR’s, then I agree completely. I didn’t get that from reading it the first time.

I agree completely comparing Graham to Richardson. But not all TE’s are Graham level, so I only think this advice applies to Graham and Gronk and young stud QB’s (Rodgers, Stafford and Newton yes. Brady and Brees, no.)

Where I see it misapplied is people drafting Luck or RG3 ahead of solid RB’s because they will be good longer. However, Richardson or Doug Martin will have a much higher trade value than Luck or RG3, and I would draft them much earlier than the QB’s, despite the longevity of the position. I can get elite performance (assuming they pan out) and then I can trade them mid-career for a huge haul. Luck and RG3 might become very good, but you will never get the same trade value out of them.

I built one team the way you advocate, going WR and QB (this started in 2008 or 2009) early and taking fliers on RB’s. I won the Championship a year or two after it started, despite stupidly taking Braylon Edwards (your 1 year wonder rule) over Andre Johnson. However, my question mark RB’s that I took fliers on only performed that one year– now I need to overhaul my RB’s completely and it is a challenge.

This summer, I built a team going against what you advocate– I took Arian Foster at 1.02 and Matt Forte at 2.11. I think that they give me the best chance to win, and while they might fall off a cliff in a few years, I can plan ahead and build around it. I would have taken Julio Jones or AJ Green or Graham or Gronk in place of Forte, but they didn’t last to me.

Cyrus Miller
Reply to  Cyrus Miller
10 years ago

I should add, ignoring injured games also extends to ignoring week 17 games where they don’t play.

Basically, I download the data from MFL for my favorite leagues– then I run a Macro I wrote that sorts Weeks 1-17 for each individual player and ignores the bottom two and the top 2 performances. (Challenging to automate only because you have to take into account that if a player is injured, they might only play 12 games or fewer.) Then it averages what is left.

I find this gives a true picture of the consistent value of a player– often people look at weekly averages and don’t see that a player got a 0 in Week 17. That 0 could cost them about 1-2 ppg.

Kenneth Leider
Reply to  Cyrus Miller
10 years ago

I’m with Cyrus here. Comparing the top 10 at each position doesn’t answer an interesting question to me. Having a top 10 QB or TE is not necessarily an advantage, while having a top 10 RB or WR probably is. This is the apple to orange comparison. The interesting questions to me are: 1) How long is a player at a position likely to score better than average. 2) How long is a player at a position likely to score respectably. In the first case comparing the top 5 QBs and TEs to the top 10 WRs and RBs, in the second case comparing the top 10 QBs and TEs to the top 20 WRs and RBs.

It’s hard for me to imagine the top 5 churn for QBs would be that much better then the top 10 churn for RBs and WRs.

Cyrus Miller
Reply to  TheFFGhost
10 years ago

(Quick aside– the RB’s were not established at the time of the draft, and other than the one year of success, I have always struggled at RB. To me, it is what made me want a solid RB in Foster/Forte to anchor my team in the new startup)

Honestly, I have enjoyed the discussion with you, but I am left wanting to do the analysis myself. As I said, I would remove the games that are outliers (both good and bad) so it removes games that the player was injured and games that the player exploded. (Remove the 250 yard game for Demarco Murray last year, for example). Alternatively, I sometimes use a different method that includes the top and bottom two scores but removes the outlier portion. For example, if the #3 game scored 20 points, and #1 and #2 were 40, I wouldn’t just ignore the fact that the player had two great games– I would lower them to about 20 and include them. Picture a box and whiskers plot, I want to remove the whiskers.

I guess here is a compromise to the other topic we are debating. In addition to the numbers already run, we can run additional calculations to make more sense of things. For example, you have the % of QB in the top 10 consistently. Now, we can figure out how many of the top 6 QB are in the top 6 consistently. Likewise, you have the % of RB in the top 10 consistently. Now we can extend it to the % of RB in the top 20 consistently.

stolenmeat
10 years ago

“I have no doubt that many of us share that sediment, myself included.” I think you mean “sentiment”. Sediment is a naturally occurring material that is broken down by processes of weathering and erosion of rock.

Mark Tavares
10 years ago

Im very interested on seeing a comment to Cyrus’ last post about taking Richardson/Martin over Luck/RG3. I posted a similar question in the Premium forums – where I suggested trading 1.04 (RG3 or maybe even Luck) for Jamaal Charles. The DLF writers all agreed that Jamaal is the better value and make the trade on Draft Day… but this article seems to disagree. If im understanding this correctly, Luck or RG3 will help my team more in the long run (I already have a Top-10 QB, Romo).

Great article and great discussion from the members/writers as well. That’s why I love this site!

