Théoden's Coding Tips

The League is a sitcom about a group of 30-something adults who are in a fantasy football league. After seeing the promo (and being a fantasy football addict), I raced to the DVR to acquire some League action. I’ve since slogged my way through six episodes and am now at the point where I just cannot bear to punish myself anymore.

I thought I was going to get a comedy that satirizes both fantasy football and the people who are involved with this national obsession. Instead, The League is a poor attempt at TV-for-guys-who-like-TV, composed of painful caricatures, pointless vulgarity, the occasional NFL player cameo, “zany” scenarios that have nothing to do with fantasy football, and a few flat jokes about the game itself.

The premise is solid, the acting is fine, but the plot lines and writing are miserable. Every now and then the show gets a chuckle out of me, but that’s just kind of like how a stopped clock is right twice a day. At the moment when the characters were in a car, screaming and yelling as a monkey was jumping frenetically around the interior, I hit the stop button, deleted the episode, and decided I was done with The League.

If you really want to bring out the comedy in the fantasy football craze, you can’t just toss together a crappy sitcom, make a few fantasy football references, and expect to bask in rays of accolades. If I had the resources and the opportunities, here is how I would do it:

Ditch the sitcom paradigm and move to something closer to sketch comedy – a Daily Show or SNL model, if you will. The drama of the NFL and weekly fantasy impact hold so much material that you could put out a show every, say, Wednesday. If actual NFL footage could be incorporated into the show, then you could have something really huge. The point is that the comedy would be focused on fantasy football and it would be current – the show would be extremely relevant to the millions who are playing this game.

A number of blogs exist that combine humor, analysis, and fantasy football fallout, but for a medium like television, The League should be disbanded.

Heyyyy-oooo!

Last year I wrote Fantasy football non-lessons learned in 2008 so maybe this is becoming an annual tradition. My sorry draft this season has hopefully taught me a few lessons that I can learn from going forward:

Avoid one-hit wonders

RB Matt Forte exploded out of nowhere in 2008 and has comparatively disappeared in 2009. I was a little scared of a “sophomore slump”, but I bought into the hype that Forte could be even better with Jay Cutler at the helm. In the end, I chased after a one-hit wonder instead of using my 7th overall pick for a reliable stud (e.g., QB Drew Brees).

Don’t bank on wide receivers

Starting with the second round, I went Andre Johnson, Greg Jennings, and Jason Witten. Johnson has produced, but Jennings and Witten have both fallen well short of expectations. The WR position in general lacks consistency, so basing your team around a core of receivers isn’t necessarily the best idea. I think it might be better to go after the late-rounders or free agents who invariably wind up rising to the top (e.g., Miles Austin, Mike Sims-Walker, Sidney Rice, the Giants’ Steve Smith).

Two upside quarterbacks do not equal one stud

Last year I had an injury prone Matt Schaub and an inconsistent Jake Delhomme. This year, my QB corps of Carson Palmer and Matt Cassel (later traded for Matt Hasselbeck) is not much of an improvement. Since his bye week, Palmer’s FP production has been a pathetic {15, 7, 16, 11, 9}. Those performances from Palmer are a big reason why I’m out of playoff contention.

Don’t break strong RB monopolies

I had the RBs for the defending Superbowl champs in Willie Parker, Mewelde Moore, and Rashard Mendenhall. The second it looked like I had a three-headed RB monster on my hands, I started dropping these guys. Currently Mendenhall is playing for another team and has more fantasy points than Matt Forte or any of my other crummy backs. I need to think very carefully in the future before I give up an RB monopoly on a good team. Though with the way the Steelers have been playing lately, I’m not sure that we can call them “good”…

Next season I am almost certainly going to take a QB in the first round. After two straight seasons of flaky QBs and 1st round busts, I want some reliable production. Aaron Rodgers, say hello to your new team for 2010: the Shambling Corpses!

In my last post, I talked about grading the rankings of fantasy football experts using the root mean square error (RMSE), and listed some results for Week 9. This time, I took the data from ten analysts and examined their performance for weeks 1-9. The numbers seem to imply a certain amount of group think:

The results are pretty discouraging… all the analysts are basically the same! The biggest difference is 1.1 between the WR performance of ESPN’s Matthew Berry and Yahoo’s Scott Pianowski. So if Pianowski tells you that Randy Moss is the #1 WR for the week, then Moss is pretty much guaranteed to be in the top 21; if Berry tells you that Andre Johnson is the #1 WR for the week, then Johnson will fall somewhere in the top 22. Is that a big enough difference to say that Pianowski is a better analyst than Berry?

