Trading or betting automation, and with what data?

We've gone to the dogs.
Post Reply
Cipobotter
Posts: 8
Joined: Tue Feb 23, 2021 10:40 am

Hi All,

I have been a lurker around here for a year now, learning the ropes through my several mistakes and the various posts.

About one year ago I have started my journey into automation betting, first on the horses, then on the dogs and football. I have mostly worked on trying to develop betting systems for dogs, although I have sometimes approached also trading. I like working on the dogs, mostly for the consistent conditions (e.g. # of races per day, # of runners, limited confounding factors, there is stability in the stochastic nature of their results) although I am aware of the problems relative with the low liquidity. I mostly do my own coding (in Matlab mostly) for both data analysis and betting, although I have been studying betangel recently. Needless to say, I am hopelessly in the red, and although it burns, I now approach the losses as part of the learning investment.

I am an engineering scientist by trade and skilled in programming and data analysis, with a good career, and, as most people around here, I do not have the time to follow the events closely, nor interest in switching from my career to trading/betting full time (unless, of course, it allows me to retire to Spain before I am 45). For these reasons I am interested in automation. I have approached betting/trading as an expensive hobby, and I usually spend a couple of hours of bot coding in the evening. I am writing because I find myself a little stuck at the moment and I wanted, it possible, to get some general hints from the people who have cracked the dogs. In the past year I have developed several mildly successful (mostly) backing bots, but none of them has endured the test of time.

Given my inclination for data analysis, I have literally dissected the betfair historical data on the dogs, looking for any kind of trend or inconsistency that could be exploited. I have been looking for inefficiencies in the BSP odds depending on favorites/volume/tracks/traps/odds of other favorites, creating any sort of custom signals, but I haven't found anything really consistent. After some time, all of the many inefficiencies that I find regress to their efficient counterparts. I am fine with this, and I understand that this is an inherent characteristics of these markets. I am however wondering if trying to model the market through its inefficiencies in the expression of market sentiment is falling, again and again, in the gambler's fallacy.

What I am trying to understand is:

a) Do successful betting greybotters (I am sure to have spotted a few here) have a number of consistent strategies (that they perhaps tweak or train over time) or do they consistently and continuously jump from strategy to strategy?

b) Does it make sense at all to create betting systems or is the money mostly in trading? I have approached trading but I am hesitant to throw myself at it, mostly because I would need data (that I can record myself from my bots) that would take a substantial amount of time and a pc always running to get. My dream would be to work also on trading once I have cracked the betting, but I don't want to start by trial and error, I need to do substantial analytical modeling before.

c) What data to use to model the market and make the bots? Is there value in the historical betfair data or do people use mostly other data (e.g form, timing)? I have seen a very nice use of form/time data in the Greyhound Mystique (impressive job!), in a logic-informed way (e.g. diff in favs times drives sentiment of shortening in price). I am trying to understand if something similar can be done for betting but going over the classic "the fastest dog wins more often", as that should already be expressed in the odds. Ultimately I am not interested in understanding which dog has the better chance of winning, but just what event is more likely than what implied by its odds, as Euler says, and I am wondering if there is value in form data for that.

I am sure these questions have been asked by other newbies like me over and over again, and I am prepared to not get any reply, but if any of you would be interested in telling their experience without spilling the beans on their successful methods that would be great . I understand that in this line of making money information is the most valuable commodity, so what I am asking is more about what is the behavioral approach to bot creation of people who have cracked the code.
Thanks to all (and apologies for the long post)!

Cheers!
Cardano
Posts: 57
Joined: Mon Aug 21, 2017 12:02 pm

I don't think you'll find anyone who is making outright betting automation backing/laying profitable.
Trading either manual/auto is the only way to be profitable.
LinusP
Posts: 1772
Joined: Mon Jul 02, 2012 10:45 pm
Contact:

Cardano wrote:
Tue Feb 23, 2021 10:03 pm
I don't think you'll find anyone who is making outright betting automation backing/laying profitable.
Trading either manual/auto is the only way to be profitable.
:lol:

Cipobotter I have sent you a PM
Anbell
Posts: 948
Joined: Fri Apr 05, 2019 2:31 am

LinusP wrote:
Wed Feb 24, 2021 6:08 am
Cardano wrote:
Tue Feb 23, 2021 10:03 pm
I don't think you'll find anyone who is making outright betting automation backing/laying profitable.
Trading either manual/auto is the only way to be profitable.
:lol:

Cipobotter I have sent you a PM
Publish it here!
liltbrockie
Posts: 11
Joined: Thu Sep 26, 2019 7:53 am

Cardano wrote:
Tue Feb 23, 2021 10:03 pm
I don't think you'll find anyone who is making outright betting automation backing/laying profitable.
Trading either manual/auto is the only way to be profitable.
I'm making about at least 5% a day doing it.... Caveat though it is using external tip data.

