Algorithmic Trading

What do you think about this field? What would be the most useful approach? I know how to program in Python3, I guess the language is just the means to carry out the application of what is truly important: The abstract concepts of statistics and finance. If you have experience in this area, could you give me a couple of tips if I want to focus on mastering Algorithmic Trading in this industry?

>>40425 You could check algotrading subreddit, they have good resources and you can ask posters there.

I second >>40426 In general, learn finance and trading first. Many textbook implementations of common analysis/indicator algorithms exist out there but that does not equal instant profit. If you do not know the steps to bake a cake you cannot teach a robot how to bake. Once you have a clearly defined process you can join the dots and implement/automate every step and path from analysis, weighting, timing, backtesting, everything. The level of math and statistics required is usually not very hard. In the case of indicators it is more about understanding how the algorithm is using X concept to derive Y. I do not know what it takes to truly master it but in my experience mastering anything is based on having solid foundations.

>>40425 This book provides a lot of details into the problem.

>>42297 Not the OP, but very cool anon, thank you! Are there any sections in there that you found particularly valuable?

This just looks like free money. If it works, why isn't every programmer doing it? Why are so many people complaining about being out of work if they could be making money in their sleep?

>>42376 Because you're still competing with Wallstreet and about every rich user on planet earth. Not to go over the experience that you must have in order to make a functional, profit-gaining AI. Also >St. IGNUcius Is that an actual lainchan name or did you just insert it lol?

>>42377 >the experience that you must have Of course you have to put in some effort in the beginning. >still competing with Wallstreet The big traders are limited by their size. Moving huge amounts of money changes the market conditions, to the point where most of the time they won't actually end up making a profit, so they just do nothing most of the time. If you're trading "just" thousands of dollars you're too small to change anything, so you can profitably buy and sell on normal fluctuations.

>>42376 Most people who start trading lose. Its similar to trying to play competitive poker but the old established players are allowed to cheat here. i.E. neobrokers have been feeding their customer data to larger banks as their income model and those banks exploited that knowledge to farm those customers positions. Then there is the usual fear of the unknown. If you run a trading bot you run a permanent online server with access to your wallet - could get hacked and emptied. And before that you might get emptied by taxes and law changes you missed. Eu ball&gagged crypto just this january and somewhere around 2019 they messed with stocks. And lastly you can't start earning with $50 because of fixed fees. Simplified: 5% profit = need to invest 20x the fixed fees just to break even. This is why traditional advice was minimum $10k to start though nowadays fees decreased. tldr: making money in trading takes serious work on top of taking risk

>>42383 >Most people who start trading lose. Most people are trying to catch the next NVDA and have no idea what they're doing. >If you run a trading bot you run a permanent online server with access to your wallet - could get hacked and emptied. When did this become about crypto? It's a bad idea anyway. The market can be irrational of course but at least stocks are based on actual things. >And lastly you can't start earning with $50 because of fixed fees. Simplified: 5% profit = need to invest 20x the fixed fees just to break even. Use a no-fee trader. There are plenty of reputable ones now. >>40443 Second this post. All that jumping directly into automated trading will do is make your fuarrrkups inhumanly fast.

>he still values currency lain, lain, lain...

>>42657 Well, Femboy Hooters isn't taking my mass collection of informational PDFs nor skill trading or bartering. Until then I'll value currency so I can eat.

>>42656 >When did this become about crypto? It didn't, a stock trading bot is also holding your wallet, a hacker can empty your stock trading account same as he could with a crypto wallet. >Use a no-fee trader. There are plenty of reputable ones now. Which one? I'm genuinely curious because when i looked i found hidden fees everywhere while they claimed no fees.

would anyone be interested in going through this book >>42297 in a matrix/discord together? i'm mostly thinking of reimplementing the code from here in rust.

I have been working with algorithmic trading for some years in my spare time, and since the beginning of the year I finally have a bot that is consistently profitable; but it has only made 10% (after taxes) so far.

>>42666 thats actually really good if you can keep it up reliably how does it work on historic data?

