Can a logical approach be applied to an illogical market?

The sport of kings.
Post Reply
mhorro
Posts: 112
Joined: Sun Feb 17, 2013 1:16 pm
Location: Cheshire

I am a computer professional with 25+ years in computer science.

I have spent 1 month logging signals in the UK horse racing markets i.e. horses hitting a certain prices, number of selections, fav price pre-off the usual variables.

In a nutshell we are dealing with animals ( horses ) and ( human or automation ) to trade the markets. I have not even factored in delays for in-play on transmissions.

So we have no logic here just a P&L over a period of time!

It is very frustrating for someone who has worked in in the computer industry to try and make sense of this.

Is gambling such a bad thing? It seems easier!!!!!!


I will not mention training mode as it only seems to test functional flows!

Thanks,

Mark
User avatar
northbound
Posts: 737
Joined: Mon Mar 20, 2017 11:22 pm

I worked as a software developer for 15 years and am now dedicating to learning how to trade. Haven't reached consistent profitability yet.

Trading is a completely different activity compared to software development and my technical / logical background probably got in the way a lot so far.

There are no zeros and ones here. There are no rules that work every single time. Trading requires developing a feel for what's going to happen, accepting that you'll be right some times and wrong others.

I agree that it is very frustrating for someone who has worked in in the computer industry to try and make sense of this.

But with time, persistence and passion, I'm confident we can make it.
User avatar
ShaunWhite
Posts: 9731
Joined: Sat Sep 03, 2016 3:42 am

It's 37 years since I wrote my first code. Was paid to do it for 27.

I think it's a big help, certainly with integration of BA into a larger system, data analysis and management, configuring automation, archiving and versioning, testing, writing your own utilities etc. The guys struggling with Guardian signals, excel and vba would kill you have the skills you have.

I designed several massively parameterised and rules based things, with live feeds, in that former life and am completely used to getting not just 0s and 1s, but getting 10 different results from 10 runs on what seems very similar. Who better than coders knows how the number of outcomes raises by magnitudes with a handful of extra variables?

A dev background is a good thing, so long as you realise the system you're looking at is more like the weather than a sausage factory. The inputs are ranges and the outputs are degrees of certainty rather than actual assertions. If the computer says yes or no it's wrong, the answer is fuzzy.

All knowledge is good knowledge. Nobody ever wished they had fewer weapons to choose from in a battle, but you do need to discard what's not appropriate.
xitian
Posts: 457
Joined: Fri Jul 08, 2011 2:08 pm

I have about 16 years experience in computing as well, along with 8 years sports trading. I was lucky enough to be profitable almost right from the start though.

My advice would be to not think about it as trading. I don't try and predict price movements. I think about it from a market efficiency point of view. So I try to answer the question "when would I expect the market to be incorrect?" or "how can I estimate a more accurate price than the market?". Then I just do the ground work of gathering data and verifying it in the simplest way I can. If I can find an answer to the above questions, then the trading part just happens itself because more often than not the market will return to where it should be. If it doesn't, then it means the market is still more accurate than you for the slice you've chosen to look at (or the inefficiency runs into in-play territory).

Use your computing advantages - and that's storing and looking at large amounts of data, being able to wait patiently for opportunities, doing complex calculations on the fly, reacting faster than manually, and not having to risk much to research ideas. I'll often spend hours, days, weeks doing ground work. Coding the trading system and putting money into it is probably less than half my effort.

