BOT testing services ?

[decomez6]

hi all,

Data mining is great provided one can build and run a reliable data base.

Q: for future refrence, does anyone offer BOT testing services for beginers with limited knowldge on data mining?

[Big Bad Barney]

What do you mean by bot testing? The only way you are going to test bots is in the same environment they are developed in...

What are your bots written with?

I think you'll find, you write it all yourself or you can't have it.

[decomez6]

What do you mean by bot testing? The only way you are going to test bots is in the same environment they are developed in...

What are your bots written with?

I think you'll find, you write it all yourself or you can't have it.

Bet angel guardian bots.
It’s possible to have a bot that shows short term
Positive results but crappy long term expectancy.
It’s great to have a database that helps in modelling but until then I trade what I see and develop a bot depending on the current profitability levels.
Sometimes I abandon ideas that could possibly develop a positive expectancy if taken through a vigorous long term test.
It would be great to have a place where you take a bot and the doctor says,‘ ‘your wasting your time mate, your bot is only good in Christmas but useless all year round!’ This are the results and invoice for the job I have done for you .

[Big Bad Barney]

Bet angel guardian bots.
It’s possible to have a bot that shows short term
Positive results but crappy long term expectancy.
It’s great to have a database that helps in modelling but until then I trade what I see and develop a bot depending on the current profitability levels.
Sometimes I abandon ideas that could possibly develop a positive expectancy if taken through a vigorous long term test.
It would be great to have a place where you take a bot and the doctor says,‘ ‘your wasting your time mate, your bot is only good in Christmas but useless all year round!’ This are the results and invoice for the job I have done for you .

Sort of like an accountant for a small business?

I don't know if you could ever waste your time. I'm inclined to think you learn from making dodgy bots. An accountant could always tell you you're wasting your time opening that pizza plaza, but how you gonna learn it was a bad idea? cause the good idea is often based on knowing a bad idea.

I believe there's a saying around that goes something like 'An expert is someone who has made every mistake conceivable'

I think the trick is somehow to make mistakes, and learn from them quickly. Which is not so easy

...but having said that, I don't have the answers....but BetAngel doesnt have any mechanism for feeding in historical data (that I know of), so your only option is in testmode/live.... not really sure why they haven't implemented such things...

I read in a marketing book somewhere once, you only have to be slightly better than your competitor in order to win

[sionascaig]

I read in a marketing book somewhere once, you only have to be slightly better than your competitor in order to win

Which in todays world usually means paying less wages.

[decomez6]

Bet angel guardian bots.
It’s possible to have a bot that shows short term
Positive results but crappy long term expectancy.
It’s great to have a database that helps in modelling but until then I trade what I see and develop a bot depending on the current profitability levels.
Sometimes I abandon ideas that could possibly develop a positive expectancy if taken through a vigorous long term test.
It would be great to have a place where you take a bot and the doctor says,‘ ‘your wasting your time mate, your bot is only good in Christmas but useless all year round!’ This are the results and invoice for the job I have done for you .

Sort of like an accountant for a small business?

I don't know if you could ever waste your time. I'm inclined to think you learn from making dodgy bots. An accountant could always tell you you're wasting your time opening that pizza plaza, but how you gonna learn it was a bad idea? cause the good idea is often based on knowing a bad idea.

I believe there's a saying around that goes something like 'An expert is someone who has made every mistake conceivable'

I think the trick is somehow to make mistakes, and learn from them quickly. Which is not so easy

...but having said that, I don't have the answers....but BetAngel doesnt have any mechanism for feeding in historical data (that I know of), so your only option is in testmode/live.... not really sure why they haven't implemented such things...

I read in a marketing book somewhere once, you only have to be slightly better than your competitor in order to win

yep you got it ! or m.O.T garage to test the' road 'worthiness of the Bot
its possible to mine data using excel, the underlaying buiding block of bet angel or purchase historical data from BF and use it to make a data base ,the only hurdle is the beginner's steep learning curve .
i am pretty sure some one is going ''put in the hours or look for something that matches your ability!''
i respect that, its not easy to be self sufficient and thats why some beginners would be willing to pay for such services.

[Big Bad Barney]

yep you got it ! or m.O.T garage to test the' road 'worthiness of the Bot
its possible to mine data using excel, the underlaying buiding block of bet angel or purchase historical data from BF and use it to make a data base ,the only hurdle is the beginner's steep learning curve .
i am pretty sure some one is going ''put in the hours or look for something that matches your ability!''
i respect that, its not easy to be self sufficient and thats why some beginners would be willing to pay for such services.

the problem you got is as soon as you cant feed historic data in, your bots can't be tested in any way, by anyone....other than the slow way...so even if one wanted to offer the service....can't.... it all hinges on what's in your head...

Building ones own back testing engine is not a trivial thing...even for people with some experience in software development.. depends how deep one wants to go and what one wants to achieve I suppose... and thats a whole nother different conversation that one

As for viability of such a service to exist, well, I dunno, I think the closest you would get is some kind of mentor ship service. I would doubt much is around due to an ROI thing and the mentality of noobs (It'd be a revolving door)..... like... try tell a teenager about the hardships of life....ya got no hope....

[Big Bad Barney]

I read in a marketing book somewhere once, you only have to be slightly better than your competitor in order to win

Which in todays world usually means paying less wages.

