General Betting
Welcome to Live View – Take the tour to learn more
Start Tour
There is currently 1 person viewing this thread.
I've been on a long journey investigating the use of thermal imaging in horse races, particularly towards the finish.
For those not familiar with the technique it involves use of infra red to determine whether or not a horse is running "hot" or "cold", and therefore the likelihood of the horse finishing strongly or fading in the closing stages. I have now managed to produce an impressive and consistent means of significantly improving the prediction of the possible outcome of races in the latter stages of in-running, and profits are flowing.
My problem is one of coverage. It is necessary to attend races with the right camera gear and set-up which limits the number of races I can cover, and also involves the full-time pain of travelling everyday to meetings.
Is any body else using this technique, and are they aware of any methods of deriving these images without the need to travel? It would be so much better to be able to do this from my armchair. The technique will withstand an element of delay - so whilst perhaps not as profitable, it would work without being track-side providing the images are not too delayed (say 2-3 seconds or so).
Ideas gratefully received.
For those not familiar with the technique it involves use of infra red to determine whether or not a horse is running "hot" or "cold", and therefore the likelihood of the horse finishing strongly or fading in the closing stages. I have now managed to produce an impressive and consistent means of significantly improving the prediction of the possible outcome of races in the latter stages of in-running, and profits are flowing.
My problem is one of coverage. It is necessary to attend races with the right camera gear and set-up which limits the number of races I can cover, and also involves the full-time pain of travelling everyday to meetings.
Is any body else using this technique, and are they aware of any methods of deriving these images without the need to travel? It would be so much better to be able to do this from my armchair. The technique will withstand an element of delay - so whilst perhaps not as profitable, it would work without being track-side providing the images are not too delayed (say 2-3 seconds or so).
Ideas gratefully received.
Show More
Nice work Artisan. Wish I could help with that, sounds like a fascinating project!
it is amazing how far people go to win
. good luck with ur problem
. good luck with ur problem
Predator would have done well at horse racing 
how large is the equipment... also what is the level of temperature you are looking for to decide if the horse wins or fades...
how effective it is compared to watching horses in paddock and see if they are sweating?
sounds like oil sands ... good in theory difficult / costly to implement...
how large is the equipment... also what is the level of temperature you are looking for to decide if the horse wins or fades...
how effective it is compared to watching horses in paddock and see if they are sweating?
sounds like oil sands ... good in theory difficult / costly to implement...
The equipment's not too bad - reasonable quality IR camera (or modified and filtered standard as a starter) connected to a laptop which automatically processes the images in real time. There's some pattern recognition software running which has been trained on previous images and data to forecast likely outcome.
I don't know how it compares to looking at horses in the paddock, although I've tried various things from the off. The most effective I find is in the last few furlongs (depending on the length of the race) when a horse goes sub say 2.0 depending on it's temperature (and on what's around it) my system either automatically backs or lays. The edge is around 30%. No other systems I've played with over the years has got remotely close to that. So it definitely works, and an average day out will more than cover the costs of the initial outlay on the kit, although there has been a great deal of work getting the right camera set up, the pattern recognition software tuned, and you need to gather some history to be able to do that.
I don't know how it compares to looking at horses in the paddock, although I've tried various things from the off. The most effective I find is in the last few furlongs (depending on the length of the race) when a horse goes sub say 2.0 depending on it's temperature (and on what's around it) my system either automatically backs or lays. The edge is around 30%. No other systems I've played with over the years has got remotely close to that. So it definitely works, and an average day out will more than cover the costs of the initial outlay on the kit, although there has been a great deal of work getting the right camera set up, the pattern recognition software tuned, and you need to gather some history to be able to do that.
and the race courses are ok with you using this equipment.. 30% is very good edge..
I don't advertise what I'm doing at courses. Although I need a laptop/notebook with a dongle - I need a good mobile connection - I don't wonder around with it open, so I guess I just look like a guy with a camera. Never had anyone question me.
I am nominating this as thread of the year !
A 30% edge on 2.00 is fantastic!
Question? From snap to back/lay decision, how much time elapses?
A 30% edge on 2.00 is fantastic!
Question? From snap to back/lay decision, how much time elapses?
Wow, a 30% edge is impressive and if you can cover the cost of the hardware in a single day why don't you hire someone to sit at each race track for you?
I'm sure that given the economy you could get someone at a reasonable rate, they don't have to know how much profit you make from the research you've done. You could set up the notebook so that it sends you the data and they don't see the trading decision.
Would it be even less conspicuous and perhaps cheaper to have your pattern recognition software running on a tablet?
I'm sure that given the economy you could get someone at a reasonable rate, they don't have to know how much profit you make from the research you've done. You could set up the notebook so that it sends you the data and they don't see the trading decision.
Would it be even less conspicuous and perhaps cheaper to have your pattern recognition software running on a tablet?
Ah yes we certainly have come a long way since backing the horse that has a dump in the parade ring.
Very impressive. Does it work for both flat and jumps? So you're looking for people to go to the course and then split the profits with? I'd be willing to give it a go if the number were right. I'd be able to do the south of england.
how obvious would it be though if you wonder around with a camera aimed at the race and a laptop on ya lap,
youd be noticed quite quickly and throw out surely?
youd be noticed quite quickly and throw out surely?
Oct 23, 2012 -- 3:30PM, Carlos4441 wrote:
how obvious would it be though if you wonder around with a camera aimed at the race and a laptop on ya lap, youd be noticed quite quickly and throw out surely?
Simply put foil around your ankles and shins.
Anyone asks, say horses are from the Xenon-sipher galaxy and as they pass, they transmit messages that you are capturing.
Bemused, they will leave you alone.
Can't give too much away but here's an image from my 'Hot Jockey In Running' software. For obvious reasons, I cannot reveal the temperature scale.[/size]
I would like to know if this technology would be possible to be adapted for use in meeting suitable women for "a good night out " ?
Learning to smell pheromones works better and you don't need the app.
should be more concerned that in the front row there is a giant who is twice the height of every other spectator.
@Fine as frog hair ... the Saville method works well the colder the better !
Interesting project. How are you identifying the runners in real time? I have it misclassifying about 1 horse in 1000 using replays when it has the whole race to build up a knowledge. I wouldnt like to use it for automated betting just using the closing stages in real time.
Post Your Reply
Please login to post a reply.
Wonder
Instance ID: 13539