AFL Mullumby Defender

Round 17 wrap + Round 18 predictions (Fremantle team 6 to score within 3 points of Defender - Power - not so much)

Well, I had originally optimised parameters to minimise the MAE in the margin of games previously played. I was introduced to Total Game Score betting, and duplicated my work, optimising the parameters on TGS. Surprisingly similar and also with differences. I decided to optimise on the MAE in individual team scores and predict results for BOTH margin and TGS in the one spreadsheet system. The bottom orange background is 2018 summary. Crazy close for GWS, Richmond, Melbourne, Bulldogs and Suns.

https://www.dropbox.com/s/4es6gklhhw3cu2c/summary%202018%20line%20bets%20Mullumby%20Defender%20to%20round%2017.xlsx?dl=0

Games Wins Row Labels Count of Season Sum of WIN Over 240 102 43% 2016 122 55 45% 2017 72 31 43% 2018 46 16 35% Under 282 158 56% 2016 76 43 57% 2017 126 71 56% 2018 80 44 55% Grand Total 522 260 50% These numbers are "unimproved" (for example not taking bets when the predicted total is within 5 points of the line), but already it appears that a betting strategy based on "Under" predictions would work over time. Stay tuned as I try to further improve the criteria. 1st move for this system might be to optimize on game total error instead of the current "margin MAE"

WELL, round 16 had ean Average Error in the margin of only a 16.4 (very good) we will see for rnd 17

A new area to expand into? Several online pundits have asked how well Mullumby Defender goes with total game score bets. In truth, I don't know. I DO know my predicted scores (and therefore game total) for every game for the previous 5 years. Given time I can pull the actual game totals AND the betting GTL (game total line?) from footywire. Is it possible Mullumby Defender will be even better at game total lines than normal line betting? I suppose it is worth a hard look. tonight  what was the bet tonight? Mullumby Defender predicted 66 to 61 Swans win (which with Swans -15 line would give a winning line bet on the Cats) for a total game score of 127. Actual score 59 to 71 game score of 130. Can I assume that is close enough (-3) that a successful bet could have been laid?

Bear in mind ALL Mullumby Defender predictions are from "PAST" data, but the parameters (for HGA etc. are optimised over 2016, 2017 and 2018 - cheating?). Mullumby Defender MAE for "Defender" (no blowout prediction), and Def Bo (blowout predictions) So how good are the MAE predictions? Well...... 2017 - a MAE of 28.14 is a clear winner by more than a full point. 2018...... still the clear winner, but by only 0.12 (but ALL models more accurate this year) It remains to be seen if the obviously WELL optimised parameters in Mullumby Defender are optimal for the future. Squiggle 2018 after round 15 Squiggle 2017 leaderboard

Below can be seen column after column of the optimisation process used. Initially, ALL parameters zeroed yields a  MAE of 32.4 Then find optimal Blowout and factor (35 and 21%) giving a MAE of 30.25 Zero the Blowout, find optimal "flat" HGA giving a MAE of 31.89 Optimal blowout in, find best "flat HGA (=8) giving MAE of 29.44 Find best individual HGA's with blowout 35, 16% ....... BEST blowout/penalty   Individual HGA and blowout prediction new MAE 27.92

WELL! First 2 line bets right -- then sadness two wrong - very small loss Round 16 FIVE selected bets

Hmm after weekly manually calculating the MAE of all the other Squiggle Models  I have decided to use the 2017 data to predict early 2018 rounds - we will see how well this works.

Well we only had one bet, and the Cameron KO in the 1st quarter of the highest ranked defender (who happened to be a full back) probably swayed the result. We lost by 1 goal and the restructure caused to Brisbane has reignited the red card debate. So, for those still on the train I have indicated the two games that made the cut, until reliability was applied (in blue) as well as two selected games for round 15.

https://www.dropbox.com/s/25uv6h3dmvlprxo/rounds%207%20to%2013%20Mullumby%20Defender.xlsx?dl=0 Dropbox link to excel

After the numerical analysis, I have been directed to "Kelly Criterion" which corroborates the work I did AND provides the theoretical basis to examine past Mullumby Defender Data to work out the best betting strategy for the future. http://www.elem.com/~btilly/kelly-criterion/ Reliability 0% means not using the 2018 developed reliability (at respectively winning and losing selected bets), so MORE bets, but lower % won. The column is filled with the ACTUAL percentage of the selected bets won that season. next column is number of games selected in the season Proportion is the "Kelly recommended" proportion to bet. Return over bets is calculated return for proportion, probability and number of games. Reliability 35% is the statistics with bets of reliability less than 35% discarded Reliability (% wins when selected for the winner * % losses when selected as loser for the loser) 40% is the Kelly recommendation for the reliability 35 % method, but in...

Well, Tables first Very much middle of an impressive field, of course, Defender focusses on detecting bets it performs best with. So how do the models like the line bets on offer? Interesting that ALL the models like the single defender selected bet for the round, and that there is disagreement on every other line bet. Also, interesting that for the first time Defender is 1 out on calling Melbourne to win.

Worth noting that both Collingwood and Hawthorn "would" be bets except that Mullumby Defender automatically rejects all bets with lines less than -44 (but gee I like them). No bet will be placed on them.

The wandering drunkard is the base problem for a LOT of Statistical Physics. Have a read https://medium.com/i-math/the-drunkards-walk-explained-48a0205d304 Essentially the mathematics of a "item" with a probability of stepping right (winning) or stepping left (losing), then repeating forever. So if we have a succession of bets lined up that we "know" the probability of winning (say 0.6) then what % of the betting pool should each bet be to maximise profit? If you bet 100% of the pool each time, very quickly you have no money at all. Conversely, if you bet 0.000001% of the pool each time, well your pool will barely change size. So for a given probability of winning, what is the best fractional bet for "the next step"? A numerical analysis solution (over a thousand simulated bets, starting with $100) yields if p=0.7 win    fraction bet = 68% and above => $0 return over time    Otimal bet fraction  = 35%   return (2x10^35) if ...