Matter of Stats

View Original

2024 - Round 12 Results - A New (False?) Dawn

In was a bad week for favourites, with just two claiming victory, allowing the targetted Head-to-Head Tipster contrarians to close the gap on Bookie Knows Best, which finds itself now tied with ENS_Linear, MoSHPlay_Marg, and MoSSBODS_Marg at the top of the Leaderboard on 68.5 from 106 (65%), and just one tip ahead of the RSMP twins

The all-Tipster average for the round came in at exactly 3 from 7, which took the season-long figure to 66.6 from 106 (63%).

Margin prediction was a little weaker this week, with mean absolute errors (MAEs) ranging from MoSHPlay_Marg’s 29.1 points per game to Bookie_3’s 37.1 points per game, with the all-Predictor average coming in at 33.5 points per game, which is up by almost 11 points per game on last week. The season-long all-predictor average is now 26.4 points per game per predictor.

RSMP_Simple’s performance of 26.1 points per game was good enough to see it climb into top spot, the other climbers being ENS_Linear (into 3rd), RSMP_Weighted (into 4th), and MoSHPlay_Marg (into 7th), who is now less than 5 goals off the lead.

In fact, the current gap between 1st and 8th is less than 7 goals.

Log probability scores were all close to zero this week and virtually the same for all six forecasters, although mildly negative for the bookie-based forecasters and mildly positive for the MoS trio. Consequently, Bookie_LPSO retained 1st place ahead of Bookie_RE, and MoSSBODS_Prob remains worst. MoSHPlay_Prob still lies in 4th and is the best of the MoS trio.

WAGERING

Round 12 was the neqar antithesis of Round 11, where almost everything that Investors touched turned to money. Only the Overs/Unders Fund failed to return a profit, and its losses were avalanched by the profits from the other two Funds.

Overall, the Combined Portfolio finished up by 12.4c - comfortably the highest single round return of the season, and possibly all-time - which leaves it down now by only 0.2c, that from a -0.1% ROT on a 1.96 turn.