Alright. Here's something to mull over across your lunchtime.
I've whipped up a version of the Elo model for the
Premier League, using the rankings source below. Let's see how it goes this weekend.
clubelo.com
I've kept the W/D/L percentages as per the international model for now (...even though the PL has a heavier weighting against away wins - 49%/28%/22%)
To make it a bit more practical let's start with a
bank of 100 units and follow the staking plan below.
High: Over 25% - Back or Lay big (3 unit)
Medium: 10% to 25% - Back or Lay (2 unit)
Low: 5% to 10% - Back or Lay small (1 unit)
No bet: Under 5% - Ignore
Rules of thumb
1. If the model proposes both a back & a lay, place
half stakes for each
2. If model proposes that there's a value angle in all 3 outcomes (e.g. back the home, lay the draw & away) aim to
back the two outcomes with the highest % diff vs bookies, as long as the bets are
consistent (e.g. don't back home win & draw)
3. For Lay bets, use the "back" odds plus 0.1, to roughly replicate the exchange
margin (typically ranges from 0.05 to 0.2 but is most often 0.1)
With the criteria above it proposes bets on 17 outcomes across the 10 games. I'd be expecting a lot of variance on the "high" category.
I think I'll start with tracking
two approaches:
A - Track the profits & losses for the approach above (17 bets this week, with stakes of 16.5 units)
B - Track the profits & losses if I only bet on the strongest recommended play in each game with full stakes (10 bets this week, with stakes of 21 units)
Saturday
View attachment 15487
Sunday & Monday
View attachment 15488