Image

The above is for Feb so far.... @ £2 stakes

new total = final p/l for the day
max new = maximum profit reached ever during the day...

I normally call it a day at 5%
Cipobotter
Posts: 8
Joined: Tue Feb 23, 2021 10:40 am

Hi Liltbrokie and all,

Thanks for your replies! Any info is useful. I am pretty sure that automated straight betting is possible, but only using a tactic that exploits the market mechanics. All the rest is hoping that what happened before continues to happens.

Liltbrokie, is that betting or trading, if I can ask? I have thought about scraping tips as they are released, but mostly from a trading perspective (so, is in the to-do list for the future).



liltbrockie wrote:
Wed Feb 24, 2021 8:40 am
Cardano wrote:
Tue Feb 23, 2021 10:03 pm
I don't think you'll find anyone who is making outright betting automation backing/laying profitable.
Trading either manual/auto is the only way to be profitable.
I'm making about at least 5% a day doing it.... Caveat though it is using external tip data.

Image

The above is for Feb so far.... @ £2 stakes

new total = final p/l for the day
max new = maximum profit reached ever during the day...

I normally call it a day at 5%
goat68
Posts: 905
Joined: Tue Jun 30, 2020 3:53 pm
Location: Hampshire, UK

"market mechanics" , another one of those secrets?
liltbrockie
Posts: 11
Joined: Thu Sep 26, 2019 7:53 am

This is straight betting.... I do not consider trading to be scalable enough for the money I want to make.

If you are straight betting you really want to be doing a mixture of back and lay bets hopefully of similar prices.

The problem with only laying is it will be very difficult to recover your losses when they happen... at least if you are backing as well at similar prices, your chance of recovery is much better.

It is a lot easier to recover from a BSP 6.5 Lay loss with a 7.0 back bet than it is to get another qty 7 lay wins.

You are then simply left with the task of making sure that you are achieving more backs wins than lay losses for the day! Easy right?

Pro Tip...do not waste your time analyzing WOM, rank, odds price, price movement in the last X seconds, money matched etc... it's all meaningless. These are not indicators.

Get into the markets early in the morning and get out as soon as you have made a fair target profit for the day.
Cipobotter
Posts: 8
Joined: Tue Feb 23, 2021 10:40 am

That's interesting... I think I am understanding what you do.... I do not completely agree on the validity of the indicators (or non-indicators) you said, again they are not going to tell you which dog wins, but are certainly exploitable.
liltbrockie wrote:
Wed Feb 24, 2021 9:41 am
This is straight betting.... I do not consider trading to be scalable enough for the money I want to make.

If you are straight betting you really want to be doing a mixture of back and lay bets hopefully of similar prices.

The problem with only laying is it will be very difficult to recover your losses when they happen... at least if you are backing as well at similar prices, your chance of recovery is much better.

It is a lot easier to recover from a BSP 6.5 Lay loss with a 7.0 back bet than it is to get another qty 7 lay wins.

You are then simply left with the task of making sure that you are achieving more backs wins than lay losses for the day! Easy right?

Pro Tip...do not waste your time analyzing WOM, rank, odds price, price movement in the last X seconds, money matched etc... it's all meaningless. These are not indicators.

Get into the markets early in the morning and get out as soon as you have made a fair target profit for the day.
liltbrockie
Posts: 11
Joined: Thu Sep 26, 2019 7:53 am

Definitely exploitable if trading... not so much for winning bets.

I have seen too many odds on favorites backed to the eyeballs only to be beaten by 100/1 outsiders to pay any attention to pre off clues......
Post Reply

Return to “Trading Greyhound racing”