3RR impact

kjduke · Post by **kjduke** » Tue Nov 20, 2007 7:59 pm

I've run the data. Here it is.

I ran stats on the overall NFFC ADP, looked at total pts scored for each of those draft picks for round 3 and round 4 (which would be the bulk of the impact of the reversal) and calculated the impact (including positional adjustment) assuming those picks were reversed for those two rounds (i.e., no 3RR). Here are the results:

slot 1 = 74 pt gain
slot 2 = 65 pt loss
slot 3 = 121 pt gain
slot 4 = 134 pt gain
slot 5 = 141 pt gain
slot 6 = 171 pt gain
slot 7 = 112 pt gain

Slots 8-14 are the mirror image of slots 1-7, since the reversal assumes picks went to the opposite teams with no reversal, so ...

slot 8 = 112 pt loss
slot 9 = 171 pt loss
slot 10 = 141 pt loss
slot 11 = 134 pt loss
slot 12 = 121 pt loss
slot 13 = 65 pt gain
slot 14 = 74 pt loss

So, in reality for this season to date, 3RR actually helped 6 of the top 7 slots while hurting 6 of the 7 bottom slots, assuming the players taken in rds 3 and 4 would have been selected in the same order had 3RR not been in effect.

However, if we look at from a theoretical standpoint, assuming the drafting teams would have selected players in the correct order from best to worst during those rounds, the impact would have been as follows:

slot 1 = 74 pt loss
slot 2 = 52 pt loss
slot 3 = 20 pt gain
slot 4 = 35 pt gain
slot 5 = 24 pt gain
slot 6 = 18 pt gain
slot 7 = 6 pt gain

slot 8 = 6 pt loss
slot 9 = 18 pt loss
slot 10 = 24 pt loss
slot 11 = 35 pt loss
slot 12 = 20 pt los
slot 13 = 52 pt gain
slot 14 = 74 pt gain

So, once and for all we can see actual impact (on average only). Now, any given league can and probably would be very different depending on where top performing players from these rounds were selected (Brady and Moss) and where the worst performers were selected (Deuce, Bulger, T Bell, Caddy, Javon and Ahman).

The bottom line is, as I said in an earlier post, having an earlier pick in that 3rd round can be an advantage, but if the correct pick was not made (Brady or Moss in this instance) the effective opportunity of 3RR will not reflected in actual results.

addendum: the raw data also shows how the impact of injuries overwhelms the actual results of 3RR as no less than 14 of 28 picks from rounds 3 and 4 have been injured for at least a portion of the season, which no doubt accounts for MOST of the disrepency between slot values for these rounds.

[ November 21, 2007, 02:20 AM: Message edited by: KJ Duke ]

kjduke · Post by **kjduke** » Tue Nov 20, 2007 8:19 pm

Order (average) of players taken in rds 3 & 4:

slot 1 = Ahman Green, Driver
slot 2 = Deuce, M Lynch
slot 3 = Caddy, T Brady
slot 4 = Brees, R Moss
slot 5 = A Peterson, Barber
slot 6 - Gates, Burress
slot 7 = Colston, J Lewis
slot 8 = Portis, Bulger
slot 9 = J Walker, H Ward
slot 10 = L Evans, Branch
slot 11 = C Palmer, McNabb
slot 12 = T Jones, L Coles
slot 13 = A Boldin, C Johnson
slot 14 = A Johnson, T Bell

Nag' · Post by **Nag'** » Wed Nov 21, 2007 1:42 am

Good work, KJDuke. I don't have time to examine or discuss right now but I want to leave with this point. If 3RR's intention was to "level the playing" field between the draft slots, based on this season, it has not done so. Not even a bit. You can blame injuries or Brady/Moss or whatever but the bottom line is the bottom line. Also, I think it is safe to assume based on the standings that if the top slots had even better FF production from their top players, they would dominate similarly to 2005/2006. The one thing 3RR did do is create the perception of draft slot equality. But if the above is all true, how long before people realize that the perception of 3RR's equality isn't supported by actual results and the top slots are still the most coveteds? And when this occurs - and it could occur as soon as next season - what has been achieved?

renman · Post by **renman** » Wed Nov 21, 2007 2:47 am

To me, STATS wont tell us anything regarding 3RR. The teams who pick the players who explode will always do better and the teams that pick the players that are busts will fail. I think we can all now finally admit that we cannot predict with much certainty who those players will be or where on a draft board they will fall.

