How I Pick Stocks.

[@JonesDawg]

This is a momentum strategy with a filter on volatility. Great idea and easy to understand. Looking forward to future vids on the topic!

[@josequinonez3079]

Hi ritvikmath! Great video! Congratulations on making so easy to understand these concepts! You said 90% of your investments go to index funds. Are you going to make a video about your strategy for that? Thanks !

[@roeiohayon4501]

Thanks for the video! (Nice intro BTW)

[@ritvikmath]

thank you :)

[@pragtinarang6680]

Great Video. Please make more videos about trading stocks. One suggestion is... something about what expectations to set or the emotional mindset to be successful in trading. For example, if you sell a stock too soon, no matter what profit you make, it always feels like a loss. Do you see?

[@ritvikmath]

Interesting idea; there is a lot of psychology that goes with trading. I'll try to address it in a future video. Thanks!

[@FB-tr2kf]

Love it rit. Esp u providing the code.

[@ritvikmath]

Always!

[@ResilientFighter]

@ritvik you should also do cointegration and Kalman filter

[@danch_ranch6594]

Hey ritvik, I discovered your channel when I was first introduced to Machine Learning and more advanced statistics. I have by now watched about 5-10% of your videos and I am amazed with your knowledge. Thank you! My question is about the series_tickers.p file; When I tried to open it in sublime as a ().py file, my dad recognized it as a hexadecimal file. So we tried to convert it to binary form, considering the script was "rb" or "read binary" (I think). So what is the advantage of hexadecimal data vs. normal csv file or binary data? Thank you once again for your time making these videos:)

[@MegaNightdude]

Hi ritvikmath, thanks for the awesome video. I was wondering what would be sensible thresholds for the mean and standard deviation of the monthly ROI (%), above and below which (respectively) a stock would seem worth buying and holding for a month? I know the caveats and disclaimers, but can you think of a good rule of thumb that has worked for you? Thanks!

[@loluser124]

Look up Sharpe Ratio

[@JJ-ck9iy]

This is interesting!

[@ritvikmath]

thanks!

[@cryptosocunswnews8950]

Very helpful video, just subscribed. I was able to run the code attached but I am wondering how I can do this for ASX listed companies. Are the same packages applicable in the case of ASX data? Thank you!

[@anindadatta164]

This will work well where trend is upwards and volatility is low, the kind of market observed in last 10 years primarily due to QE. However, for more volatile EM, some short term and medium term MAs could be added to the model for better results.

[@asif09ansari]

Nice video Ritvik! What is the exit strategy?

[@lorenzobaggi161]

Incredible introa hahahh

[@dasundesilva5588]

Please do more strategies like this

[@bapireddy5790]

I think that you are assuming past returns reflect future performance. What's your reasoning for this assumption

[@ritvikmath]

Good question! Clearly this is not always true, especially when dealing with individual stocks. However, this method attempts to find stocks with a consistent pattern of high returns within a recent time window. Of course we could see that the next day, some event can happen that causes the price to drop significantly.

[@philipphabicht497]

Great video! But how do you make your own ticker?

[@derekfun75]

Thank you for sharing the codes and the insight. I followed exactly your codes inJupyter notebook and I cannot generate any graphs after running the code. NO charts show up. And I also incorporate %matplotlib inline and still no charts shows up. Do you have any clue?

[@danch_ranch6594]

I have the same problem. Let me know if you find anything. Thanks! EDIT: I think the problem lies within the plotting function itself. Instead of using the series_tickers.p, and converting the hexadecimal file into binary form, I tried making a csv file with some random stocks tickers I know, and it went through without any errors, but there still was no plotting. I thougt the problem was the series_tickers.p file, but using logic, it wasnt. So my theory is that the plotting function is the error

[@hunterfeagans590]

Are you using an arithmetic or geometric mean? I suppose since you are using rolling 1-mo returns that a geometric mean would be optimal -- considering compounding returns.

[@mihirbhatia9658]

Do you have any experience with APIs that use minute level data to make short term predictions to day-trade ? I haven't been able to find any free and reliable source of tick data.

[@ritvikmath]

Hmm .. I haven't used such an API but if I come across one, I'll let you know!

[@mihirbhatia9658]

@@ritvikmath cool thanks. Your channel is amazing.

[@shawnroy1933]

I think the Yahoo Finance API provides reliable and real-time data during open market hours. Or maybe Robinhood/TD Ameritrade/E-Trade/one of those guys have an API to get real time data and submit orders via API.

[@FunnyDIYerGardner]

How did it perform In last 1 year in 2022. Would you be able to share performance of the code

[@zollen123]

As an amateur like many of us, do you think it is a good *exercise* for us to try to use SARIMAX to develop our own stock picking strategy?

[@MrAlket1999]

How can I modify the code in order to get the stock that lost the most ground in the same period no matter the volatility ?

[@ritvikmath]

If you want to look for stocks that have significantly decreased over the period, try changing the line starting with "if avg_return > min_avg_return" to "if avg_return < min_avg_return" and then set min_avg_return to some negative percent like -0.1

[@MrAlket1999]

@@ritvikmath Thanks, it works :-) In the link below there seem to be a strong loser for the period considered: finance.yahoo.com/quote/clvs?ltr=1

[@MrKhanRizwan]

This code no longer works, as yahoo finance is not working with the datareader, any thoughts on how to resolve that problem?

[@gigz54]

Hopefully you sold that YNDX before it got to today's price!

[@Sam-on3eq]

Awesome video! Wondered how it would look in R, so I made a R shiny app based on your Python code. In the process, found out that Numpy returns population standard deviation by default. App might have some slow load times and a few bugs I missed. github.com/parmsam/simple-stock-picker/ App was too slow when I attempted to integrate avg return and dev return sliders, so I opted out of those.

[@2A9D8F]

great job mate

[@philippe.geukers]

I backtested the algorithm, and it didn't offer a consistent return. On a single stock like AAPL or Nvidia, it doesn't perform as well as just buying the stock, and on the hole economy it stumbles with the market and doesn't bounce back with it. I tested from 2017 to 2020, and with different parameters, but none of them gave a positive alpha. Do you have any other details or ideas that might make it profitable?