Admin
Reply to  Mark Tavares
10 years ago

I’ll let Ghost argue his side since I’ve already argued w/ him on this article 🙂 But yes, I’d take a proven Charles for the 1.04. Not saying that 10+ years of potential QB1 wouldn’t be a better value long term considering the churn of the RB position, but position scarcity & perceived trade value play too big a role for me to ignore it.

Kenneth Leider
Reply to  TheFFGhost
10 years ago

This article has inspired me to do some number crunching myself. I considered doing a VBD calculation by position, and multiplying by an average longevity to get a relative value calculation. My concern is that position longevity (and QB in particular) is probably being under calculated because the dataset is too small. Rodger, Newton, Stafford could all have many more years of elite status that are not being included in your calculation because you don’t have their future data yet. If the average tenure is about 8-9 years you might need decades of data to reveal that.

I wonder if you could better measure churn with the inverse of this measurement, how many players do not repeat in the elite. Try to determine how often the opportunity to find an elite player in each position presents itself. Or maybe try to predict movement into the elite by looking at past rankings.

Brady Lawson
10 years ago

Interesting read. I think this analysis would be better if you looked at elite production instead of top 10. Would obviously be more difficult to do but I think it would be more telling. The QB position, for example, has a big drop off after the first 5 players and I don’t think it makes sense to include 6-10. I also think coming to the conclusion that TE and QB should be drafted early is a little flawed due to the fact that most leagues start 1 QB and TE and 2 or more WRs and RBs; skewing the value. Plus, there’s no point in taking a TE early not named Graham or Gronk. I would say taking a QB or TE over a RB if you value them roughly the same makes sense, though.

Basically I just think this research needs a little bit of modification. I wouldn’t argue with taking a top 5 QB or top 2 TE over a RB or WR early. But if it’s something like QB7 vs RB10, taking the RB makes a lot more sense.

Brady Lawson
Reply to  TheFFGhost
10 years ago

Very true about WR depth. I generally try to stay away from early round WRs unless I think they’re special talents. In a startup draft I did this summer, the top 5 QBs were off the board by 1.9. I was at 1.10 and ended up with Ryan Mathews which I considered a great value. Would you have suggested that I took Gronk or Graham there? I didn’t even really consider it at the time. I planned to get one of the 2 at 2.03 but they both went between my picks and I ended up with Fitzgerald (another great value in my mind). I felt like that was a strong first 2 rounds, but your research would suggest otherwise. Maybe I’m just misunderstanding the article, but it seems as though it’s missing something. Are you saying that you should take the #1 QB or TE over the #1 RB? (And continue along that way taking QB and TE 2/3/4 etc. over RB 2/3/4 etc?) Mostly I’m just trying to figure out when this TE and QB over WR and RB is supposed to come into play.

Brady Lawson
Reply to  TheFFGhost
10 years ago

Thanks for the explanation, really helped me understand a bit better. I thought you were suggesting that draft boards be completely changed to feature top QBs and TEs in round 1 instead of top RBs. I understand now that you are merely pointing out that longevity should be used to distinguish between 2 players valued similarly. I still would have a hard time drafting Gronk over Mathews, though, just based off of trade value. I think this could really help me in the trade department, though. Trading a top RB and average TE for a top TE and average RB, for example. I’ve definitely been RB oriented (as I’m sure many have) for the majority of my FF tenure.

The league I alluded to earlier was actually just a standard PPR league. Rodgers went 1, Brady went 2, Stafford went 3, Newton went 8, and Brees 9. It was a very unexpected turn of events. Was actually originally planning to go Newton or Stafford followed by Gronk or Graham. Expected Rodgers, Calvin, Mccoy, Foster, Rice, Mathews, Fitz, Richardson, and either Newton or Stafford to be off the board. But I guess that’s what makes drafting fun!

Chris Russell
10 years ago

I think the key is that you have to be drafting in the same spot (late first, early second) for this to work out well where you actually draft the QB and TE (without reaching). Perhaps what should be said here is to draft the most valuable player always…and then trade for the player you need. So take MJD but trade him for Cam Newton and RB X. Take Chris Johnson but trade him for Gronkowski and RB X or WR Z. That is of course if you subscribe to the theory your team will be stronger long term for these trades.

In fact, in the draft I referenced above I took Brady with my 3rd round pick and traded him for Harvin and Dalton. If I had passed on the best player during the draft I would have just taken Harvin for no upgrade later.

Always draft the most valuable commodity.

Sensei John Kreese
10 years ago

I wish they would make an episode of Quantum Leap, where Sam goes back in time and fixes some of my bad trades.

To Top