Final Season Rankings

The above rankings are for the entire 14 week season.  While my weekly rankings are meant to show how a team is playing right now based on their last 4 weeks, this rankings shows the entirety of a team's season.  The "act w" (actual wins) column shows their actual wins for this year.  The "expected result" column shows how we expect that team should have done given their season long Estimate Win Percentile.  Anyone familiar with advanced baseball metrics will recognize this as expected wins.  Obviously the difference is what we call luck, although there may be more to it than that.

Drow and Lagers were the class of the league this year as was pretty clear all season.  Old No. 7, Malleus, and Hamburglers were the only other teams that the model expected would have a winning record at the end of the year.  Malleus were easily the unluckiest team this year but they also were essentially neck-and-neck with Old No. 7 -- who also did not make the playoffs.

Note that there are two reasons that the .500 percentile mark does not fit neatly between the 6th and 7th ranked teams.  The biggest reason is that I use 2 year's worth of data to as the context for calculating EWT in order to give me a bigger data set.  And scores were generally down this year compared to last.  The median total score this year was 70 points.  That is the lowest since 2008 (67).  Last year it was 73.  Offensive median was 43.  Also the lowest since 2008 (42).  Median offense has been 48 for the last 3 years prior to this year.  Median defense was 24 which is only a point lower than last year.

Above we see a little more detail on how teams ended up where they did.  All numbers are medians if not otherwise indicated.  So Drow's median total score was 84 (a variation on a power score).  Then we see that his median offensive score was 52 points, his median percentile score on offense was .63, average percentile offensive score was .63, variance of percentile offensive scores is .30.  The line across the top is the league wide medians in those columns for this season.

A couple maybe obvious things on this table.  When a team's average is below a team's median it means that they had larger spikes below their median.  An average that is above a median indicates larger spikes above the team's median.  And perhaps obviously, a lower variance indicates greater consistency.  If there are any stat geeks out there, variance here is actually standard deviation and excel calculates standard deviation from the average not the median.

Below are the scores for every game played this year (its a wide image, you might have to click to make bigger).  Color-coded by percentile as always: red = 83rd percentile or roughly top 2 every week, blue = 66th percentile, green = 50th percentile, yellow = 33rd percentile, tan/grey = 17th percentile, black = bottom 17 percent of scores.  What that means in actual scores in the pic to the right.

Finally, a quick look at the divisions.  "W +/-" is essentially how many wins the division had as a whole against teams in other divisions.  "Expected w +/-" is how many we should expect based on their EWP and then luck is the difference.

Not a shock that the Left came out with the highest average EWP for all four teams.  Team scores suggested that the Middle would be the number two division but it just did not play out that way in the results, with the Right Division performing a lot better than expected against the other divisions.  And that matches what we know about the two luckiest teams this year (Bill and Mimes