Anyone who ever tells you they have a 100% error-free program or script that does anything much more than print out “hello world” is certifiably deluded.

I discovered tonight, as I was doing the various updates for this blog post, that my script to estimate player value had been ignoring each player’s most-recent game. It doesn’t have much effect (the correlation across player ratings before and after the fix is +0.997) but it is screamingly obvious for all the Round 1 debutants whose estimated ratings stayed at 45.

Anyway, here’s the latest from MoSHPlay, assuming that the currently designated “emergencies” won’t slip onto the bench (narrator: some of them will).

So, with the exception of the Swans v Crows game, MoSHPlay is tipping the same teams as MoSHBODS, with about a 5 points per game average absolute difference in margins.

More updates to come.

FRIDAY 6PM UPDATE

Updated MoSHPlay forecasts are below. Note that these have changed for every game, despite the fact that some teams are unchanged relative to the teams that were announced last night. Spreadsheet error? Yep. (It’s always such fun the first few weeks of a new season …).

The changes of most consequence are that MoSHPlay is now tipping Sydney and West Coast rather than Adelaide and GWS.