Amusingly, if you sum up the errors for each analyst, you’ll find that Matthew Berry, face of ESPN’s fantasy football team, is the worst. Still, they are all so close that being best or worst doesn’t have much meaning.

All this homogeneity had me wondering if there was a bug somewhere in my code. As a sanity check, I decided to see how different the rankings of the analysts were from one another. To do this, I took the average of the standard deviation of the rankings for each player. The result:

So you end up with an average standard deviation of about 4 for each player ranking. That is a pretty small number when you consider that each of these analysts is ranking 40 running backs and 50 wide receivers. If you include only rankings in the top 10, then the number plummets to 1.42704758.

Even more interesting is the very clear illustration of group think. The ESPN guys have an average deviation of about 3 amongst themselves, as do the Yahoo guys. But combining the two groups, the number increases by about 15%. It’s all very fascinating; the conclusion I draw is that these analysts are not significantly different from one another.

Notes about methodology

When calculating the RMSE table, I leveled the playing field by using the top N rankings from each position; see the table below:

The other thing to note about the scoring is that I capped the maximum error for situations where a player’s actual output put them outside the top N rankings. If they were outside of N, then I squashed their ranking to N+1. This was to avoid the situation where a player posts a goose egg and therefore winds up amongst the hundreds of other players with zeros.

Was it worth it?

I had fun doing this as a hobby project, but I was disappointed not to uncover a “super analyst”. All these guys are doing better than if you were to draw names out of a hat, but I wonder how much better they are than the average Joe Football. I am half-tempted to try ranking players myself and see if an amateur can compete with the experts.

UPDATE: added the ESPN guys.

Last year I considered writing some code that would grade the performance of fantasy football analysts. This year I’ve actually finished up the code and can post a few results. What I am doing is very simple:

1. Download an analyst’s weekly rankings.
2. Download the actual player rankings (based on fantasy point production).
3. Compare the two for each position using the root mean square error (RMSE).

The following table shows the RMSE of CBS Sports’ Jamey Eisenberg and Dave Richard for Week 9. Similar to the Netflix Prize, the lower the number the better. In this case, the RMSE tells you how many rankings off the mark the analyst tends to be.

EXAMPLE: Using the table below, we see that Dave Richard’s RMSE for quarterbacks is about 8. So if he says that a QB is going to be ranked #10, then you can be pretty confident that the QB’s actual rank will be between #2 and #18. That may seem like a large error, but keep in mind the random, uncertain nature of pro football: Tampa Bay beat Green Bay last week; Cedric Benson is the 7th top RB; Smith Smith of the Giants is the 7th top WR; the Titans are 2-6. Predicting this stuff is nearly impossible, but I have a hunch that some analysts will be better than others.

• ^C = CBS Sports
• ^E = ESPN

I will look to post data covering more analysts later this week!

Here’s the scenario: player A has two QBs on a bye in Week 8, and player B has two QBs on a bye in Week 9. Both players want to keep their QBs, but they also need to cover their respective bye weeks. To solve the problem, they work a temporary trade where they swap one of their quarterbacks before Week 8, then swap back after Week 9. Is this collusion? Is this a legal trade?

Intuition says that something fishy is going on, but producing a rational argument is difficult. The arguments I found on the web were frustrating, boiling down to nothing-statements like, “It’s illegal because it’s unfair.” The problem is that the classical example of collusion involves a lopsided trade, where Team A becomes weaker in order to make Team B stronger. But in the case where the teams are borrowing from each other at equal value, the balance of the trade is even.

So I was left trying find a logical reason behind the following two assertions:

• ALLOWED: a permanent trade that is mutually beneficial to both owners.
• NOT ALLOWED: a temporary trade that is mutually beneficial to both owners.

For me, the key was to view the trade graphically. Consider two teams, each having a roster of 9 players. Here are what the teams look like before a trade:

When dealing with a permanent trade, any number of players move from one team to the other, but the teams are still separate entities:

But when we deal with a temporary trade, then during the time period in question, the two teams are overlapping their rosters:

By performing a temporary trade, these two teams have effectively done the following:

• circumvented the roster limit by storing a player on each other’s bench.
• gained access to quality players at no cost.
• formed an alliance that, while acceptable in a game like Risk, really is not in the spirit of fantasy football.

I think those reasons are sufficient to nullify any temporary/mutual borrowing trades.