You could be right there...but everything in life has to be prefixed with... 'it depends' ... I suppose it's about keep your customer happy enough that they will pay you...but no more happier than they need to be

Hypothetically, lets say Betangel wanted to put backtesting facilities into the software. They've already thought about PIQ and half the stuff necessary to actually do it, so they got a leg up on most people.... Off the top of my head there might be a few reasons why it might be a bad idea for them.

- The software is already 'better' than the next competitor, so people will pay for it anyway.
- It costs money for features, more so in maintenance than development...
- Complexity decreases reliability...
- Complexity changes your target demographic...
- Betfair might not dig it...

[sionascaig]

I read in a marketing book somewhere once, you only have to be slightly better than your competitor in order to win

Which in todays world usually means paying less wages.

You could be right there...but everything in life has to be prefixed with... 'it depends' ... I suppose it's about keep your customer happy enough that they will pay you...but no more happier than they need to be

Hypothetically, lets say Betangel wanted to put backtesting facilities into the software. They've already thought about PIQ and half the stuff necessary to actually do it, so they got a leg up on most people.... Off the top of my head there might be a few reasons why it might be a bad idea for them.

- The software is already 'better' than the next competitor, so people will pay for it anyway.
- It costs money for features, more so in maintenance than development...
- Complexity decreases reliability...
- Complexity changes your target demographic...
- Betfair might not dig it...

I think this was covered in another thread (not the points you make above)
just cant remember which one...

- bf data costs
- licencing requirements

I think came up and suspect it would be a fairly expensive exercise with ongoing maintenance to produce a robust tool.

If you don't want to pay for bf data (from bf) there are tools on the forum that allow you to capture for modelling purposes - not used them though so can't really say what's possible...

Think I can remember a NigelK spreadsheet that could capture data second by second and allow you to rerun a market...

[the4droogs]

the problem you got is as soon as you cant feed historic data in, your bots can't be tested in any way, by anyone....other than the slow way...so even if one wanted to offer the service....can't.... it all hinges on what's in your head...

Building ones own back testing engine is not a trivial thing...even for people with some experience in software development.. depends how deep one wants to go and what one wants to achieve I suppose... and thats a whole nother different conversation that one

As for viability of such a service to exist, well, I dunno, I think the closest you would get is some kind of mentor ship service. I would doubt much is around due to an ROI thing and the mentality of noobs (It'd be a revolving door)..... like... try tell a teenager about the hardships of life....ya got no hope....

Sounds like what's needed is this feature I posted in the suggestions area

[decomez6]

yep you got it ! or m.O.T garage to test the' road 'worthiness of the Bot
its possible to mine data using excel, the underlaying buiding block of bet angel or purchase historical data from BF and use it to make a data base ,the only hurdle is the beginner's steep learning curve .
i am pretty sure some one is going ''put in the hours or look for something that matches your ability!''
i respect that, its not easy to be self sufficient and thats why some beginners would be willing to pay for such services.

the problem you got is as soon as you cant feed historic data in, your bots can't be tested in any way, by anyone....other than the slow way...so even if one wanted to offer the service....can't.... it all hinges on what's in your head...

Building ones own back testing engine is not a trivial thing...even for people with some experience in software development.. depends how deep one wants to go and what one wants to achieve I suppose... and thats a whole nother different conversation that one

As for viability of such a service to exist, well, I dunno, I think the closest you would get is some kind of mentor ship service. I would doubt much is around due to an ROI thing and the mentality of noobs (It'd be a revolving door)..... like... try tell a teenager about the hardships of life....ya got no hope....

There are a couple of excel sheets In the forum that will use betfair / betangel live feeds Record P/L and project it on a graph. I think the issue is a private data base which is run and maintained by the user to develop and model ideas which are then converted into bots Which are then tested and proven to be profitable or not.
Most beginners wouldn’t stumble upon a winning streak and fall into the gamblers fallacy and end up being fooled by randomness ,the service would help mitigate the problem of dwelling on loosing strategies which are demoralising, take and waste precious time.
The more you get into trading , the more you realise it’s a very one man job, you are the CEo,the accountant, the it department, the think tank, the research, and finally the execution department..
Having so many Balls floating In the air is quite overwhelming for a serious beginner and is often the case He will drop while putting too much effort on another.
Most people are in willing to share thier Market edge but it’s only an edge if only it’s proven and tested. That’s where the service would plug in the gap between delegation and shared profitability. It’s a win win , for the service provider, he gets to see other traders edges and partake in the profit. There is nothing stopping him from using the traders ideas for his own benefit and the trader is more than willing to share his ideas , it’s better to make less profit than no profit.

[decomez6]

Building ones own back testing engine is not a trivial thing...even for people with some experience in software development.. depends how deep one wants to go and what one wants to achieve I suppose... and thats a whole nother different conversation that one

Sounds like what's needed is this feature I posted in the suggestions area

That’s great suggestion and I have also checked on other similar threads. It’s clear BA is a gateway to BF and is designed to be an execution tool . Changing the BF live-streams into saved streams Pricy hence commercially Unviable and Will also require a change of terms and conditions between BA and BF.
What I keep asking myself?
The markets are efficient and when traded randomly The outcomes are random results that will eventually break even minus commission. So in effect if one was trade all markets in all entirety you will only loose the commission imposed on winning trades.
That being the case it’s easier to simulate the market using artificially computer generated Data. It doesn’t have to be a horse racing data, it just need to mimic an efficient market with efficient overruns.
May be i am missing the devil in the details, so I stay corrected.