A heuristic explanation is this: If you consider using volatility in a model such as geometric brownian motion (stocks generally behave geometrically/exponentially), then you end up with the relationship ln(S(t)/S(t-1))=drift*dt +volatility*shock. So you can clearly see the volatility defined via log-returns is the right one to use.

Hello, I hate to bother, but might you help me with something? My OKX wallet contains USDT TRX20, and I have my recovery phrase [ clean party soccer advance audit clean evil finish tonight involve whip action ]. How can I send it to Binance?

It is already some time since I made this video so don't recall exactly but it is almost surely closing.

I suppose you are using python 3.12? Have a look here: stackoverflow.com/a/77638756/2743931

@@xkolm49 thanks, yesterday reinstalled anaconda and everything worked

Hey, sorry for the interruption, but I need your help. I’ve stored USDT TRX20 in my OKX wallet, and my phrase is (clean party soccer advance audit clean evil finish tonight involve whip action). How can I transfer it to Binance?

Pardon me, could you give me some advice or help with this? my OKX wallet holds USDT TRX20, and my recovery phrase is (clean party soccer advance audit clean evil finish tonight involve whip action). Can you explain how to move it to Binance?

Hi, I apologize for interrupting, but could I ask for your help? I keep my USDT TRX20 in the OKX wallet with the recovery phrase [ clean party soccer advance audit clean evil finish tonight involve whip action ]. How do I send it to Binance?

Yes, you are right. But it won't be easily extensible and you don't see the day-by-day logic. The purpose of the video is to teach, not to write the shortest possible code. You can always optimize your backtester.

@@xkolm49 Thanks for answer. I write strategies from scratch for a while and I ended up with the thought that loops are the best (of course with some restrictions like using only numpy arrays). But this code is hard. REALLY HARD! I don't understand it at all. It's so overcomplicated... For example: - little things like dates-mask to get the previous day. Why don't you use "for i in range(1, closes.shape[1]" and just "iloc[i-1]"? Simplier and even faster, cuz loc is slower + unnecessary dates-mask. Or just lines like line 40... Does it teach newbies? Don't think so... At least cuz "to_frame" method takes unnecessary argument "name" and you can make it more readable... But still I don't understand logic of this line and don't want to. Code is overcomplicated. - big things like 3 dataframes! 3! For what? I see some logic here and so on... But at least 2 dataframes just aren't needed, because your main output is only one of them, and other stores day-by-day data that you need ONLY when you fill day-by-day your main output-dataframe. So... Why don't you use variables? There is no reason to use them, but there is a drawback in memory (not crucial nowadays, but still), in speed (cuz of slow locs)... And ye, for daily data and one ticker there is no big difference, but... Any big intraday data and several tickers will these drawbacks shine. Well, if you wanted to teach - you should write a signal function and pass it to a function with the main loop where you call it with arguments you need... Well, it's great with OOP, but still - THAT would be more extensible. But for THIS strategy - my example-code is simplier, faster, all strategy logic fits, you can check strategy for all tickers you are interested in at once, really fast and efficient. And the main benefit - less risk of mistake. Because there are no loops with locs, maths and so on... At least, this code can be used to check their day-by-day loop function. Well, I understand that I'm nobody for you and wrote a lot of bad things about your code... But if your code is right and, as I can see, - some people says you helped them a lot. Then I think that you did a great work! P.S. Ye, I forgot to shift signals xD)

If I add 'yahoo' while using DataReader instead of the get yahoo data function I get the error; ValueError: time data 'google' does not match format '%Y-%m-%d' Note, this is all from your original script but this error to me implies that it shouldn't be there, perhaps i'm on the wrong version of pandas, all is updated.

But like said, when I removed 'yahoo' or 'google' I instead recieve the KeyError for 'symbol'

Спасибо!

When you run, it stops at the breakpoint, then enter 'n' it goes next line and shows variable on variable explorer.

you can use only 'int main ()'

speed up ++