If you’re curious, I originally began exploring this idea when commenting on this blog entry.

Just like last year, I am finding that what I thought was an awesome draft has turned out to be less than mediocre. The season is halfway over in my 16-team league, and I remain unimpressed with my top 10 draft picks:

1. Matt Forte (7th overall) – I thought I was getting a great deal by having Forte drop out of the top 5 and into my hands. Drew Brees was on the board, but my thought was that an elite RB was too good to pass up. Of course, what has happened is that Forte has been a bust – you don’t spend a 1st round pick on the #30 RB in fantasy point (FP) production.
2. Andre Johnson – an exception to my poor draft, the first WR off the board has paid off. Johnson is the #1 WR in my league, meaning that finally one of my players has lived up to their draft expectations. If only this happened more frequently for me…
3. Greg Jennings – deciding to secure two elite wideouts, I took the 5th WR in the draft with Jennings. Yes 5th off the board, but currently 32nd in FP production. Roddy White, Marques Colston, and Chad Ochocinco are all looking really good in hindsight.
4. Jason Witten – my league has no TE position, but does have a WR/TE slot. My thought was that Witten should produce like a high-end receiver, but being the #59 WR/TE isn’t what I had in mind with a fourth-round pick. Witten consistently gets a number of short targets by Tony Romo, but the TE’s last three FP totals of {3, 4, 5} just aren’t cutting it.
5. Willie Parker – ugh, see my 10th pick.
6. Carson Palmer – there was a big run on QBs early in the draft, but I decided to be patient and go for a middle-tier QB later on. Palmer is currently the #7 QB in the league and just came off a 5 TD game. He’s been up and down throughout the season but I’ll feel pretty good if I end up with a top 10 QB in the sixth round.
7. Marshawn Lynch – Lynch fell really far in the draft due to the suspension. When Fred Jackson performed well in his absence, I began to worry that Lynch was not going to pay off in the way I had hoped. His post-suspension output of {4, 12, 7, 10} isn’t spectacular, but certainly nice for a 7th rounder.
8. Kevin Walter – I just dropped Walter this week. Coming back from the injury with a 16-FP game, I thought I might have something good on my hands. However, I can’t reserve a bench spot for a WR whose last four games have been {4, 3, 2, 2}.
9. Matt Cassel – meh, a low-end draft pick for a low-end QB. Can’t really complain about this one, especially since I’ve just traded him away for Matt Hasselbeck who hopefully turns things around and stays healthy.
10. Rashard Mendenhall – I had the foresight to draft Parker’s handcuff, but not the patience to retain him on my bench. In the early goings of the season, it looked like Parker would be the guy for the Steelers. Now Mendenhall is the #16 RB and playing for one of my opponents.

Despite the lackluster draft and the Mendenhall blunder, my team is still 4-3. I’ve made good moves with free agent DSTs, cashing in on a 25-FP performance from Denver, and a massive 43-FP from the 49ers. Surprisingly, kicker Lawrence Tynes has been a solid free agent acquisition, averaging over 10 FP/game. I also managed to have a good run with Glen Coffee, using the RB to help me through a couple of bye weeks.

The second half of the fantasy football season is going to be crucial for my team, the Shambling Corpses. If Forte, Jennings, and Witten can get going, then I should be playoff bound. If not, then I’m just going to have to keep hitting that DST lottery.

I just lost Week 15 which means I’m out of the Championship game and will be playing for third place. I sort of choked because I made a bunch of lineup tweaks and started these guys:

• QB Shaun Hill [10 FP]
• WR Davone Bess [2 FP]
• K John Carney [6 FP]

When I should have stuck with my usual crew of:

• QB Matt Cassel [31 FP]
• WR Bernard Berrian [10 FP]
• K Rob Bironas [14 FP]

Those moves cost me a shot to win it all. The problem is that I had very logical reasons for making these decisions. For example:

Matt Cassel had put up -2 FP against the Steelers, barely had a decent 19-FP game with a last minute TD against the miserable Seahawks, missed practice time grieving over the loss of his father, and was about to face a Raiders team ranked in the top ten against the pass. Meanwhile, Shaun Hill had two TDs in 4 of his 5 previous games and was facing a Miami D ranking in the bottom half of the NFL against the pass. What happens? Cassel throws a career high 4 TDs while Hill is shut out of the endzone.

The problem is that you can look over the stats, make decisions that seem logical, and then have everything blow up in your face. You want to play the matchups, but most of the time they don’t mean shit. You want to listen to the experts, but they’re right barely 50% of the time. You want to apply some sort of reasoning to the whole process, but there is so much randomness that it doesn’t seem to matter.

