If any of your friends or colleagues build statistical models for a living and pretend that they don’t hurt when their preferred models let them down, I’ll let you in on a secret: they’re lying, it hurts, and quite a bit.

That said, it was hard to be critical of MoSHPlay_Marg this week, at least as a Head-to-Head Tipster, where it bagged an equal competition high of 7 from 9 to leave it now 3 tips clear of any of its rivals.

Those who know me best though, will understand how little store I place in forecasters whose only claim to fame is an ability to pick winners, because that talent is maybe 3 parts ability and 2 parts luck. The best forecasters are calibrated enough to tell you not just which team they think will win, but by how much they expect they win, and how confident, in probabilistic terms, they are about a team’s chances.

So, in contrast, as a Margin Predictor MoSHPlay_Marg finished last this week on mean absolute errors (MAEs), which was enough to shunt it back also into last place on the Leaderboard, now almost 12 goals behind the leader That hurts way more than a 7 from 9 tipping performance.

Bookie_9 did enough with its 28.5 ppg MAE to retain 1st spot, whilst Bookie_LPSO’s 28.6 MAE was sufficiently less-bad than others’ to see it finish the round in 2nd, now just over 13 points off the lead.

The round’s best MAE of 27.6 ppg belonged to RSMP_Weighted who found themselves, post round, in 4th place.

All of the Head-to-Head Probability Predictors again returned positive log probability scores this week, best amongst them MoSSBODS_Prob, but by only a tiny amount. It still sits in 1st place, ahead of MoSHPlay_Prob whose probability estimates will surely, eventually, consume it and burn it to the ground. MoSHBODS_Prob currently sits in 3rd place forming a barrier between the MoS trio and the Bookie-based trio.

WAGERS

Round 12 produced another loss for Investors, the 4th in 6 rounds, though only the 3rd-largest of the season.

That loss was, unusually, entirely attributable to the Line Fund, who this week registered only a 1 from 4 performance, which saw it shed almost 3% of value.

Overall, the Combined Portfolio’s price decreased by 1c, leaving it with a price of $1.088, this result coming from a 13.4% ROT on a 0.658 turn of the Portfolio Funds.