Indicator suggestion

[Charunks]

Hi guys!

I was hoping someone would have a suggestion to help me out. I am currently trying to configure gunbot to buy intradaily for a small profit based on cycles theory by Walter Bressert. So far I use MACD as a buying strategy, and a tight TSSL as a sell with a long stochastic RSI crossover as a 'upwards volatility indicator. For this to work it is essential that tssl can do its job without other indicators interfering with the selling process.

Im looking for an indicator that tells gunbot to buy at the upcycle; so it buys into the strength of momentum.
Normally a stochastic can do that perfectly, however gunbot only looks down from the buylevel. I've tried a 40 periods RSI  with a cross up of 50 (cycle up RSI +50, cycle down RSI -50). Allthough this  "seems" to work, I get the feeling gunbot is actually looking for the crossover and not as much to the fact that the RSI is higher than the buy level (= exactly how the wiki describes it).

I've been theorizing over EMA, MFI and ADX.  All great indicators, however... EMA brings nothing new to the table that MACD doesn't do, MFI only seeks opportunities below the buy level (and doesn't consider cycle information -> bad sync.). ADX comes with a sell condition as far as I can see, which is counterproductive.

I was hoping that maybe someone might have a suggestion that could complement my strategy to increase the probability to buy into the strength of a cycle.

Thanks!

regards.

[Aitor]

Hi, in my opinion, better than theorizing, you can directly check how the buy and sell points are affected by the strategy changes. You can get the Backtesting Addon here:

https://gunbot.shop/product/backtesting-addon-for-gunbot/

So you can easily tweak you strategy until the buying points behave like you want to.

[Charunks]

Hey m8, thank you for the tip  .

I would not use the backtesting addon. It has many biases and when comparing results from live trading action from the bot with the tool afterwards its not even close.

I'll study and experiment some more and see where it takes me .

[Aitor]

Hey m8, thank you for the tip  .

I would not use the backtesting addon. It has many biases and when comparing results from live trading action from the bot with the tool afterwards its not even close.

I'll study and experiment some more and see where it takes me .

It's should be very very close. If it doesn't, you are missing something. Maybe you are running the backtesting with different candle types or maybe the delays are not the same. Note that backtesting process the data once per candle, when the candle is closed. In other words, if you run a 5 mins candle, the data will be processed every 300 seconds. So if you run the bot and put a 10 seconds delay, the bot would be much more reactive than you can see in the backtesting, and the results would be different.