As many of you know, building a bot in Bet Angel's automation feature is a dream come true. With no coding knowledge or experience, you can implement your favourite strategies in a surprisingly short amount of time using an intuitive and powerful UI. Building the bot is fantastic. Testing the bot, on the other hand, is a nightmare. Without the power of unit testing that you would find in a traditional coding environment, the only way to test your bot is to point it at a real market and hope the conditions you are trying to test eventuate. With a bot containing a lot of edge cases and exit strategies, this can be extremely time consuming, not to mention boring as you stare at the screen willing a particular market condition you are testing to occur.
To fill this gap I propose the concept of Mock Markets. The user would set parameters of the mock market, from volume to liquidity to the time of bets being placed. The user would then press play on the market and the parameters he/she set would be used to mock the api calls and the market would play out exactly as the user specified. They would then be able to point a bot at the market and see how it reacts under a controlled set of conditions. You would even be able to flick between the one-click, ladder screens etc.
I understand that this would be a very big job, but I think it could add a huge amount of value to the BetAngel product. It's potential application would go far beyond just bot testing too. Imagine you are on the BetAngel Academy listening to Peter's dulcet tones. Then at the end of the video, he gives you some Mock Markets to download. These Mock Markets are designed to simulate different facets of whatever lesson it is that the video was teaching, and the student could test themselves against that market and review their actions in line with an explanation of what they should have done.
I'm sure people could think up even more powerful uses of such a feature. I'd love to know what you all think.
Mock Markets
-
goldfinch77
- Posts: 199
- Joined: Wed Aug 05, 2020 11:45 am
I totally agree, as a relative newbie I am struggling to A: learning how to drive the bot and B: waiting to see if i got it right.
So I would be very much in favour.
Goldfinch
So I would be very much in favour.
Goldfinch
Without mentioning any names, AN Other product offers this as a time travel functionality. Rather than "mock markets", previous markets history data is available to download at cost and you can run your automation against those in sped up/slowed down time - i.e 1000's of markets a day.
Obviously past performance is no guide to profitability but to be able to test bot rules against real market data would be invaluable and I'd happily pay decent money for it in BA
Obviously past performance is no guide to profitability but to be able to test bot rules against real market data would be invaluable and I'd happily pay decent money for it in BA
i actually proposed this about 4 years ago and it was explained in detail the reasons why it would not be economically viable...
viewtopic.php?f=20&t=11983
as you can see, it did have good support.
viewtopic.php?f=20&t=11983
as you can see, it did have good support.
Shame.jimibt wrote: ↑Fri Nov 20, 2020 11:30 ami actually proposed this about 4 years ago and it was explained in detail the reasons why it would not be economically viable...
viewtopic.php?f=20&t=11983
as you can see, it did have good support.
Interesting how another product can provide it though. Either they are bypassing something at BF or they have found a way to make it viable, at a reasonable cost to the user.
Again, it would be an invaluable way to test bot functionality against a real market so could see the uptake being quite decent