In this week's Varsity Numbers, Bill Connelly takes a page out of baseball's playbook and attempts to isolate power from efficiency.
09 Jan 2013
by Tom Gower and Mike Kurtz
Ken Leibensperger: My friends and I have set up a Dynasty Fantasy League in which there is a large payout for the team that is deemed the best team after five seasons of fantasy football. We have just finished season three and are realizing we don’t have great parameters regarding how to prove which team had the strongest five seasons. I am looking for advice on a formula that would be able to take the total regular-season points per team, total regular-season record over five years and somehow weigh that with how many championships a team won over the five years. Any help or advice on a way to do that would be great.
Mike: Why doesn't this guy's league just use Pythagorean wins and then add extra points for deep trips in playoffs?
Tom: I think the concept of Pythagorean wins makes sense in actual sports because the team you're trying to measure has some control over how many points they allow. The same is not true in fantasy football. My favorite simple way of doing things would be to use a team's weekly record against every other team in the league. Yahoo! (at least) runs these numbers for each team at the end of the year. So I can see that my team this year would be 38-87-1.
Mike: I ignore basically everything Yahoo! tells me aside from points, wins/losses, and what my current roster is, so I'm not sure to what you are referring.
Tom: It's now in the Recap feature.
Mike: Oh. My least favorite feature.
Tom: I don't find it particularly useful either, but that "your record playing every team every week" information is now in there. So if I finished fifth in points that week, my weekly record that week in our 10-team league would be 5-4.
Mike: I see.
Tom: This gets rid of the vagaries of match-ups.
Mike: I'm not really sure what that gets you, though. You're still just comparing points for to points against, merely against a larger pool, which gets you back to Pythagorean wins, basically.
Tom: Sure, I'm trying to minimize the effect of randomness. And if you want to reward winning championships (I assume determined the more traditional way), you can just give a team bonus points. Say, "number of teams in league" points for each playoff win or something. If you want to use Pythagorean wins, I'd suggest not using an individual team's points against but an average of points by other teams in the league.
Mike: Well, at that point you may as well just use points for, since the average for the rest of the league is going to be nearly identical for each team. That isn't a bad idea, though, since points for is really what fantasy is all about.
Mike: I guess the question, then, is how to rate playoff performance and reward it, as Ken seems desirous of.
Tom: One of the things you have to guard against, though, is making the fifth year uncompetitive. There are a couple ways of doing this.
Mike: I suppose it's just a matter of adding wins to the pythagorean total.
Tom: One of them is just to have single-year prizes instead of just a single big prize for all five years.
Mike: That isn't what they're going for, though.
Tom: Many fantasy leagues deal with the problem of people, especially those not in contention late in the season, not bothering to do things like set their lineups. In a five-year competition, the whole last year is essentially that "late in the season" period. I want to find a way to minimize those people, as well as dissuade against intentionally tanking. Weekly records would solve that, since two competitive teams would both finish ahead of the out-of-contention teams.
Mike: Since we're fiddling with Pythagorean wins, how about this: add points against. Some kind of "league average" points against, rather than counting it against the team.
Tom: Or just "total points over average" in any given year?
Mike: Or something like that, yes
Tom: How would you value postseason wins, though?
Mike: I don't think you can value them as high as regular season wins, except maybe championships. It depends on how long your playoffs are. I could see .5 per playoff win and 1 for championship win.
Tom: I mean, if you went to points above average.
Mike: You could prorate the points above average score (which doesn't include playoff games), although that would return different numbers for each team. Honestly? Just pick a number of points for each game that sounds about right, possibly based on data from the first few years. You could get more complicated, but then everything becomes a mess and it'll still be just as arbitrary.
Tom: If I had my druthers, I'd suggest not using postseason performance at all, given the randomness involved there. Just regular season weekly records should give you a good idea of which teams have been consistently very good for five years. I probably am an anomaly as far as my low regard for playoffs as a means of determining the best team, though.
Mike: You are a lousy consultant. It's like someone came to you to make plans for a bridge and you come back with plans for a ferry.
Tom: I said "If I had my druthers" for a reason. He mentioned he wanted us to include postseason performance. My preferred method has a way to do that. (Though I should note, you could get bonus points for making the playoffs, and bonus points for each postseason win. That's probably better, if you want to reward making the playoffs.)
Mike: Anyway, I'm assuming from this email that he wants to reward both playoff success and championships. I think championships should be rewarded more generously than simple playoffs wins. So that's something to keep in mind when you come up with your arbitrary point values.
Tom: Well, in that case you can give bonus points for making the playoffs, bonus points for non-championship playoff wins, and double or triple points for winning the championship.
Mike: I'm not sure why you would reward making the playoffs, as opposed to playoff wins.
Tom: Well, they don't have to.
