An Insightful Autocorrelation Analysis Walkthrough - Step by Step

  Рет қаралды 1,310

Crypto Wizards

Crypto Wizards

Күн бұрын

Пікірлер: 29
@ShawnJiang1125
@ShawnJiang1125 3 ай бұрын
Great video! Very inspiring content. I'll have my team working on machine learning and neural networks take a look soon to see if there are more possibilities! Shaun, your updates never disappoint!
@CryptoWizards
@CryptoWizards 3 ай бұрын
Shawn! Thanks for the encouragement as always!
@alexwatkin4263
@alexwatkin4263 3 ай бұрын
solid work
@CryptoWizards
@CryptoWizards 3 ай бұрын
Thanks Alex!
@FreddyMinn
@FreddyMinn 3 ай бұрын
I've created both Stastical Arbitrage bot and Charts web version Using JavaScript that base on your Udemy courses. To exit I use Average True Range values as Trailing Stop Loss for both positions. Million Thanks for sharing your secrets ❤
@CryptoWizards
@CryptoWizards 3 ай бұрын
Nice work Freddy! Also thank you for sharing your insights with the community!!
@ministryNoiz
@ministryNoiz 3 ай бұрын
Thanks for your video! Have you researched all trading pairs on some exchange? If so, how many pairs have a transition percentage greater than 60? What is the highest percentage you have seen?
@CryptoWizards
@CryptoWizards 3 ай бұрын
I haven't done it for each exchange, but 60% to 70% is not uncommon. My hunch is that as the pairs become more traded, the autocorrelation disappears. What causes it is open for debate!
@ay5960
@ay5960 3 ай бұрын
Most assets with good liquidity will have 0 auto-correlation, but the future returns are not independent of past returns. There are correlations between the magnitude of returns (volatility clustering), hence you get auto-regressive models like GARCH to model volatility. Might be helpful for options.
@CryptoWizards
@CryptoWizards 3 ай бұрын
💯
@proo3968
@proo3968 3 ай бұрын
Amazing stuff mate! Do you think people use it as a feature for ml?
@CryptoWizards
@CryptoWizards 3 ай бұрын
Good question. I could see a scenario for using ML to learn from trades of assets that have autocorrelation. Although for analysing a single asset, the use of a state space model for autocorrelation could be better for ML perhaps. Will be covering this later on.
@NuruddinKhan-w8r
@NuruddinKhan-w8r 3 ай бұрын
can i have the Python Code for Data:
@CryptoWizards
@CryptoWizards 3 ай бұрын
I just used the CW XTrader App to get the data for this video.
@CryptoWizards
@CryptoWizards 3 ай бұрын
Also forgot to mention that you can download the spreadsheet with the link in the comments to get the data
@alexjacob612
@alexjacob612 3 ай бұрын
Sir can you create a funding fee bot pls? Spot /futures arbitrage
@CryptoWizards
@CryptoWizards 3 ай бұрын
Sounds like a good project for when the tool upgrades are done.
@alexjacob612
@alexjacob612 3 ай бұрын
@CryptoWizards it's a good strategy but not many people talk about it I don't know why.
@jon-uk1468
@jon-uk1468 3 ай бұрын
Great video and it nicely follows on from the last one. How can I combine this algorithm with z score and half life calculations ? And is python now the right language to use as you have mentioned Rust in the past.
@CryptoWizards
@CryptoWizards 3 ай бұрын
Indeed, with the spread we actually use something called an ECM, which am building into the upgrades coming to analyse what’s known as the “speed of adjustment”. It’s easy to understand once past the up and coming videos so will cover it later on. For now, we want a spread to be mean reverting and we can analyse similar to autocorrelation through what’s known as an AR process. In terms of programming language, it doesn’t really matter unless doing HFT. I like Rust because it’s very robust and fast, but the learning curve on it is steep. Python has more libraries for statistical analysis already built etc, so faster to deploy ideas. Sometimes I test ideas in Python and then write the final implementation in Rust. Of course ChatGPT speeds this all up now.
@jon-uk1468
@jon-uk1468 3 ай бұрын
@@CryptoWizards thanks again for replying. I will study your reply in more detail tomorrow. Are you updating the z score app to provide this additional data in the tables ? This sounds like a more reliable way of finding pairs to trade. This would be a massive help as finding the pairs to trade with reliable data is key here. I'm ok at taking profits but not very good at finding assets to trade.
@CryptoWizards
@CryptoWizards 3 ай бұрын
Absolutely, the ZSCORE tool updated coming are extreme. This is one big reason for making these latest videos, to help everyone become aware of what all the new metrics mean. Thanks again for such helpful comments. Will do a YT short to cover programming languages too...
@jon-uk1468
@jon-uk1468 3 ай бұрын
@@CryptoWizards Thank you for replying and you are welcome. I am fascinated by the date science and as an ex coder from the good old days , I have a real interest to learn and share my learning too. As a lot of the members on CW discord do. Are you going to add another channel to discord ? I am happy to help mod it foc . Also happy to beta test foc if you need any additional help / resources. I'm on as Jon UK . Either way I'm really looking forward to seeing the next videos and the new z score app 😃
@CryptoWizards
@CryptoWizards 3 ай бұрын
kzbin.info4yRVza7aSgs?feature=share
Correlation (Pearson) Explained with Excel Walkthrough
20:57
Crypto Wizards
Рет қаралды 776
The Trillion Dollar Equation
31:22
Veritasium
Рет қаралды 10 МЛН
Мен атып көрмегенмін ! | Qalam | 5 серия
25:41
СИНИЙ ИНЕЙ УЖЕ ВЫШЕЛ!❄️
01:01
DO$HIK
Рет қаралды 3,3 МЛН
Attention in transformers, step-by-step | DL6
26:10
3Blue1Brown
Рет қаралды 2,1 МЛН
Estimating Relationships Using Linear Regression
24:22
Crypto Wizards
Рет қаралды 757
The Oldest Unsolved Problem in Math
31:33
Veritasium
Рет қаралды 13 МЛН
Complex Fibonacci Numbers?
20:08
Stand-up Maths
Рет қаралды 1 МЛН
You might never look at another price chart again...
11:00
Crypto Wizards
Рет қаралды 5 М.
But what is a neural network? | Deep learning chapter 1
18:40
3Blue1Brown
Рет қаралды 18 МЛН
Time Series Talk : Autocorrelation and Partial Autocorrelation
13:16
Correlations, Autocorrelations and Correlogram
12:02
Gmaz
Рет қаралды 54 М.