Value betting several runners in a race.

[Tiger]

I have been working on what I'm happy to call true value for the greyhound races at various tracks and traps.

If I'm confident that I have found value odds for both traps 1, & 6 for examples should I put bets on both dogs or pick just one?

What happens if I found value odds for 3 or 4 more of the 6 dogs?

Do I not erode my value or profits by placing multiple single bets on a one winner event?

I usually place bets with a target profit rather than a fixed stake. Is this still the right approach for multiple back bets in the win market event?

[ShaunWhite]

Bet everything you see as value, dog markets are small so as you scale up the stake will limit itself pretty soon.

But while you scale up, Kelly staking, or partial Kelly would give you a guide, although simply starting small and gradually increasing every few hundered bets would be a fair call too.

[Anbell]

I have been working on what I'm happy to call true value for the greyhound races at various tracks and traps.

If I'm confident that I have found value odds for both traps 1, & 6 for examples should I put bets on both dogs or pick just one?

What happens if I found value odds for 3 or 4 more of the 6 dogs?

Do I not erode my value or profits by placing multiple single bets on a one winner event?

I usually place bets with a target profit rather than a fixed stake. Is this still the right approach for multiple back bets in the win market event?

You should bet all of them that have value, even if it is 6 out of 7.

In theory you should back all so that you receive the same payout (not the same win) if the animal wins. (this can be refined further)

[wearthefoxhat]

The big syndicates (not necessarily greyhounds) back more than 1 selection in a race that is "value" to them.

Syndicates like CIF, Zeljko-style ops, or the HK giants run full probability models for the entire field and then compare those to the market’s odds in real time. If more than one runner shows positive expectation > threshold, it triggers multiple bets.

In fact, not backing all value runners would actually lower their EV.

---

On a side note, I remember when Jennings bookmakers on the high street did a Cheltenham festival offer for their chosen race of the day (2003), where you could back any runner at double the SP. Great for the casual punter, maybe back the favourite at 2/1 and get 4/1.

However, when a few of us got together with a spreadsheet and a bankroll, we went around all the Jennings bookies, backing every runner and took the prices available. Then, 10 minutes before the off, made adjustments to the any late steamer. A drift was a gift. (a 120% book becomes a 60% book)

There were only 3 days for the Cheltenham festival back then and on day 3, one of the branches got pissed off and refused to take any more bets, so we just added those bets to the final shop, then watched the race.

It was an edge based on overall maths, it worked as long as we got the maths right, and got all the bets on.

---

[sniffer66]

The big syndicates (not necessarily greyhounds) back more than 1 selection in a race that is "value" to them.

Syndicates like CIF, Zeljko-style ops, or the HK giants run full probability models for the entire field and then compare those to the market’s odds in real time. If more than one runner shows positive expectation > threshold, it triggers multiple bets.

In fact, not backing all value runners would actually lower their EV.

Not just the big syndicates. I'm doing almost exactly this using ML models and capturing real time odds data. I'm laying but the value model is the same against the calculated probability

[ShaunWhite]

The big syndicates (not necessarily greyhounds) back more than 1 selection in a race that is "value" to them.

Syndicates like CIF, Zeljko-style ops, or the HK giants run full probability models for the entire field and then compare those to the market’s odds in real time. If more than one runner shows positive expectation > threshold, it triggers multiple bets.

In fact, not backing all value runners would actually lower their EV.

Not just the big syndicates. I'm doing almost exactly this using ML models and capturing real time odds data. I'm laying but the value model is the same against the calculated probability

Agree, similar. I avg about 10 bets on avg 2.5 dogs per race and just over 4 horses, but could be none or all. But I'm not using fundamentals. Live data -> TFT model - > signal from -1 full lay to +1 full back. Fairly std stuff these days really.