mhorro, it sounds like you're logging the sorts of data I'm talking about, but at the same time it sounds like you're doing it fairly aimlessly. I'd personally store the tick data as is to database. Then later you can run programs to spit out whatever data or slices you're interested in investigating. I come back to my raw data stash time and again, and it's useful to have it going back for months. You have the skills to collect and store it, so you should. Otherwise you're going to be paying £200 for a month of data from Betfair.
northbound wrote:
Mon Apr 16, 2018 9:23 pm
accepting that you'll be right some times and wrong others
Yup, spot on, northbound. Be right more often than you're not adjusted for the amount you win/lose on each attempt. It can be something quite small that's enough to tip the balance in your favour though. But it's best to strive for something which generates many "signals" or opportunities each day. I'd personally prefer a strategy where I can have 1000 rolls of the dice where I think I have a 1% edge than something with 10 rolls where I think I have 5% edge. I'll know quite quickly in the first case if I'm wrong or not, but not in the second. It's not unrealistic to find ideas with those sorts of daily instances either. There's around 300 horses running in the UK each day, maybe 1000 in AUS, or 700 odd greyhounds, or 50-200 football games. I won't say exactly how many bets I place a day, but it's more than 10,000 (spread across many different strategies). If I only have a tiny edge that's enough to make a lot of consistent money with so many rolls of the dice. Develop a strategy with a tiny edge, and be a casino essentially. A casino is naff if it's only got 2 customers a day.
mhorro wrote:
Mon Apr 16, 2018 9:13 pm
Is gambling such a bad thing? It seems easier!!!!!!
If you have a lot of bets per day like I mentioned above, then "gambling" isn't so much of an issue. If you're beginning, then I'd still advise thinking of strategies where you close your positions before the off, as you probably won't want to deal with the big ups and downs of returns that comes with taking the extra risk in gambling. But in the end trading is just gambling except that you have one extra trade to close a position before the off anyway.

Hope some of that helps.
User avatar
Derek27
Posts: 23471
Joined: Wed Aug 30, 2017 11:44 am
Location: UK

mhorro wrote:
Mon Apr 16, 2018 9:13 pm
I am a computer professional with 25+ years in computer science.

I have spent 1 month logging signals in the UK horse racing markets i.e. horses hitting a certain prices, number of selections, fav price pre-off the usual variables.

In a nutshell we are dealing with animals ( horses ) and ( human or automation ) to trade the markets. I have not even factored in delays for in-play on transmissions.

So we have no logic here just a P&L over a period of time!

It is very frustrating for someone who has worked in in the computer industry to try and make sense of this.

Is gambling such a bad thing? It seems easier!!!!!!


I will not mention training mode as it only seems to test functional flows!

Thanks,

Mark
I've been gambling on horse racing since the age of eight, and got my first job as a computer programmer at the age of 17.

I think your problem Mark, is the way you relate the two. When I started programming in IBM COBOL (for financial and insurance firms) I didn't try to apply my experience in horse racing to the financial programs I was writing because it was irrelevant.

I recall from a previous thread that you had designed a software/database application to log information of every race and produce the best bets of the day based on form or the information your database has. That sort of approach is futile as it requires knowledge of horse racing.

Having IT skills can be immensely useful applied to trading but it's the core of your strategy has to be sound. Simply analysing data looking for trends can be like roaming the streets hoping to find money!
User avatar
ShaunWhite
Posts: 9731
Joined: Sat Sep 03, 2016 3:42 am

Quick tot up and we're already up to 100 yrs of IT experience and God knows how many trading and looking for angles. If we weren't all doing similar things then between us we'd have a hell of a toolbox, instead of an ill-conceived tangle of aborted ideas and half finished projects... Well that's what my folders look like anyway.

Cursing myself now for not building a proper db a long time ago, data just in excel and incomplete to make it manageable. If I'd known I'd still be here I'd have invested the time. My db skills are pretty rusty to be frank. It's a long long time since I did anything from scratch and I'm clueless about the new api stuff too :oops: Being in IT a long time isn't sometimes as useful as being fairly new to it.

Jeez I felt quite handy until I thought about it. :cry:

I did the first IT lessons our school did, the IT room had one of these...ONE, for about 15 boys. With optional punch card reader and 1200baud acoustic coupler...the actual computer was 20mls away. (1200baud is 0.0012Mbps for our younger readers, over 40,000 times slower than our 50Mbps!). Year 2 they got 3 PETs, mindblowing, it had a screen!