What 3RR does is provides a SENSE OF CONTROL over ones draft position and that provides piece of mind. It takes away the SENSE this is all about luck. These are good things that likely appeal to fantasy players.

King of Queens · Post by **King of Queens** » Wed Nov 21, 2007 2:47 am

Originally posted by Nag':
Good work, KJDuke. I don't have time to examine or discuss right now but I want to leave with this point. If 3RR's intention was to "level the playing" field between the draft slots, based on this season, it has not done so. Not even a bit. You can blame injuries or Brady/Moss or whatever but the bottom line is the bottom line. Also, I think it is safe to assume based on the standings that if the top slots had even better FF production from their top players, they would dominate similarly to 2005/2006. The one thing 3RR did do is create the perception of draft slot equality. But if the above is all true, how long before people realize that the perception of 3RR's equality isn't supported by actual results and the top slots are still the most coveteds? And when this occurs - and it could occur as soon as next season - what has been achieved? Coming in 2009: Team with 14th pick in 1st round gets the 1st pick in the 2nd, 3rd, 4th...and every round thereafter.

Greg Ambrosius · Post by **Greg Ambrosius** » Wed Nov 21, 2007 3:05 am

Originally posted by Nag':
Good work, KJDuke. I don't have time to examine or discuss right now but I want to leave with this point. If 3RR's intention was to "level the playing" field between the draft slots, based on this season, it has not done so. Not even a bit. You can blame injuries or Brady/Moss or whatever but the bottom line is the bottom line. Also, I think it is safe to assume based on the standings that if the top slots had even better FF production from their top players, they would dominate similarly to 2005/2006. The one thing 3RR did do is create the perception of draft slot equality. But if the above is all true, how long before people realize that the perception of 3RR's equality isn't supported by actual results and the top slots are still the most coveteds? And when this occurs - and it could occur as soon as next season - what has been achieved? At least with KDS as part of 3RR, Alex, folks will have a choice to "covet" those spots if they want or move to other areas of the draft. We all know that even with LT2 being the top choice this year, more than 50 owners this year didn't have No. 1 as their top KDS preference. That may seem unreal to you and me, but with 3RR not EVERYONE wants to be at the top. Some folks (I believe it was two) this year got the No. 1 pick because owners chose to move down with their KDS preferences.

So what has 3RR achieved? I don't know, but the combination of KDS/3RR has given owners more say in their draft slot position and it allows folks to plan their draft strategies with some hope in mind that they'll draft where they want to. As we know, some folks go WR-WR-WR and to do that you almost need to be at the bottom of the first round and KDS/3RR allows for that.

Anyway, I agree that the top teams would have been stronger if LT2 and SJax and Gore had performed better, but looking at the bottom of the first round you could say the same thing for teams 12-14 if Maroney, Travis Henry and Steve Smith had performed better. You think the top picks stunk? Look at the bottom of Round 1 and the start of Round 2 for stinkers!!

It's all good. Arguing about the merits of 3RR isn't worth it as I can tell you it's here to stay in the NFFC and it should be a fun off-season of trying to figure out where the best spot to draft from in 2008 will be. Let the facts state what they may about 3RR at season's end and the same for randomly selected leagues. Both methods are good. We've just decided to go with 14 teams and 3RR, like it or not. Us Cheeseheads are just weird that way.

Quahogs · Post by **Quahogs** » Wed Nov 21, 2007 3:12 am

It's possible people believe you're polluting the ground water with such radical FF ideology Greg. Mayhap they're afraid it may taint their 1-12 random wells in the future??

joetreff · Post by **joetreff** » Wed Nov 21, 2007 3:33 am

Greg if possible can you arrange things so that every team is tied at the end of the year?

