I have an idea I want to run buy you

[John Verbrugge]

Hello traders,

I have an idea I want to get validation on before I start working on it.

My goal is to get buy and sell signals for BITX, a 2X Bitcoin ETF.

BITX has daily resets, so I don’t want to hold it over a day.  Clearly this is a short term strategy.

Idea: 

• Build a strategy based on the Bitcoin 15-minute bars or 1-hour bars
• Strategy must have a high win-to-loss ratio
• The trade must close by end of day
• Trades should be short term, so 4 to 6 bars for one-hour, or 10 to 20 bars for 15-minute bars
• Average bar loss < ?

I will use the strategy as a signal, so when it tells me Bitcoin is a BUY, I will buy the BITX.

I am using this newsletter's ideas to guide this idea: https://www.adaptrade.com/Newsletter/NL-ProfitStrategy.htm

Has anyone tried this kind of thing yet with Builder?  Am I missing some metric or constraint?

Any ideas or suggestions would be helpful.  

Thank you!

JV

[MikeBryant]

Sounds like a great use of the program. My only concern would be the correspondence between BTC and BITX. If you're building the strategy based on BTC, you would want to make sure the strategy is not exploiting some feature of BTC that is not available in BITX. For example, if you're using market orders, do both symbols use the same closing and opening times? If not, it could affect the entries and exits.

[John Verbrugge]

Hi Mike, 

That's a good point.  BTC trades 24/7, and BITX is an ETF on one of the American exchanges.  So, to get around that, could I use the Trading Logic trade entry and exit after fields?  Would that be sufficient to ensure I only enter during market hours?

BITX has only been trading since June 2023, so there might not be enough historical data to use that as my data source; I think I need to stick with BTC.

Regarding my build metrics parameters, I think adding "ideal trade" as one of the metrics might be a good idea as well.  What do you think?

Thanks,

John

[MikeBryant]

Sure. You could use the entry and exit time options to restrict the times to the times that BITX trades. Another option would be to change the session times on BTC on the Build Symbols table via the Symbol Properties window; i.e., change the session times on Symbol Properties AFTER adding the symbol to Build Symbols.

Ideal Trades as a metric is always available but can be difficult to use effectively.

[John Verbrugge]

I will try changing the session times on BTC on the Build Symbols table first.  I like that better.  If it gives me trouble, I'll try it the other way in the Trading logic.   

Stay tuned... I will have some results to show in a few days. 

Thank you!

[John Verbrugge]

Here's what I came up with:

Build Objectives
Minimize Complexity, weight 0.700
Maximize Net Profit, weight 1.000
Maximize Corr Coeff, weight 1.000
Maximize Trade Sig, weight 1.000

Build conditions
Drawdown <= S7,000.00 (Train)
Win/Loss Ratio > = 1.500 (Train)
Ave Trade Length between 15 and 60 (Train)
Corr Coeff >= 0.9500 (Train)
Trade Sig >= 95.00% (Train)

This one had 99 trades, so it traded infrequently (on 15 the minute bars).  I'm not thrilled with it, but good enough for testing the idea. 

Net profit:  $43,325

Drawdown: $3,918

To see if this was an equity doubler, I took the trade list and calculated 2x Profit/Loss.  So my equity formula was as follows:

NewEquity = NormalEquity - ProfitLoss + DoubledProfitLoss

Here's a few lines from the Excel sheet:

Do I have that right?  At the end of the trade list, my "equity with 2x" is almost exactly the same as the normal equity.  Is that because my wins and losses are both doubled?  So, there's no point in trading this strategy?

What build metric could I change to make this work?  

Thank you!

[MikeBryant]

Have you considered changing the symbol properties to see the 2x results in Builder; e.g., doubling the point value?

[Nick]

Your performance (in green) isn't that great. Thats why you aren't seeing much difference between the 2x trading.  There seems to be a lot of curve fitting in your strategy given the majority of profit is on the training data.

Nick

--
You received this message because you are subscribed to the Google Groups "Adaptrade Builder" group.
To unsubscribe from this group and stop receiving emails from it, send an email to adaptrade-buil...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/adaptrade-builder/138ce9af-2120-4791-b810-04bd4eb89c54n%40googlegroups.com.

[John Verbrugge]

Mike,

That sounds like a great idea to try.   Again, this is something that had never occurred to me, so I really appreciate the "inside information" on this program.  Thank you. 

Best,

John

[John Verbrugge]

Hi Nick,

Yes, I agree it's not great in the validation phase.  I am working on doing another strategy that gives me better numbers for the testing and validation phases.

Thanks,

John