- Instead of individual players, owners pick teams.
- Instead of racking up points for touchdowns, yards gained, and all other manner of things which allow players to help their teams win games, points are awarded for miserable quarterback play. This includes interceptions, fumbles, low passing percentages, off-the-field transgressions, and many other things. In fact, having very good stats (300+ yards, 3+ touchdowns, etc.) will result in negative BQBL points.
You can check out the website for more details, but this article is for those who are already in a league and whose drafts include the option of choosing your position.
Allowing owners to choose their own draft positions is not in the official BQBL rules, but I definitely recommend it. The draft positions for the first year of our league were determined by a bar game tournament (two-sided bracket with alternating darts and shuffleboard rounds). It was a blast, but a couple owners intentionally threw matchups to get certain draft positions. This year, we did a poker tournament and allowed owners to choose their draft positions based on the outcome. Not only did it add another wrinkle in draft decisions, but it gave all the owners an incentive to try their best in the tournament, making it more fun.
I finished second-to-last in the poker tournament (only because one owner couldn't make it and was automatically dropped to last place), so I didn't have much of a choice. However, for the first few guys, there was some discussion surrounding which draft positions were the best. Everyone had his own theory, but the prevailing opinion was the first position was the worst since it also meant you were forced to take the last pick of the draft. Since each team must be started at least four times a season, many owners thought this could have a significant effect. After hearing all this, I thought I would try to find a stat-driven way to determine the best draft order using the statistics from the 2012 BQBL season.
The first and most obvious stat to use is the average team score over the season. But that can easily be skewed by a few really high scores, so a second metric is needed to counteract that. What I came up with was a multiplier of the number of weeks where a team achieved a minimum score. I needed an indicator of the level at which the team actually gave you a realistic chance of winning your matchup. So I took the minimum winning score of every matchup for the year (one point higher than the loser) and averaged them. That left me with 17.2, the theoretical weekly score which would net you a .500 season. Since each owner must start two teams per week, I halved that and was left with 8.6, which I rounded up to 9 since there are no decimal scores in BQBL. So the final formula is:
(Average weekly score) x (Number of weeks where score was equal to or higher than 9) = Adjusted score
I thought about adding additional factors, but they all seemed to skew the numbers in the same direction or in a way that didn't contribute to the overall goal. The list that resulted from the formula above passed the reality check: when I looked at it, the teams at the top were the ones that I was scared to face that entire year, the ones in the middle depended on the matchup, and the ones at the bottom were on everyone's bench most of the year. That list is included at the bottom of the article in case you want to look through it. One of the important things to remember is that the goal of this number is to present the independent value of that team, not to reflect how it will be used during the season. Individual starting decisions are dependent on way too many factors to include here, including what other teams are on an owner's roster.
Now that we have our total adjusted score, we can calculate the values of the specific draft positions under different scenarios for that year. The most obvious first step is a perfect draft: one where everyone has full knowledge of the adjusted score of each team. While anyone who has ever played fantasy sports knows that is impossible, it's one way to see if the stats from this season will indicate any underlying draft position pattern. Also, it's just as impossible to factor in things like owner knowledge, biases, and level of intoxication.
| Position 1 | 745.74 |
| Position 2 | 716.77 |
| Position 3 | 609.40 |
| Position 4 | 540.59 |
| Position 5 | 527.18 |
| Position 6 | 538.93 |
| Position 7 | 541.38 |
| Position 8 | 545.04 |
You can see that the first three positions hold an advantage over the rest of the league. Again, this assumes perfect knowledge of the outcome of the season before it even begins, but it reinforces the idea that dominates standard fantasy leagues: a single good pick in the first round can make or break your entire season.
Because the scores in the BQBL are tied to franchises instead of individual players and trading isn't part of the official game, one could argue that each pick in a BQBL draft carries far more riskt than traditional fantasy leagues: individual players taken in the first round have a relatively known amount of talent, and barring injury or another catastrophic change, their roles are fairly stable. Even in the event of a major change, most leagues allow you to trade or drop the player and get another. However, as proven by Arizona during this sample year, a team's fortunes can change very quickly: they were a BQBL bench team for the first ten weeks, but when their starting quarterback was injured in week 11, a rotating cast of inept quarterbacks filled in an starting scoring insane amounts of BQBL points. It works the other way as well: Cleveland was the first overall pick in our 2012 draft, and while their quarterback was very bad in some games, he didn't score nearly as consistently as we all thought he would and ended up 11th in average weekly score. Either way, unless you've instituted trading in your league, you're stuck with your drafted teams; no waiver wires or free agents here.