3RR helps in perception only, that was obvious from the beginning. Everyone and I mean everyone had 4 chances to draft Braylon Edwards and lots had 5 chances. Almost everyone had 8 chances to draft Welker, 3 chances to draft Brady etc...
The good players still ended up with better and deeper teams.

ToddZ · Post by **ToddZ** » Wed Nov 21, 2007 3:51 am

Here's my take...

The following is fact. It can be demonstrated year after year in baseball, football and basketball. If you use END OF SEASON statistics, and populate teams rosters giving them the most valuable player at an open position available, and do this in a regular snake draft fashions, when you total the value on each team, the first few teams ALWAYS, ALWAYS, ALWAYS have more total value, usually in the order 1,2,3.

Another way to look at it is if you graph the value of each round, there will be a slope from 1 to 14, then 14 to 1, then 1 to 14.

Furthermore, the distribution of the value is such that the slope is STEEPEST in round 1, decreases in round 2, decreases more in round 3 etc. Eventually, it is essentially flat. That is, the first player taken in the round is just about equal in value to the last. There is a difference, but it is small, almost insignificant compared to the differences in the first couple of rounds.

The idea of a snake draft is to even things out. But as suggested, the VALUE DIFFERENCE between the first few players is SO HUGE that it cannot be made up by snaking each round, leading to the teams with the first few picks being the most valuable.

In football, the actual stat that generates this list is simply the number of fantasy points each player socred, adjusted in the value-based-drafting manner.

In baseball and football, since the scoring is roto-style, you use a "dollar value" as if the league we an auction. (NOTE - Since it is roto, "more value" doesn't necessarily equate to "more points").

If you do the populating the rosters exercise using 3RR, you indeed find the difference between the top team and the bottom team is diminished. Why? After round 1, the early picks have a decided value advantage. Round 2 reverses, so the teams with the turnaround picks have the value advantage, but the relative difference is greater in round 1, so teams 1,2 and 3 still have a total value advantage after the second round. In 3RR, the END TEAMS again go first, FURTHER cutting into that advantage (and also not adding onto the advantage if it were a regular snake). But even in 3RR, the EARLY ADVANTAGE still leaves teams 1,2 and 3 with the overall advantage, but the top to bottom difference is reduced.

Again, USING END OF YEAR STATS, the above is fact, for football, baseball and basketball.

The question now is HOW PREDICTABLE ARE THE FIRST FEW ROUNDS? If they are VERY PREDICTABLE, then 3RR is useful, if not necessary to even the playing field. If they are MOSTLY UNPREDICTABLE, then 3RR has little effect.

Another way to look at it is how close is the ADP to the end of year rankings per sport, at least for the first few rounds.

In basketball, it is VERY CLOSE, hence why 3RR is basically necessary to level the playing field.

In baseball, it is not quite random, but there is much variability. This along with the dynamics of fantasy baseball (larger rosters, separate hitting and pitching pools, longer season) render 3RR not so necessary.

That brings us to football. How predictable are the first three or so rounds? PRIOR TO 2007, you can argue that it is predictable enough to make 3RR more than perceptually useful, it has a tangible effect of leveling the field. Throw 2007 out the window. If someone were doing a 10 year study on football ADP versus end of year rankings, they would throw the 2007 data away as an outlier.

So you can't look at it year by year, that is too small and myopic a sample. To REALLY determine if 3RR has an effect, you need several years worth of comparing end of season values to the ADP of that year.

Tom Kessenich · Post by **Tom Kessenich** » Wed Nov 21, 2007 3:55 am

Originally posted by Todd Zola:

So you can't look at it year by year, that is too small and myopic a sample. To REALLY determine if 3RR has an effect, you need several years worth of comparing end of season values to the ADP of that year. Well said and this is reinforces the point that Greg and I have made that 3RR wasn't just about what LT and LJ did last season. This was about our years of experience in fantasy football (as well as all of yours) and seeing the realities that exist in this hobby and implementing an approach that we believe levels the playing field.