Mike: Especially since inferior teams points-wise may end up in the playoffs over superior teams based on the vagaries of head-to-head matchups during the regular season
Tom: Sure. I'm fine with just bonus points for playoff wins and maybe double or triple or quadruple bonus points for winning the championship. One of my recommendations is that if you do measure performance relative to an average, it should be the average of the other teams in the league. Fantasy football as generally constructed is rivalrous -- if I have Adrian Peterson, you can't have him too. I think that method does a better job of showing how well you did constructing your team as compared to how well the other people did constructing their teams.
Mike: All right. So where have we ended up, as far as a measure of success for poor Ken?
Tom: Well, I think ultimately you have two options for what you want to emphasize: points or wins. If you want to emphasize points, you can find points relative to a league average, and use cumulative totals relative to average. If you want to emphasize wins, you should find a way to normalize wins, to remove the vagaries of yearly schedule strength. I like weekly wins. Pythagorean wins relative to a league average (adjusted or not) is another good way to do this. You could add playoff wins, with extra weight for championship wins, to any sort of regular season win metric. And most wins, well, wins. If you want to measure things by points, you can value postseason and championship wins at some level of points. To an outside observer, that amount of points is not obvious.
Anyway, Ken, thanks for the question, and I hope this helps.
It was a week of relatively modest scoring for staff member leagues. Mike is currently in the lead, but seems unlikely to stay there for long after losing five players, including all of his wide receivers. Better positioned is second-place Danny, who benefited from multiple Seahawks. Tom had the top performer of the week in Arian Foster and is in third. With only Mike and Rivers (down both running backs and his kicker) suffering significant losses, the race is still wide open.
|FO Playoff Wild Card Round Results|
|QB||Aaron Rodgers||Tom Brady||Peyton Manning||Russell Wilson||Colin Kaepernick||Matt Ryan|
|RB||Ray Rice||BenJarvus Green-Ellis||Arian Foster||Marshawn Lynch||Adrian Peterson||Frank Gore|
|RB||Alfred Morris||Vick Ballard||Michael Turner||DuJuan Harris||Knowshon Moreno||Stevan Ridley|
|WR||A.J. Green||Eric Decker||Andre Johnson||Demaryius Thomas||Michael Crabtree||Roddy White|
|WR||Reggie Wayne||Sidney Rice||Randall Cobb||Wes Welker||Brandon Lloyd||Jordy Nelson|
|WR||Pierre Garcon||Golden Tate||Julio Jones||James Jones||Torrey Smith||Brandon Stokley|
|TE||Jermaine Gresham||Aaron Hernandez||Jacob Tamme||Dennis Pitta||Rob Gronkowski||Tony Gonzalez|
|K||Justin Tucker||Josh Brown||Mason Crosby||Stephen Gostkowski||Steven Hauschka||Matt Prater|
Joe Flacco and Anquan Boldin were your Best of the Rest stars in the wild card round, combining for 42 points. Sorry, Christian Ponder selectors, you do not receive Joe Webb's 15 points. The top Best of the Rest running back was John Kuhn, who like Shonn Greene in 2009 was not actually selected by any Best of the Rest team. Bernard Pierce (10 points) and Michael Robinson (8) were the top backs actually selected by a commenter. Congratulations to Rhombus for picking Michael Jenkins, whose 15 points was the second-best score of the weekend by any wide receiver. Popular picks Owen Daniels, Shayne Graham, and Texans D/ST all served the people who picked them well. Zac is the current Best of the Rest leader with 77 points, followed by Sid with 65 and Levy with 60.
A reminder that you can find your squad from the FO Playoff Challenge right here. Pierre Gracon's score is currently incorrect but will be fixed next week.
Keep Chopping Wood: In some ways, what happened Saturday night in Green Bay was really not Joe Webb's fault. He is not the one who made him, a very raw passer, a backup quarterback. He was not the one who tried to run a game plan that seemed to be designed for Christian Ponder. Nevertheless, the Minnesota Vikings started 21 of the 22 players they did in their win over the Packers the week before and were completely uncompetitive. That, plus elements of his performance that included a seeming indifference to being sacked and a willingness to attempt to avoid sacks by throwing nearly vertical passes behind his own line of scrimmage, is enough that we find him a deserving winner even considering the things that were not really his fault.
Mike Martz Award: Mike Shanahan declared this week that a coach who ignored a doctor's recommendation to not play a player should be fired. Which is kind of nice, considering the pre-game note by team physician Dr. James Andrews that he simply was not allowed to examine Robert Griffin before he re-entered the regular season contest against Baltimore. It's not just that Shanahan tried to break Griffin, or even that he continued to play Griffin after he was ineffective throwing the ball. It was that of the Redskins 24 plays from when they held a 14-3 lead to the time Griffin left the game, 15 of them were pass plays or Griffin runs. As of the most recent reports, Griffin is undergoing surgery for a re-torn ACL and newly-torn LCL. Shanahan has stated that he believes Griffin re-injured his knee the play before he collapsed, attempting to bend over to pick up a low snap. Which makes perfect sense considering all the limping Griffin was doing for the game's previous two quarters.
4 comments, Last at 13 Jan 2013, 12:26am by MurphyZero