Target for season 2025

Each season has a fair bit of randomness built in, so it is hard to see improvement just from one set of numbers, but the goal is to do better than the numbers below.  

You can see that I started focusing on MAE and TipsBots a few years ago, with some succeess

Season     Tips           MAE      TipsBits  Squiggle relative to S10

2024    |    61.6%    |    26.76    |    28.20    |    95%
2023    |    67.6%    |    27.02    |    22.72    |    96%
2022    |    69.1%    |    25.90    |    21.24    |    91%
2021    |    67.1%    |    27.43    |    19.10    |   103%  ** 2nd place
2020    |    67.7%    |    ??.??     |    17.15
2019    |    65.2%    |    ??.??     |    13.26
2018    |    64.2%    |    ??.??     |    19.56
2017    |    57.7%    |    ??.??     |    11.07

I am not sure what happened last year, the tips were quite low but the tipsbits were a significant improvement.

 ZaphBot is back for another season

This year (2025) marks a big change for the algorithm.
Up until now I have been using an algorithm that is more or less the same as one I created back in the early 1980s, a simple algorithm that just looks at prior results of matches - no extra info other than the scores.

This year, I have moved to what I call the Team Synergy model (I just made that up while typing this).

The way it works is, for a given match I rate the two teams based on how well pairs of players in the team have played in the past.  That is 462 pairs per team (assuming 22 players listed for the game).
I then sum these scores up, compare to the other team, and pick a winner.
A few other things are considered, like how well the team does at this ground, but the main source is this synergy.

No idea how it will go, but we'll find out over at https://squiggle.com.au/