People moan about the internet but it's f'kin science fiction made real for some of us who've waited 20 minutes for a typewriter to plod through "Hello ShjuN" ....D'oh, hand me another punched card !
220px-Teletype_with_papertape_punch_and_reader.jpg
You do not have the required permissions to view the files attached to this post.
Last edited by ShaunWhite on Tue Apr 17, 2018 2:07 am, edited 1 time in total.
User avatar
Derek27
Posts: 23471
Joined: Wed Aug 30, 2017 11:44 am
Location: UK

ShaunWhite wrote:
Mon Apr 16, 2018 10:57 pm
It's 37 years since I wrote my first code.
Having thought about it, it's exactly that long ago when I wrote my first program - on the ZX Spectrum. :)
User avatar
ShaunWhite
Posts: 9731
Joined: Sat Sep 03, 2016 3:42 am

Derek27 wrote:
Tue Apr 17, 2018 2:07 am
ShaunWhite wrote:
Mon Apr 16, 2018 10:57 pm
It's 37 years since I wrote my first code.
Having thought about it, it's exactly that long ago when I wrote my first program - on the ZX Spectrum. :)
ZX81 got me into my first job really.
The kit was £79 vs £99 for one already made so dad got his soldering iron out. Amazed it worked, it was more braised than soldered really, you heated the iron up on the cooker.

..ahh jumpers for goal posts eh Derek. :)

I bought an unopened ZX81 kit about 10yrs ago (a guy in the states bought all the old stock) and it's in the loft waiting for Antiques Roadshow 2050.
User avatar
ShaunWhite
Posts: 9731
Joined: Sat Sep 03, 2016 3:42 am

I just found his site http://www.zebrasystems.com/nyc/index.p ... x-sinclair...i think I paid about £100 for mine 10 yrs ago.
I should have bought all he had! I seem to remember he had about 150.
User avatar
Derek27
Posts: 23471
Joined: Wed Aug 30, 2017 11:44 am
Location: UK

I've still got my old BBC Micro B (the 'B' means 32K). I hope that will be worth something sometime - it's not much use now obviously but it's one of those things you just can't throw away.
User avatar
jimibt
Posts: 3641
Joined: Mon Nov 30, 2015 6:42 pm
Location: Narnia

shaun -make that 125 years then :D

started out modifying sequencers (passport designs midi sequencer), then graduated onto paid jobs doing vba in the late 90's, followed on into the 2000's with .net. still doing it to this day but have been sidelined into the brave new world of AI, but not as a coal face developer, more a process integrator. as part of my current remit, saw this and thought it would probably be a great fit for the fuzzy world of sports trading: https://sdtimes.com/ai/dimensional-mech ... -language/

anyway, have found for sure that my numerous .net attempts at capturing the market in my piston/bouncy castle all work to an extent, but really there is a divine variable sitting above it all that is not easily captured in traditional programming approaches. AI/ML may be the best way fwd imho...

right, talking of soldering irons, used to also take old drum machines to bits and reconfigure them by patching across circuit boards with jump wires (the hihats on the TR 606 were too clunky, so a little wire bridge to another part of the circuit board thinned them out nicely!!) - wild experimentation over risk is the name of the game!! :D
PeterLe
Posts: 3715
Joined: Wed Apr 15, 2009 3:19 pm

xitian wrote:
Mon Apr 16, 2018 11:05 pm
I have about 16 years experience in computing as well, along with 8 years sports trading. I was lucky enough to be profitable almost right from the start though.....
Nice post James, some jewels in there
spreadbetting
Posts: 3140
Joined: Sun Jan 31, 2010 8:06 pm

Maybe too many
User avatar
Euler
Posts: 24700
Joined: Wed Nov 10, 2010 1:39 pm
Location: Bet Angel HQ

I started out as a Z80 /6502 coder. The first thing I ever did was model football matches while I was a teenager. I used this to collect a ton of dividends on the football pools.

I've always found that modelling a market is the best way to go if you can describe something then you can anticipate it.
mhorro
Posts: 112
Joined: Sun Feb 17, 2013 1:16 pm
Location: Cheshire

Many thanks for your comments, this is just my view and I am probably wrong.

But I will be returning to gambling as I cannot make the transition.

Thanks,

Mark
Post Reply

Return to “Trading Horse racing”