Greyhound Mystique

[jimibt]

Anyone seeing unpredictable results over the last few days ?. My backing strat usually runs at a strike rate of around 17% (15k sample). Since yesterday I've been running at around 4%. Obviously could be just variance but wondering if the weather might be a factor...

I did Sunderland today with 4 x winners and 2 x 2nds, and Nottingham provided 5 x winners and 2 x 2nds from 12 races each (all bets above 2.02) not necessarily market favourites, but my own selections.

Some tracks seem better suited to favourites below 3s to win, managing to place consistently.

CS

therein lies the tale!! i'm TRYING to AVOID conventional wisdom... WIP !!

[wearthefoxhat]

Couldn't resist posting this one....Scurlogue Champ

https://www.youtube.com/watch?v=Dz0Utl30A0U

[Archery1969]

Has anyone ever used Neuroxl Predictor ?

Possibly a complete fluke and I would never have bet on this dog to either win, place or shorten in price pre-off.

But it won.

[Realrocknrolla]

Has anyone ever used Neuroxl Predictor ?

Possibly a complete fluke and I would never have bet on this dog to either win, place or shorten in price pre-off.

But it won.

Saw this post earlier and just checked my model out.

My model predicted this a top finish also, and I included it in my dutching selections for this race.

To be fair the dog ran a fast time last time out also.

Trap 2 was definitely a false favourite.

[lavenham]

Hi,
With the shortage of decent horse racing lately due to bad weather, I decided to revisit greyhounds. I came across this interesting topic, which has about 79 pages of high quality input, spreadsheets etc. Sadly there has been no input since 2021 and I was interested to know why. Has everyone found a better approach which they do not wish to share or have they abandoned greyhounds all together? The last greyhound input I could find was Dallas's 'Leap frog' servant, which might possibly work well with Archery 1969's approach.
There are a host of 'scraping' approaches using Python but not sure if they work anymore, but if they do then they would be a great way to check out different angles.
So - where have all the experts gone?
Thanks

[Archery1969]

Hi,
With the shortage of decent horse racing lately due to bad weather, I decided to revisit greyhounds. I came across this interesting topic, which has about 79 pages of high quality input, spreadsheets etc. Sadly there has been no input since 2021 and I was interested to know why. Has everyone found a better approach which they do not wish to share or have they abandoned greyhounds all together? The last greyhound input I could find was Dallas's 'Leap frog' servant, which might possibly work well with Archery 1969's approach.
There are a host of 'scraping' approaches using Python but not sure if they work anymore, but if they do then they would be a great way to check out different angles.
So - where have all the experts gone?
Thanks

I for one worked out that I was trying to make things too complicated. Each race day is different. Better to trade them by using a combination of “find the gap” and “leap frogging” plus some additional adjustments. Which you can apply to any low liquidity markets, not just greyhounds. The key is working out when the value disappears, which actually turned out to be very simple.

[Atho55]

Different approach I am presently looking at is comparing Rank of Odds v Rank of Volume and if it can help in determining the outcome of the race. Lowest odds would be Rank 1, Highest Vol Traded is Rank 1.

Looks like this

Rank of Odds v Rank of Volume.png

[sniffer66]

Different approach I am presently looking at is comparing Rank of Odds v Rank of Volume and if it can help in determining the outcome of the race. Lowest odds would be Rank 1, Highest Vol Traded is Rank 1.

Looks like this


Rank of Odds v Rank of Volume.png

Have you looked at BSP and Volume in the Historical data ? Assuming BSP is going to be close to a few seconds from the off.

[lavenham]

Thanks for replies. Interesting to see Archery 1969 has simplified things.
My best, most consistent results on greyhounds is simply to look at where the traded volume is and a simple 2 sec 50 DP candle chart, from around 4 minutes mark. Sort of similar to my horses approach with a few indicators omitted due to shortage of data on greyhounds. Mostly concentrate on dog with biggest market share and lower price, usually the 1st or 2nd Fav. As ever, the better the grade of race the more consistent the result.

[ilovepizza82]

I've been exploring whether there's a connection between volume and the number of races won, seeking any potential advantage. I've written a simple Python code to analyze this, but I'm uncertain about its accuracy. Here are some findings for the year 2023:

- The dog with the biggest volume won 12375.0 times out of 36567 races.
- The average odds of the races won: 2.97
- Number of times a dog with the biggest volume won three times in a row: 420
- Number of times a dog with the biggest volume won four times in a row: 87

Yea I know, nothing spectacular and It doesnt take into account rating so its probably useless but wondering if its somewhat correct in calcs.

[wearthefoxhat]

I've been exploring whether there's a connection between volume and the number of races won, seeking any potential advantage. I've written a simple Python code to analyze this, but I'm uncertain about its accuracy. Here are some findings for the year 2023:

- The dog with the biggest volume won 12375.0 times out of 36567 races.
- The average odds of the races won: 2.97
- Number of times a dog with the biggest volume won three times in a row: 420
- Number of times a dog with the biggest volume won four times in a row: 87

Yea I know, nothing spectacular and It doesnt take into account rating so its probably useless but wondering if its somewhat correct in calcs.

If you breakdown your data by track, you might find an edge.

Each racing manager grades' their races slightly differently, some like to see a runner win 2+ in a row, others put them in harder races.

[Crazyskier]

Hi,
With the shortage of decent horse racing lately due to bad weather, I decided to revisit greyhounds. I came across this interesting topic, which has about 79 pages of high quality input, spreadsheets etc. Sadly there has been no input since 2021 and I was interested to know why. Has everyone found a better approach which they do not wish to share or have they abandoned greyhounds all together? The last greyhound input I could find was Dallas's 'Leap frog' servant, which might possibly work well with Archery 1969's approach.
There are a host of 'scraping' approaches using Python but not sure if they work anymore, but if they do then they would be a great way to check out different angles.
So - where have all the experts gone?
Thanks

I still bot the dogs daily, though it's usually a few tried and tested systems for Oxford and Swindon mostly as almost every other track has now lost their edges for my systems which involve backing or laying to win / place and letting them running the races.

CS

[Jason89]

I've been exploring whether there's a connection between volume and the number of races won, seeking any potential advantage. I've written a simple Python code to analyze this, but I'm uncertain about its accuracy. Here are some findings for the year 2023:

- The dog with the biggest volume won 12375.0 times out of 36567 races.
- The average odds of the races won: 2.97
- Number of times a dog with the biggest volume won three times in a row: 420
- Number of times a dog with the biggest volume won four times in a row: 87

Yea I know, nothing spectacular and It doesnt take into account rating so its probably useless but wondering if its somewhat correct in calcs.

The calculation sounds about right, because they are winning about a third of the time, the favourite is usually the one with the highest volume and they win about a third of the time.

[Archery1969]

Hi,

Been trialing something new to use when Greyhound markets are first formed. But no reason why it couldn't be used on all markets. It looks for value based on the odds being offered and how they change as money enters the markets.

I don't care about form, previous history, just how the market changes. Bets are placed at random times from x seconds to y seconds and amounts offered are also random from x amount to y amount. All designed not to spook the markets.

Be interesting to see how the professional market makers react although I guess they probably wont care. But I will be jumping in front of them if value is percieved to be there.

Below is just a screenshot of the workings, the BF back/lay and amounts are just random to makesure my calculations were working. Its now being coded in Python accessing the BF API by some coders as I want maximum speed. Eventually it will run on a VPS.

Lets see where this one goes.

[lavenham]

Will be interesting to see how this works out. Certainly a different approach. Good luck with it and hope it is profitable.