Wednesday, November 27, 2013

Play fantasy football: In progress


 Happy Waiver Wednesday!

While a common Thanksgiving tradition is to have football on in the background while prepping for the massive feast, my family didn't partake in such traditions. In fact, excluding the occasional Super Bowl (for social purposes only) I had never watched a full football game until this season. A few months ago, on a whim, I agreed to participate in a fantasy football league. I figured I would learn a bit here and there (my first goal was to learn what wide receivers and running backs are), casually set my line-up once a week on Sunday mornings, catch a game every once in a while... real chillax stuff. It would take, what, an hour a week? But then... numbers.

Some background for the folks who are unfamiliar with fantasy football... 

Before the beginning of the football season, you get together with some friends and create a fantasy football league. A draft takes place, in which everyone takes turns picking football players for his/her fantasy football team. Each week, your team is pitted against the team of another friend in the league in a head-to-head match-up. You score points based on how well the individuals on your team performed in their real games. For example, you get a fraction of points for every yard that your player ran, and you get points for every touchdown he scores. You compare the total points of each team to determine the winner of that week's match-up. Eventually, you build up a record of wins and losses (as well as a gaggle of arch enemies) and establish a rank in your league. Towards the end of the regular NFL season, the highest-ranked teams face each other in the single elimination playoffs. The last team remaining is crowned the winner and gets to prance around in a field full of rainbows and unicorns.

So... back to numbers... 

Football is full of stats, and I found myself perusing weekly projections and "expert" rankings and weather reports and injury reports and depth charts and red zone targets and on and on and on... Naturally, I started looking for my own questions that could be answered within football stats.

Here is what I wanted to explore: How often do league-specific settings change the outcome of a head-to-head match-up, in comparison to standard settings?

More background for the folks who are unfamiliar with fantasy football... 

There are standard, default point settings for every fantasy football provider (e.g. Yahoo, ESPN, NFL, CBS), but each individual league can adjust its own point settings. For example, in standard scoring, a quarterback would be credited 4 points for every touchdown he throws. In our league, however, our league commissioner set it so that the quarterback would receive 6 points instead.

In a more extreme example, our league awards 5 bonus points for each of 40+ yard completions, 40+ yard passing touchdowns, 40+ yard rushing attempts, 40+ yard rushing touchdowns, 40+ yard receptions, and 40+ yard reception touchdowns. This means that if Aaron Rodgers throws to Jordy Nelson for a 50- yard touchdown, Jordy will get points for his yards (50*0.1),
plus 6 points for getting the touchdown,
plus 5 points for the 40+ yard reception,
plus 5 points for the 40+ yard receiving touchdown, or a total of 21 points.

Considering that nearly all of the top 20 wide receivers average between 10 and 20 points per week (and the next top 40 wide receivers average 8 to 10 points), the bonus points can cause huge fluctuations in scoring. How often would scoring adjustments affect the outcome of a match-up? Would these changes in head-to-head outcomes affect who gets a playoff berth?

Let's find out!

Coming from academia, this paragraph is the obligatory Methods section. I downloaded 12 weeks worth of league rosters and player statistics from Yahoo fantasy sports using the Yahoo query language (thanks for the tip, LN!). I parsed the XML data and did all subsequent data munging in Python. Finally, I imported the data into Tableau Public.

After some finagling (which, I believe, is the technical term for what I did), I ended up with this visualization of all our league's games thus far. Also, here are our league settings compared to standard settings. Differences are highlighted in blue.

Out of 60 games analyzed, there were 4 instances in which using standard scoring would have changed the outcome of a match-up. Team #8 (TR) seemed to be most affected by this, so in the visualization, I set the default view to her team. Note that league scoring is always in blue, and standard scoring is always in orange. In weeks 10 and 11, the dark blue bar (Team #8's score) is lower than the light blue bar (her opponent's score) whereas the dark orange bar is higher than the light orange bar. This means that she lost her league match-ups even though she would have won those match-ups had we used standard scoring.

