Seetu for Amibroker Users

[RowanC]

This is hopefully something useful for Amibroker users that I stumbled on recently.

Collective2 is a site that tracks individuals trading strategies. It's been around along time and is reasonably well known.

Recently they have launched their own backtest engine called Seetu.

The interesting thing for Amibroker users is that it uses AFL. So anyone familiar with AFL will be up and running immediately. And it's free to use.

It's a bit like Quantopian and their Python/Zipline setup for those familiar with that.

While Seetu is still in development you can currently backtest US stocks and forex although I believe the data is only daily at the moment. However as I understand it, they will soon be offering minute data and a wider range of products including futures. C2 does track ASX stocks so hopefully they will incorporate them into the backtester also.

So my thinking is that it might be a useful way to roll out a strategy in real time to test, or for those who don't have data they can do simple backtests. I still think Amibroker with Premium Data is far better, but for testing strategies in real time by just copying and pasting your code it could be pretty handy.

Certainly worth a look for those interested.

 

[Gringotts Bank]

This is hopefully something useful for Amibroker users that I stumbled on recently.

Collective2 is a site that tracks individuals trading strategies. It's been around along time and is reasonably well known.

Recently they have launched their own backtest engine called Seetu.

The interesting thing for Amibroker users is that it uses AFL. So anyone familiar with AFL will be up and running immediately. And it's free to use.

It's a bit like Quantopian and their Python/Zipline setup for those familiar with that.

While Seetu is still in development you can currently backtest US stocks and forex although I believe the data is only daily at the moment. However as I understand it, they will soon be offering minute data and a wider range of products including futures. C2 does track ASX stocks so hopefully they will incorporate them into the backtester also.

So my thinking is that it might be a useful way to roll out a strategy in real time to test, or for those who don't have data they can do simple backtests. I still think Amibroker with Premium Data is far better, but for testing strategies in real time by just copying and pasting your code it could be pretty handy.

Certainly worth a look for those interested.

Good shares are rare. Thanks for this.

 

[Gringotts Bank]

Uses AFL or something C-like? I'm seeing something a bit different.

 

[RowanC]

Uses AFL or something C-like? I'm seeing something a bit different.

It definitely uses AFL. I've tested a few simple strategies that I've literally just copy and pasted in.

However there might be a few things that vary in terms of setup options but overall is virtually the same from my experience.

Here's the language guide.

https://collective2.com/seetudoc/index.html

 

[Wysiwyg]

Great idea to collect a large data base of trade strategies. Win/win at 70/30 though.

 

[RowanC]

Great idea to collect a large data base of trade strategies. Win/win at 70/30 though.

Overall I think it has the potential to be a pretty good setup for developers.

I personally really like the Quantopian structure too.

 

[Gringotts Bank]

It definitely uses AFL. I've tested a few simple strategies that I've literally just copy and pasted in.

However there might be a few things that vary in terms of setup options but overall is virtually the same from my experience.

Here's the language guide.

https://collective2.com/seetudoc/index.html

Can you show me a screeny of where you pasted your code please? I can't find the place to do that.

 

[howardbandy]

Greetings --

Some of the examples have SetTradeDelays set to values other than 0. Be very careful when analyzing, testing, and generating signals when the trade delay is anything other than 0. It is easy to get confused. And easy to create systems that look good on historical data but are not tradable.

Best, Howard

 

[RowanC]

Can you show me a screeny of where you pasted your code please? I can't find the place to do that.

So this is under the develop - seetu tab. Code in the strategy tab, choose products in the target tab and the results are in the statistics/chart etc

 

[RowanC]

Greetings --

Some of the examples have SetTradeDelays set to values other than 0. Be very careful when analyzing, testing, and generating signals when the trade delay is anything other than 0. It is easy to get confused. And easy to create systems that look good on historical data but are not tradable.

Best, Howard

Hi Howard

I'd also say another one of the big issues would be the fact that the test universe of stocks is based on current constituents of the key indexes. Ideally they would need to offer perhaps the ability to filter stocks by market cap.

That's why Premium Data are so good IMO.

 

[Gringotts Bank]

So this is under the develop - seetu tab. Code in the strategy tab, choose products in the target tab and the results are in the statistics/chart etc

View attachment 68393

Thanks again.

 

[Gringotts Bank]

It's good to see the top developers can make money. But compared to the top Youtubers - who earn in the $millions pa. for making videos - the income is quite small.

Howard, do you sell anything on C2?

 

[howardbandy]

It's good to see the top developers can make money. But compared to the top Youtubers - who earn in the $millions pa. for making videos - the income is quite small.

Howard, do you sell anything on C2?

No, I do not have any systems running on C2. I registered there, and may post something someday, but nothing so far, and nothing now.

Best, Howard

 

[howardbandy]

Hi Howard

I'd also say another one of the big issues would be the fact that the test universe of stocks is based on current constituents of the key indexes. Ideally they would need to offer perhaps the ability to filter stocks by market cap.

That's why Premium Data are so good IMO.

I'll second the value of Norgate Premium Data. They are very high quality at a very reasonable price.

Survivorship / membership / selection bias is serious. Doing it wrong results in serious underestimate of risk and overestimate of profit potential. Diffusion indexes work well, but rely on accurate and stable selection of the issues in the group. Norgate's membership information related to delisted issues and historical membership is outstanding.

Best, Howard