So what is the lesson here? There is no lesson. Wanna see the top draft picks of the newbie owner who is going to the Championship game? Here are his first five rounds:

1. Eli Manning
2. Willis McGahee
3. Marvin Harrison
4. Vince Young
5. Fred Taylor

If you can go to the playoffs after a draft like that, what the hell is the point of playing at all?

I guess I shouldn’t complain so much. After all, I made the playoffs and had the potential to play for the whole ball of wax. The season was so much more enjoyable with the tools and gameflow that CBSSports provides as opposed to the garbage that Yahoo puts out there. I mean even consider a “premium” league for 2009 to get a little more customization.

The Shambling Corpses will rise again!

I went into the 2008 fantasy football season thinking that I’d scored a hole-in-one with my draft:

1. LaDainian Tomlinson
2. Terrell Owens
3. Earnest Graham
4. Wes Welker
5. Jason Witten
6. Willie Parker
7. Jake Delhomme
8. Patrick Crayton
9. Colts DST
10. Matt Schaub

Eleven weeks into the season, these guys have been bad enough to land me 11 out of 12 in my league in overall fantasy point production. Let the bitchfest begin:

1. LaDainian Tomlinson – leave it to me to break the best player in fantasy football. LT hasn’t even been a top ten back this year, and his lack of consistency has been a killer. After a 10-FP game in Week One and a 3-FP game in Week Two, I sat LT… he blew up for 20 FP on my bench in Week Three. I gritted my teeth and bore it when he put up 5 FP in Week Five, 11 FP in Week Six, and 6 FP in Week Seven. At that point I needed a win and just couldn’t run him out there for a game in London halfway around the world. He exploded for 22 FP on my bench. I picked up 10 FP from him against a terrible Kansas City rush defense, then benched him for a 15 FP performance against the Steelers. It’s bad when a player underperforms, worse when that player’s performance is so flaky that you have to make week-to-week decisions on him. That’s not why I drafted LT second overall.
2. Terrell Owens – unlike LT, I’ve put TO in my lineup for each of the 10 games he’s played. Early on I had a chance to trade him for Philip Rivers, but no more. Here is TO’s FP production over his last five games: {3, 3, 3, 7, 3}. That’s terrible. That sort of production from a 2nd Round WR is a team-killer.
3. Earnest Graham – I dropped Graham last week for Cadillac, and now it looks like Graham might be done for the year. Graham gave me #2 RB numbers, but I had been hoping for a lot more. At 737 all-purpose yards and 4 TDs this season, he’s fallen well short of the 1,222 all-purpose yards and 10 TDs from last season – even though he had the same number of starts in both seasons.
4. Wes Welker – another Tom Brady casualty. The yardage has been there for Welker, but his lone TD will require some padding if he wants to hit the 8 touchdowns he had in 2007.
5. Jason Witten – like TO, Witten produced as expected early on. Then there was the Romo injury along with Witten’s rib injury; his last five games: {5, 4, 0, 0, 3}.
6. Willie Parker – a great deal in the sixth round… or he would have been if he hadn’t spent five games benched with injuries.
7. Jake Delhomme – ugh. The first of my “upside” quarterbacks. You have good Jake with games like these: {15, 23, 19, 19, 19}; and bad Jake with games like these: {3, 3, 3, 0, 9}. I played the matchups well and didn’t get burned by Delhomme until these last two weeks. Against a 2-8 Oakland team and a 0-10 Detroit, Delhomme totaled 170 yards, 2 TDs, and 4 picks. If 9 FP is the best I can get out of Delhomme against the winless Lions at home, there is no way I can trust him down the stretch.
8. Patrick Crayton – bust.
9. Colts DST – bust.
10. Matt Schaub – my second “upside” quarterback. When Schaub went through a mid-season stretch of {24, 20, 29}, I thought I had a QB whom I could ride for the rest of the season. But after a 1-FP Week Nine performance when he injured his MCL, Schaub is pretty much done for the rest of fantasy football in 2008.

So what looked like a phenomenal draft has turned into a bunch of scrubs. It’s not just that one or two players have underperformed… they all did.