Team #8 had 2 games (weeks 10 and 11) that were affected by our league-specific scoring system.
2 of her recorded losses would have been counted as wins had we been using standard scoring.

A breakdown of team rosters for week 11 in the lower half of the visualization window gives a better idea of what happened:

  • For Team #8 (TR): LeSean McCoy, Harry Douglas, and the formidable Cincinnati defense profited from our league settings. However, her quarterbacks, Christian Ponder and Eli Manning, were heavily docked for their sacks, fumbles, and interceptions.
  • For Team #5 (OF): While the quarterbacks, Big Ben and RG3, also struggled with sacks, fumbles, and interceptions, they combined for 7 40+ yard bonuses, which converted into 35 additional points.
On Team #8, Christian Ponder and Eli Manning were continually penalized for their multiple sacks, fumbles, and interceptions.
On Team #5, despite also having sacks, fumbles, and interceptions, Ben Roethlisberger and Robert Griffin III heavily benefited from the 40+ yard bonuses.

It seems as though sacks and the 40+ yard bonuses make the most difference in the outcome of the head-to-head match-ups. I thought that the defensive teams' "Tackle for Loss" statistics would make more of a difference because they greatly increase a defensive team's total points. However, it seems to affect all defenses in the same way, so it's pretty much a wash. On the other hand, certain quarterbacks are more prone to being sacked, while elite quarterbacks throw 40+ yard passes at least every other week.

Ultimately, does it make a difference? 

There have been 60 match-ups in our league so far, with 4 changes in outcome. Here is how it would affect our standings:


Our league happens to only send the top 4 teams to the playoffs at the end of the season, so as it stands, the same teams would be in the playoffs. However, given that we still have 3 regular season match-ups left (in our particular league), and that other leagues might send the top 8 teams to the playoffs, the scoring settings can indeed change the course of a team and its chances of an appearance in the playoffs. It might be worthwhile to deviate from standard rankings, and target players that are more likely to benefit from league-specific bonuses and devalue players that are more prone to league-specific penalties. Just some food for thought.

Speaking of food... I might just end up participating in the football-on-Thanksgiving tradition this year for the first time. Happy Thanksgiving everybody!
-T

P.S. I'm team #3 if you want to do some fantasy football roster stalking and evaluate my waiver wire pick-ups or make recommendations (unless it's to suggest that I drop Roddy White, which I finally did this morning after much deliberation).

P.P.S. Special thanks to JM and SS for watching games with me and fielding all my newbie football questions.

P.P.P.S. To Aaron Rodgers: Please be healthy soon, OK?

-------------------------------------------------------------------------------------

Update!

Our regular, 15-week-long season ended this past weekend so playoff berths are locked in! While Teams #9 (CL) and #3 (me!) had already clinched 2 of the 4 playoff spots, it came down to the last exciting week of match-ups to determine the last 2 teams for the 3rd and 4th spots. Ultimately, Team #6 (JS) held on to 3rd seed, and Team #4 (BK) secured 4th seed.
We go back to the original question of this post, which is whether league scoring settings altered the destiny of any teams, and it turns out that it makes an incredible difference:


Notice the 4 teams with the same 8-7 record hovering around the playoffs cut-off! A win or a loss for any of those teams would have changed the outcome of the standings. Both Teams #6 (JS) and #4 (BK) would have dropped a full 3 spots and out of the league championship running had we been using standard scoring.

What I believe to be the clutch match-up was week 13, when Team #4 (BK) edged out Team #2 (CB--the same CB from my mittens post!), as seen in the blue bars on the left. However, had our league used standard scoring, Team #4 would have lost. Once again, the discrepancies came down to quarterback woots and woes, as seen on the right side of the figure.
The result is that the dream was kept alive for Team #4 (BK), while Team #2 (CB) struggled and eventually had to settle for 6th place.
In any case, I'm heading into the league playoffs, where anything can happen. Keeping my fingers crossed for an epic Aaron Rodgers return!
-T