Careful with automation

[thebarra1]

Last night after fair bit of practise I decided to automate the Real Madrid v real betis game on lay the draw if profitable automation.
There was no limit put on the lay price therefore it wiped out my account.
Near the end of the game about 85 mins Madrid had all the pressure but didn't look like scoring so I green up to save some of my money.
Then in injury time betis scored a winner.
My point is be careful. Look at the odds first for laying.
Also is it possible to insert a condition limiting the amount of odds i.e.lay no more tha say 5.0

Thebarra

[ShaunWhite]

Try using a Fixed Odds condition on the selection's lay price?

Your admission about your mistake (and lots of people wouldn't admit to it) should be a clear reminder to everyone to test, test & test again.

Writing rules is no different to writing software, it's just a series of instructions, so it's essential to consider every possible set of circumstances, however unlikely and in various combinations. Test each scenario, first on paper (as live data won't necessarily present the strange 1 in a 1000 situations), then in test mode, then with very small stakes. Will it cope where no money is being offered? where the overound is huge? where there's insufficient volume to exit? when there's a suspension and a reformed market? where there's a penny at 1.1 and the next money is at 1000, etc etc.

Writing software (aka rules) that work is easy while the world behaves in an orderly logical way, the difficulty lies in coping with the real world and all of its quirks and foibles. If nothing else I hope it gives people who struggle with 3 lines in a Guardian rule a little more sympathy for coders who create entire complex systems and then let them loose on 100s or even millions of users.

[ruthlessimon]

Writing rules is no different to writing software, it's just a series of instructions, so it's essential to consider every possible set of circumstances, however unlikely and in various combinations. Test each scenario, first on paper (as live data won't necessarily present the strange 1 in a 1000 situations), then in test mode, then with very small stakes.users.

Exactly, 100% agree. Why was fiure given different advice? I really hope barra didn't lose 4 figs learning that

[ShaunWhite]

Writing rules is no different to writing software, it's just a series of instructions, so it's essential to consider every possible set of circumstances, however unlikely and in various combinations. Test each scenario, first on paper (as live data won't necessarily present the strange 1 in a 1000 situations), then in test mode, then with very small stakes.users.

Exactly, 100% agree. Why was fiure given different advice? I really hope barra didn't lose 4 figs learning that

Trouble is Si that a lot of people (and not suggesting barra is one of them) think that's all too much effort, combine that with an over inflated belief in their own abilities...and BANG! A good coder doesn't earn 600 a day + for nothing. My bro-in-law is in the job I left 10yrs ago and is now pulling in £750! Mind you when one good idea can save/earn a company millions it's a modest rate compared to the guys in the front office.

[ruthlessimon]

Trouble is Si that a lot of people (and not suggesting barra is one of them) think that's all too much effort, combine that with an over inflated belief in their own abilities...and BANG! A good coder doesn't earn 600 a day + for nothing. My bro-in-law is in the job I left 10yrs ago and is now pulling in £750! Mind you when one good idea can save/earn a company millions it's a modest rate compared to the guys in the front office.

Certainly does take a lot of work to get through all the development stages. But had this been tested on the sim, the fundamental error in the strategy would have been picked up - because he'd have blown a demo bank - & acted upon the mistake, no harm done.

I can understand if the simulated results had been extremely good week on week, then going live - on a consistent stake. But it just felt to me reading the post, he just went live immediately, applying something brand new to the strategy (automating it) & unfortunately missed something vital.