Learn to code and use trading bots like me : codealgotradin... Get A Free Trading Algo 👨💻: codingtips.jac... My Trading Brokerage : www.interactiv... My Instagram : / jacobdamaral
10 minutes is pushing the minimum for trending systems. General rule is < 15 minute bars useful for scalp and mean reversion trading. >15 minutes for trend trading.
Nice video because Walk-Forward is CRITICAL to testing testing stability BUT I diverge with you on using NEGATIVE WF settings. In my testing negative WF results will usually return >50% chance negative results (seems like WF is >50% predictive of minus results too). I'm running cluster so I pick the most positive 30-day or 60-day result. Or I'd rerun the sim hoping for a current POSITIVE settings.
Hey, just came over from Kevin's channel, great content!! Been developing on NT for over a year now. How many iterations do you like to stay under for a strategy not including time series optimization?
I’m kinda confuse, if you are using the latest parameter settings, what’s the point of doing the previous stages of optimisation as you basically not doing any comparison nor any picking of the previous parameter settings? Am I missing something
Great question James while we do use the latest iteration for running live, previous walk forward iterations are important to find edge in a good strategy.
Can you apply the same thinking for walk forward optimization by using say spy for the backtest and all other tickers to forward test it? also when you walk forward and optimize it every tear that you have traded aren't you slowly optimizing for the latest data? thanks for the video.
Great video. Sure makes more sense than the past methods I used lol. Now I realize why my back testing results were never accurate.
Glad I could help!
10 minutes is pushing the minimum for trending systems. General rule is < 15 minute bars useful for scalp and mean reversion trading. >15 minutes for trend trading.
Nice video because Walk-Forward is CRITICAL to testing testing stability BUT I diverge with you on using NEGATIVE WF settings. In my testing negative WF results will usually return >50% chance negative results (seems like WF is >50% predictive of minus results too). I'm running cluster so I pick the most positive 30-day or 60-day result. Or I'd rerun the sim hoping for a current POSITIVE settings.
Nice video, thank you. Please also make Monte Carlo analysis for robustness too
Hey, just came over from Kevin's channel, great content!! Been developing on NT for over a year now. How many iterations do you like to stay under for a strategy not including time series optimization?
Hey Jonathan, welcome! Kevin is a great guy. I like to stay under 500 iterations, usually if you're using less than 5 parameters it's quite easy.
I’m kinda confuse, if you are using the latest parameter settings, what’s the point of doing the previous stages of optimisation as you basically not doing any comparison nor any picking of the previous parameter settings? Am I missing something
Great question James while we do use the latest iteration for running live, previous walk forward iterations are important to find edge in a good strategy.
So the most recent result, 2022 for this video, would be most trustable metrics to use then?
Can you apply the same thinking for walk forward optimization by using say spy for the backtest and all other tickers to forward test it?
also when you walk forward and optimize it every tear that you have traded aren't you slowly optimizing for the latest data?
thanks for the video.
Very informative 👍 thank you
No problem!
Nice! I think you set the gap up increment to 1 instead of the 0.5 you intended 🙈
Oops I'll double check
Hi Jacob could you make a video where you teach how to create a front running bot for pancakeswap. Thanks
I can try!
First time I heard about this
Really? Good I could teach you something.
Top