Really good video, it's important that beginners know these basic strategies that can assist them with their trade career. It requires a lot to trade actively, I guess many make losses due to lack of skills to handle their tradings.
That's true. Doing technical analysis on your own time can be very demanding and boring. Thank God for all them pro brokers. I'm glad I don't gotta do that anymore.
Damn your content is exactly what I was looking for!!!! You MUST create a training course. Also interested to know if I can make more money working with Trading Bots than currently as a senior software engineer. Knowing that you came from Computer Science background I am very interested in your views about that. Would you say your gains are restricted by the amount of money you are willing to risk in your bots? I have an OK life but want to provide more to my family. Thanks!
Hi! Thanks for the great videos you put out. 1. A question reg. the rolling windows: Is there a reason to implement it by hand (other than showing the mechanics for educational purposes), instead of e.g. using a `pandas.DataFrame` and its `rolling`-method, e.g. `df.rolling(window=xxxx).max/min()` to detect extrema? 2. I guess a more general question out of curiosity: Do you use `pandas` at all? (Why or why not?)
This is the video I needed I had no idea how to build the target set, I was thinking in classifying as Buy, Hold and sell signals with a time series network. Hope it works
Very cool, I should have mentioned this in the video. I forgot. I tried training a classifier to predict local tops and bottoms a while back. I found it effective to give it some leeway, label the candle(s) before and after the actual bottom/top as well. That way the model doesn't get punished if its almost right. I didn't get great results overall, but I think its a reasonable approach.
I am trying to find the highs and lows ALTERNATING based on number entered I tried your code but sometimes I get 2 or 3 highs before another low. How can I make sure its alternating. I am trying to determine the percentage up or down it goes each way to determine what total percentage if I buy at each drop
this is the kind of info i would like to test . i went looking for a different way of looking for stock volatility. there was an anomally detection video on youtube that looked good. so i adapted it to stocks. it works pretty good. much better success now at picking stocks that have to potential to go up.
If you’re going +1 in index to find a top or a bottom, how do you use this strategy in the backtest? That’s a lookahead bias unless you’re saying we would use these tops/bottoms as a trigger for the trade
There is no problem in this backtest. In fact, if it shows a top in the live data, it is definitely going to the bottom(same as bottom to top), but I am not sure about waiting for any signal to buy or sell. because ask this function "What is the price of the coin when we receive the signal?" I'm not sure about the answers we will get. because we can draw candles for 1 hour or 15 minutes. When thefunction says us to BUY THIS ONE or SELL THIS ONE, its already too late...
When building a trading application you can't look forward in time, only backwards. The loop goes through each candle. if the loop is on the candle of 01-23-2020, we can't use data from 01-24-2020. Only data from prior dates. The scipy implementation does look forward, so it must be used with care.
thank you for your videos, very clear and perfect. can you make about the detecting Three drives chart pattern? I am working on it and used rolling windows before. but have never been successful. so if you could give me some advices, that would be awesome.
Yeah, I may do a more broad video for finding many harmonic patterns/Fibonacci stuff. Harmonic patterns seem to be a popular topic. In the past I have tried but they did not appear profitable.
@@neurotrader888 thank you. I wish i could talk to you about my team in private. We had so many good results but our main strategies needs these type of patterns as a trigger for combining other indicators.
It's great to discover this channel and thanks for your video. I'm now picking python for stocks screen and pattern recognition. Your methods are useful for identification of invented head & shoulder pattern with following uptrend movement, especially by locating the lowest/higher points and straight line method with distance measurement. I will start to try it in weekend, ha.
@@neurotrader888 but can I share my strategy with you so you code it for us to use....because I know nothing about coding and I know you are the code of god .....my strategy is not that difficult.....even if you can write a machine learning to look for my setups too it will help so that I wouldn't have to be going through all the charts looking for my setups or Is there a way for an AI to look for setups and send alert whenever it sees the setup.
