another DO question

[js190]

I have a strange result where for an account size of 150k running a backtest from rob_system/run_system.py and looking at the DO output, I have some positions BBCOMM long, shorts in RUSSELL GOLD_mico, KOSDAQ
but running the same config in prod through the runSystemCarryTrendDynamic I get no positions when doing the update_strategy_orders, I would have expected it to want to open similar given 0 starting positions.
Could it be that the forecasts are not strong enough to open new? But the DO backtest had positions due to the forecasts not being weakenough to reduce?

[robcarver17]

Do you have any orders at all eg in other markets?

[robcarver17]

How does the tracking error compare to the buffer?

[js190]

ah wow it's sensitive if i change to tracking_error_buffer: 0.01 I get a few tentative positions.

Don't panic Mr Mannering!

[TobiasAntiGravity]

I'm assuming this happens because of the size difference in the notional trading capital?

[js190]

I'm not sure I follow what you mean?
By lowering the buffer I get trades, so I think that's job done.

[js190]

@TobiasAntiGravity are you now live?

[TobiasAntiGravity]

@js190 - My apologies - no context at all to that question - I am on vacation - and working on getting the system live - got it live now as a matter of fact. Have 3 orders executed - but nothing is getting filled so having to solve it manually.

My question was - why did you have to lower the buffer error , compared to the level that Mr. Carver uses in his system. I was wondering whether or not the difference in notional_trading_capital was the difference.