Wanting to simplify it even more, I put the teams into brackets where every 50 points of the adjusted score equals 1 point in the bracket (0-49=1, 50-99=2, etc.), we are left with an even flatter picture with a small advantage given to the top two positions:
| Position 1 | 17 |
| Position 2 | 16 |
| Position 3 | 14 |
| Position 4 | 13 |
| Position 5 | 13 |
| Position 6 | 13 |
| Position 7 | 13 |
| Position 8 | 14 |
The basic takeaway from this is pretty obvious and is true in most fantasy leagues: luck and your relative ability to predict performance of the teams have much more to do with your success than your draft position. There are sleepers, but they're as hard or harder to predict in this format than in traditional fantasy leagues. It's worth noting that the top three teams of 2012 were all taken in the first round of our draft and over 50% of the final points were scored by teams picked in that same round, so make your first pick count.
Just for grins, I put the teams with their adjusted and bracket scores into the actual 2012 draft order to see what came out.
| Position 1 | 317.40 | 9.00 |
| Position 2 | 575.72 | 14.00 |
| Position 3 | 435.85 | 11.00 |
| Position 4 | 1023.04 | 22.00 |
| Position 5 | 584.57 | 14.00 |
| Position 6 | 873.75 | 20.00 |
| Position 7 | 384.55 | 10.00 |
| Position 8 | 570.15 | 13.00 |
Somewhat surprisingly, the owner at draft position 6 ended up winning the league with Kansas City leading the way. The score of the owner in position 4 is inflated a bit because he had Arizona during their ridiculous stretch in the last 1/3 of the season. I suspect that the values for these positions will change pretty randomly depending on the year/draft.
For another view, I put the teams with their adjusted scores in our actual 2012 draft order and grouped them in fours: Group 1 was picks 1-4, Group 2 was picks 5-8, and so on with Group 8 being picks 29-32. I then totaled the adjusted score for each group and divided it by 4 to give the average adjusted score of a team within that group.
| Group 1 | 1,116.54 | 279.14 |
| Group 2 | 1,340.98 | 335.25 |
| Group 3 | 505.36 | 126.34 |
| Group 4 | 825.16 | 206.29 |
| Group 5 | 340.71 | 85.18 |
| Group 6 | 304.78 | 76.20 |
| Group 7 | 261.90 | 65.48 |
| Group 8 | 69.60 | 17.40 |
Totaling the "average prospective" scores for every draft position wouldn't be worthwhile since breaking them down like this ensures there are only two groups of scores: draft positions 1-4 and 5-8. Here are the results:
| Positions 1-4 | 588.01 |
| Positions 5-8 | 603.27 |
Not a big enough difference to be significant. It's interesting that the average score of Group 2 is actually larger than Group 1, so the first-pick advantage doesn't exist in this data set. If that held true in additional sets, it might indicate a later pick is better for real life drafts because highly-touted prospects which go in the first few picks are overrated (NFL draft, anyone?). That said, the difference between Group 3 and Group 4 more than makes up for the previous gap, indicating there were high-value picks in Group 4.
Here's the same analysis for the scores of a perfect draft from last season:
| Group 1 | 1,662.90 | 415.73 |
| Group 2 | 959.28 | 239.82 |
| Group 3 | 757.75 | 189.44 |
| Group 4 | 548.66 | 137.17 |
| Group 5 | 392.25 | 98.06 |
| Group 6 | 296.79 | 74.20 |
| Group 7 | 138.71 | 34.68 |
| Group 8 | 8.69 | 2.17 |
| Positions 1-4 | 737.90 |
| Positions 5-8 | 453.36 |
Again, we see that in a perfect draft, you're better off getting one of the first few picks since they give you a significant aggregate point advantage. The difference between Group 1 and Group 2 (175.91) more than offsets the difference between Group 7 and Group 8 (32.51), so for this particular data set, the argument that being forced to take one of the last four picks overrides any benefit gained by having one of the first four picks simply doesn't hold weight.
I wish I would have found some amazing revelation that gave a more surprising answer, but dem's the numbers. If you think your fellow BQBLers are savvy and will make correct draft choices, you're best off getting one of the first two picks and grabbing a team who you think will be the standout of the season. If you don't have much confidence in everyone's predictions, maybe take a slightly later position so you can grab two of the top 9-12 picks. Regardless of your draft position, you're probably only going to start your fourth round pick four times and against very specific matchups, so I'm not convinced that should sway your decision that much.
I will say I am limited by only having one draft's worth of data; once a few more years go by or I get my hands on draft info from other leagues, I might see a more interesting or definitive trend emerge.
Table of adjusted scores for all teams during 2012-2013 season.
|
No comments:
Post a Comment