?? Work at Pickup Points � Flexible schedule

A machine learning approach to stock trading | Richard Craib and Lex Fridman

Lex Clips 134 views

Download Video

Comments

🎯 Key points for quick navigation:brbr00:02 b📉 Machine learning can automate stock trading by detecting price breakouts/bbr01:22 b📈 Breakout indicators help identify significant price movements in trading/bbr02:10 b🖥️ Python coding is utilized to create automated trading strategies/bbr03:39 b📚 Open-source code is available for developing trading algorithms/bbr05:24 b🛡️ Risk management strategies are crucial for successful automated trading systems/bbrbrMade with HARPA AI
Comment from : @imranaalam

loads of BS
Comment from : @Suckerx911

Lex misspoke "reduce exposure to red lights" because he probably just spoke to Andrew Huberman about red light therapy
Comment from : @nexus7c0

Historical data is useless for stock price
Comment from : @mkahraman82

with the front of machine learning and algorithmic trading this guy is running a crypto scam
Comment from : @Rolololollo

Parasites talking ABT their scams
Comment from : @FirstNameLastName-fv4eu

'Amazing video, A friend of mine referred me to a financial adviser sometime ago and we got talking about investment and money I started investing with $120k and in the first 2 months , my portfolio was reading $274,800 Crazy right!, I decided to reinvest my profit and gets more interesting For over a year we have been working together making consistent profit just bought my second home 2 weeks ago and care for my family"
Comment from : @mayacho4910

I remember when lex used to breakdown the bjj for yoga guy’s fights brFor people that want to go down trading rabbit hole, don’t it’s a wast of time brYou need to read the reports and listen to earnings calls brAI can’t preform high level judgment, the only way robots might make you money is trough arbitrage brAll models are wrong some are useful
Comment from : @michaeltabet9223

This guy is smarter than me so I’ll just make fun of his hair
Comment from : @Eric_In_SF

Interesting idea but having an intimate understanding of the data is the most important pre-requisite to attempting any type of ML approach Not being able to read the column names or contents is a massive handicap Removing human intuition from feature engineering is like asking a doctor to cure a patient without seeing them
Comment from : @justindehorty9739

There are no trading strategies (including data) that work that you can just buy, for two reasons:brbr- Nobody sells a working tradong strategy because a working trading strategy is worth more than money you can sell it forbrbr- Any advantage that a working trading strategy has will quickly disappear as soon as it becomes known to others
Comment from : @riccello

Stay high as long as possible😂
Comment from : @timzimmerman1310

@ 9:40 what he means is the more you expose your Model to obstacles and set backs and failures the more it will learn adapt and develop Similar to humans
Comment from : @DeepFrydTurd

Data is important because it allows you to predict the future with higher accuracy and steer your operations towards that direction and capture unmarked territory and profit from it first Like Christopher Columbus And now with the recent developments of Machine Learning Ai models and new optical hybrid CPUs and GPUs we are getting closer to an intelligence faster and possibly greater than our own and that's not taking into account NeuChip Biological Ai and Neuromorphic Computing
Comment from : @DeepFrydTurd

They're 'giving away data' That's just a marketing way to put it
Comment from : @the_notorious_bas

great content and insight
Comment from : @raphael_mogger

love you
Comment from : @kevin-jm3qb

denial hair
Comment from : @user-po7te6hn2s

Ben Goertzel‘a SDAO will bring ai trading to everyone as of December 2021
Comment from : @yellowledbutter

Democracy is stupid
Comment from : @code5829

He talks about 2 being a lot Already there you know he is a scrub
Comment from : @socrates_the_great6209

Investing is counted in yearsnot monthswhat a fool he is
Comment from : @socrates_the_great6209

How would you predict that Apple would get back Steve Jobs?
Comment from : @eanerickson8915

I literally can't tell if this guy has a british or american accent
Comment from : @improcrastinating8063

This man is a bs merchant
Comment from : @jaydee8069

if it's over 3 months this doesn't make sense events like covid and other crisis greatly influence it (and you can't predict that with stock data)
Comment from : @MrFlexNC

Good thing Richard Craib was coherent
Comment from : @Jangellz

Hard to believe his “data” is anymore special than data found on the web Anything that is publicly traded is basically public info Furthermore, stock price only has a few data points per day: open, close, high, low, volume (adjusted close if needed) making it a minimalistic dataframe In the end, if people attempt to “solve” the data, they’ll end up adding columns like moving averages, stochastics, OBV, RSI, MACD, DMI ending up backtesting reversal/pivot points based on those metrics interacting with one another If he adds market beta like jobs payrolls, FX flows, US treasury yield curve, he’ll end up with 10s of linear regression models (not even considering polynomial lines) that attempt to give a signal at the same time Seems like he is trying to reinvent the wheel when the wheel’s already been invented Good luck nonetheless
Comment from : @sweealamak628

;)
Comment from : @LarryPanozzo

Tying to learn how to trade automatically without learning how to trade manually brilliant (sarcasm) You'd be better off to hire a bunch of trainees, pay them minimum wage and fade them
Comment from : @dachanist

Eminem just became really smart
Comment from : @ferdinandbrand1

It reminds me on the Goldcorp Story and Don Tapscott many years ago
Comment from : @klauszinser

How is a machine learning approach any different from a 'traditional' time-series forecasting approach to stock trading?
Comment from : @ctothaz

If this guy plays his cards right, he'll be rich af in no time
Comment from : @AmadinEnobakhare

People no longer visiting companies to see and understand their daily function, how they create value? If so, it's only speculation which is vulnerable to manipulation
Comment from : @chillbeach7322

I like the idea, but don't see it working in the long term because of game theory principles Personally, I see more value in preserving and executing my own alpha rather than submitting signals to numerai However, I will still run my trading system on $nmr because regardless of whether I'm right or wrong, profit is profit
Comment from : @AGI_CEO

This approach will have limited success First problem is only looking at 10-15 yrs worth of data which is essentially a long bull market Other problem is with anonymizing the data is that you can't derive any meaning Understanding what the data means is essential to making a decision and ML/AI is still to basic when it comes to understanding Understanding the data is how macro economists (Dalio, Drunkenmiller) and subject matter experts (Buffet, Chamath) get their edge
Comment from : @khurramtahir687

I wonder what his or Numerais opinion on the 10 year yield is!! 😱🤑
Comment from : @djsvideodiarys

Basicslly Momentum exploited onto Other Features?
Comment from : @ClawDragoon

"Stay high for as long as possible"
Comment from : @horseheadbutt

It's not investing - it's trading brYou create no value - you extract the value that others have created brFocus on something that benefits humanity, not your bank balance
Comment from : @the-selfish-meme7585

I wonder how they guarantee that poor performance can be attributed to the model as it might also be due to a large difference in data given and live data On top of that, ensembling all these models does not seem arbitrary Interesting stuff though, will look into this more!
Comment from : @M0481