42) Research Study of Correlation between Trading Timeframes

[@leonjbr]

Martyn: this is amazing. I never thought about this possibility but it is great. Thanks a lot for this content. As allways egearly waiting for the next video.

[@christopherjamieson4921]

Martyn’s videos are awesome. Huge fan. Great work, very insightful. I’ve coded my first multisymbol EA which incorporates asset correlation and the CAGR/MDD performance metric thanks to these videos.

[@jocollo979]

Hey Martyn. Now it's clear why you suggested to use filter indicators in a 4 times higher timeframe than the trading one. Awesome job with those heat maps. They will be very useful. I am expecting to see a heat map with all that you are explaining together: assets+asset classes+tf+strategies. That heat map will be HUGE!! 😁😁 Thanks! 👏🏼👏🏼

[@gianorozco7532]

your on another lvl martyn thank you for saving me a ton of research! your videos will not be in vain

[@seanjohnson7145]

Amazing, thank you once again

[@mohamadalikolahdouzmohamad5877]

Pure Gold...

[@m.ataide]

Thanks for the amazing job!

[@batatakhizou4509]

as alway nice video, but this make's me wonder how the spread is going to affect my performance on lower time frames, need to test to see by my self i guess.

[@p.s.r.krishna4250]

When you are doing a strategy (presumably the same strategy) for all timeframes you discussed, if we keep same lookback periods for all the indicators in that timeframe, each of those strategies are different w.r.t. the dynamics of price action. When you are comparing M5 to H4, correlation is expected to be low, isn't it? So it would be nice if we know what has been done to the lookback periods of the indicators. For example, a 2 period RSI on a H4 chart is similar to a 72 period RSI on a M5 chart. Please clarify.

[@simonfalta6294]

hello martyn, i really enjoy the videos darwinex puts out but if you want to get more views you should have shorter titles and put the number of the lesson at the end of the title the youtube algorithm will recomend it more that way. (just an idea) its very wierd how that works, maybe youtube should reoptimize their algo : )

[@donbangert]

Interesting. Quite a number of strategies I've seen advise looking at the higher timeframe in determining trade direction for the lower timeframe (H4 to H1 for instance). Your heat map indicates that, at a time difference of 4, there's actually very little correlation. Or am I reading the results wrong?

[@batatakhizou4509]

well that is the case but even with 3 times the current time frame the correlation still low (around 0.1), you can adjuste position size also according to the openned position on other time frames also, but this for the moment is not feasable for me (a lot of coding and no time for it)

[@a0nmusic]

the video wasn't clear on when you were comparing the equity change from M5 and M6 and calculation the correlation - are you comparing the equity change on a M5 bar to bar and an M6 bar to bar? or are you normalising to say a daily pnl change? If I were trying to calculate how correlated the different time frames were i would first map each to a common timeframe: so M5 and M6 would map to M30 => how did the equity change every 30mins across the 2 strategies. that way you can do a fair comparison. If i were correlating M5 to H1 i would just use H1 as the common bar. so how was the pnl change across 12xM5 bars to 1xH1 bar. For example say trading a super terrible basic 5vs8EMA crossover. The EMA on the M5 would hopefully capture the majority of a trend change of direction, whilst the H1 bar strategy would take abour 3-4 bars before it changes position.