Fortunately I did a decent job taking advantage of my high waiver priority to make a few key pickups:

• RB Steve Slaton – only one fantasy point less than LT’s total, Slaton has provided consistency even while Ahman Green eats into his carries.
• WR Bernard Berrian – has more total fantasy points than Randy Moss, Terrell Owens, Lee Evans, and Braylon Edwards.
• Baltimore DST – the #8 DST in total scoring, Baltimore’s 143 FP is a huge gain over the Colts’ miserable 85 FP.

Those free agents have kept me from sinking beneath the waves, putting me at 6-5 with an outside shot at making the wildcard. Every week I think, “this is the week we’re going to let loose with both barrels”, then Delhomme throws four picks, or TO gets 33 yards and no touchdowns, or LT rushes for only 41 yards.

Still… surely, this is the week when it all comes together.

Another fundamental complaint that I have with fantasy football, aside from the WR/RB flex position, is the way your win-loss record is determined. In “standard leagues” you are pitted against another team each week – score more points than your opponent and you win; score less points and you lose.

The flaw with this system is that you have almost no impact on what your opponent scores. You aren’t really playing against the other team – there’s no interaction whatsoever. In the long run, this system might work out. If you put up good numbers, then over time your record will reflect that. But there is no long run. The fantasy football regular season is 14 games – not enough time for the numbers to average out.

I’m in a three-division, 12-team league where the average team score is 83.9 fantasy points (FP). Right now we’re ten games into the season. If I tell you that a given team is scoring 83.7 FP/game, then you might guess that they have about a 5-5 record. Well no, they are 7-3 and leading their division, ahead of two teams that are scoring 89.4 and 87.7 FP/game. Yeah, 7-3 isn’t so far from 5-5; you’d even expect it given such a small sample size, but as I’ll get to, that’s kind of the point.

Oh, and the aforementioned 87.7 FP/game team is 2-8 and has faced a ridiculous 103.0 FP/game. CBS Sports (better than Yahoo) has a great stat called “Breakdown” where you can see how a team would have fared if they had played all opponents. This team that is 2-8 (20% win pct) would be 63-47 (57% win pct) if they played all teams each week.

That’s a perfect illustration of the absurdity of the system. A random scheduling fluke made the difference between this guy being a playoff contender, and being all but mathematically eliminated from the post-season. Meanwhile, the team with the league’s best record (8-1-1) has faced a pathetic 74.3 FP/game, though their weekly average of 97.5 FP/game means that they don’t need an advantage like that. The fact is, these two teams have played nearly the exact same opponents, but there is a difference of about 30 FP per game in the scores they have to beat. That’s wrong.

Performance in fantasy football should be directly related to how well you draft, trade, acquire free agents, and set your roster. Plenty of randomness is already introduced through the game of football without adding in this luck ‘o the schedule. I would say that I’m observing a fluke in this 2008 season, but I’ve been on the receiving end of this crap in 2006 and 2007 so I know it’s nothing new.

Lately I’ve been hearing about some interesting hybrid systems. They keep the head-to-head aspect, but add in an overall component. For example, teams who place in the top half of scoring for the week might get an extra win. This is the sort of system I might look for in 2009.

A quick list of fantasy football players that had me running the other way in the 2008 draft…

QB Matt Hasselbeck: CBS ranked Hasselbeck as the #12 QB going into 2008, summarizing with, “He’s a low-end No. 1 Fantasy QB worth a pick between Rounds 7 and 8.” I’m not sure why I felt like I couldn’t trust Hasselbeck at the QB position, but my hunch turned out correct. A few sorry performances followed by injury have landed Hasselbeck’s 2008 season in the gutter.

QB Marc Bulger / RB Steven Jackson: These guys sucked last year and they suck this year. After a 3-13 season in 2007, I just wasn’t buying the Rams hype. Now it’s 2008, the Rams are 2-7, and I’m still not going to touch these guys when 2009 rolls around.

RB Brian Westbrook: This guy can put up monster stats and makes up a huge part of the Eagles offense. Still, I always shy away from Westbrook because I don’t want to deal with a player who is consistently Questionable. Looking at the high points, he’s had fantasy scores like these: 21, 27, 14, and 32. But then you look at the other five games he’s played/missed: 1, 0, 0, 9, and 5. Maybe he’s worth it, but I just can’t stand that boom-or-bust performance at the running back position.

RB Larry Johnson: Coming off a 2007 season where he held out and then got injured, and into a 2008 rebuilding year for the Chiefs, I really didn’t want to end up with LJ on my roster. With 10 NFL weeks in the books, LJ’s 417 yards, 3 touchdowns, and various suspensions are going to land him far from the #10 RB spot where CBS had him slotted.