total brownlow

We’ll have plenty more Brownlow betting pointers this week. To put your name down for our limited Total Brownlow package release – just click here.

The Brownlow is this Sunday night, and our Total Brownlow package will be available again as per last season. Last season’s hard work in setting up our two different Brownlow simulation models working paid off, and we had 73 bets for a +11.5u win at 11.1% POT. We also released full price ratings on most Brownlow markets so clients could find their own bets.

Given our groundwork last year, we’re already done with pricing all markets, and will continue to monitor and price new Brownlow markets appearing. For Total Brownlow, full price ratings on every market will be provided along with a bet list. We’ve also had enough time to add the markets we didn’t price up last year, so we have added Quinella, Boxed Trifecta and Boxed First Four bets and price ratings available for all combinations. There is now absolutely nothing that our Brownlow simulation models don’t price.

As always, you have to be very careful of looking at projected vote counts and trying to eyeball value. You put two players together with the same projected count, and one can end up being $1.75 and the other $2.32 on H2H rated prices. It’s because the projected count really should be at the centre of a range of different vote possibilities that needs simulation to get the pricing right.

Total Brownlow: Pricing pitfalls

To illustrate with a hypothetical extreme example: one player with 15 x 1 vote projections vs another player with 5 x 3 vote projections in a head-to-head match-up.

They could end up in a significantly different position than simply assuming H2H prices should be 2.00 / 2.00, because both have the same vote count.

When you extend this problem into group props, the idea of finding value between 5-6 players – when they all have unique round-by-round vote projections and unique round-by-round simulation distributions – is a very tough ask. Especially when you’re betting into 125%+ markets.

You can’t get a better example than Neale to win Brownlow.

Every Brownlow leaderboard has him winning, but what’s his true price to win? The market generally has him $1.25, but maybe he’s really $1.10? Or maybe even as high as $1.55? You don’t know unless you simulate the whole count with all players thousands of times, and count the amount of times he won to form a probability/price estimate.

What not to bet

Here’s a great example of the huge amount of black holes for your money in Brownlow group props:

Nic Naitanui Top 10: Book price is $1.90, our price is $21.54. -91.2% edge

Scott Pendelbury Top 10: Book price is $3.00, our price is $44.30. -93.2% edge

What you should bet

We’ve already released (to AFL Tips members) a Top 10 play where we are very different to the bookmakers. We always simulate everything twice across two very different kinds of simulations. One simulation prices him for Top 10 at $6.50, and the other (arguably stronger simulation) prices him at $2.86. To be fair there is a good chance of dead heat rules compromising the odds, but even so we still have a tremendous overlay.

He has almost no chance to make Top 3, but at fourth position both simulations start to see a small peek on this player, and then he starts to come in very hard from sixth position onwards. If the Brownlow was a huge horse race with 10 paying positions he’d be classed as a big time place bandit!

What the books are doing is a similar variant to the crap they used to do years ago on horse racing – having a 1/4 win odds yardstick as the starting point for place odds. This kind of extrapolation is no good at all. That’s why this Top 10 play is such a good price: the books are not doing the necessary work. They probably have a rough 1/10 guide and use their win odds framed at a cushy 130% and then think they will be fine offering roughly 1/10th of it for the Top 10.

The umpires’ award

Above all, remember it’s an umpire’s medal.

These umpires and figuring out how they respond to player performance is at the heart of the whole problem. The more work here the better, and figuring out individual umpire preferences can be a big edge. Often the biases of a key senior umpire like Brett Rosebury or Ray Chamberlain can have an outsized impact on that game’s votes if they are matched with junior umpires. Network analysis across all games over multiple seasons is one way to reveal the key influencers among umpiring crews.

This umpire knowledge then feeds back into the simulation models. There are many hundreds of slight changes all over the two differing simulations from this umpire bias analysis. A player’s 3-vote estimate for a given game could vary signifcantly because of a favourable umpiring crew. Two percent doesn’t sound like much, but when you add up many hundreds of slight perturbations like this across the entire playing cohort and across the entire season, and you can end up at price points on Brownlow markets that can be very different to the bookmakers, and much better.

We’ll have plenty more Brownlow betting pointers this week. To put your name down for our limited Total Brownlow package release – just click here.

Pro Golf Preview: 2020 US Open

Full preview of the US Open with The Golf Insider Limited time offer – double your monthly membership It’s for…Read More

Pythagorean Theorem: Forecasting the NFL with high school maths

Bill James brought the Pythagorean Theorem to light many years ago as a way to indicate a baseball team’s future…Read More

NRL Pro Preview: Dragons vs Titans

Dragons vs Titans preview from rugby league expert Steve Green including recommended bets. Steve Green has four straight winning seasons…Read More