>>42667 >how does it work on historic data? Not sure what you mean by that. If you mean "how is the performance in a backtest of the last 20 years", the answer is I don't know. Contrary to most people, I don't think that it makes much sense to try to create a bot that performs well in that timespan; the markets changed way too much and you can tell if you look at the data. Furthermore, I collect the price data myself, because all of the historical data you can get is "smoothed" and "cleaned", i.e. rigged. The guy in pic related talks about this, but basically I saw in my broker that if, for example, for today I get no price for EUR/USD at 21:37 (because no one traded in that time), on the next day the broker will display a price in the chart. This seems to be common practice since the 80s, and my bot reacts to live prices, not old prices. I simulated it - this makes a huge difference. If you mean technically, I story my price data in a MySQL database and pull it from there to compare the live trades to the simulated model.

>>42669 >2137 najsu Body, as per usual, too long and not short enough.

>>42669 was talking about backtesting yes One of the benefits of backtesting is that its nice to see what happens on the lets say worst 1 in 1000 or 3000 days. You can have a strategy that's pretty good pretty good pretty good and then one bad day just ruins you. Interesting to see what would have happened to you during i.e. covid. its probably true that a great algorithm circa 2005 isn't going to be a great one today. Depending how long you've been doing it maybe you even literally saw what happened during covid though IDK.

>>42671 During covid in the beginning, my bot would have lost about 200$ and only made money from there. While I started working on this more-or-less in 2014, I wasted so much time in the wrong areas and therefore started for real just in 2021. Otherwise I would have made so much more money. >one bad day just ruins you I limit my bot to 1 trade per day, therefore this does not happen to me.

>>42677 >I wasted so much time in the wrong areas What should be avoided?

anyone interested in forming a team to make a submission to this challenge? the prize money isn't so bad. https://www.kaggle.com/competitions/mitsui-commodity-prediction-challenge/overview

>>42678 >What should be avoided? I will kinda dodge the question because what I originally meant is something different. You can laugh at me. Basically, back in the early crypto days there were websites where you could exchange crypto currencies anonymously, just by entering the amount and two wallet addresses. I did not trust any of the exchanges where you need to register, so my dumb idea was to just program my own blockchain clients to handle all the transactions. Took me about a year for bitcoin and litecoin and right when I was finished, the laws were changed so that these kinds of trades were not allowed anymore and the sites started to require ID. Now I know how to implement a blockchain from scratch, but who needs that. This is possibly not what you wanted to know, so let me give you a different hint: Something that I do not see anybody doing, is: Compare your live trades against your backtest. Basically, after your bot closed a trade, run your backtest on the data it used. Your backtest should perform the same as your bot or worse (!). If not, there is some issue with your live bot implementation (or your strategy does not work live in the first place). I looked at countless frameworks for placing trades/running backtests and read forum posts, articles etc. and no one seems to take about this. The second tip is, start with something really simple. No need to learn complicated math. For instance, lots of markets that are not forex have a natural upward bias, e.g. gold or stock indexes. These biases are kinda unavoidable due to inflation. So for example, you can create a signal based on the closing prices of the last three days. If they are higher each time, open a long position. This is something really simple that works (you can try this in simulations yourself) if properly adjusted (window size, time offset for the daily price, profit target).

>>42665 irc.lainchan.org/#programming welcomes you

>>42669 Any books to recommend for getting general/background trading knowledge?

>>42666 >I finally have a bot that is consistently profitable Do you think you'll ever make enough to live on?

>>42297 >>42665 i would be interested. I was reading that book years ago and i remember it being about agentic architecture in Common Lisp. But tech changed so much recent years and i would be really interested in using Langchain and LLMs for implementing agents described in this book

>>42782 >i would be really interested in using Langchain and LLMs for implementing agents described in this book that sounds very interesting. solar_zonal on discord

>>42666 Any resources that you'd recommend or is all your knowledge a massive collection of tidbits here and there?

>>42669 Hey, I'm also working on backtest data collection as a side-project. Can you advice on how to build such system? It's tedious to work with a bunch sources to make them somewhat similar but still have like prices each 15 minutes.

I got my divergence detecting functions to work. https://www.babypips.com/learn/forex/divergence-cheat-sheet

BattleCodes is an onsite, broadcast-grade trading competition happening Oct 7 at OS NYC (Manhattan). - All trades live on BNB Chain mainnet via Aster - PnL-based leaderboard updated in real time - Bracket eliminations (10 → 6 → 3 → 1) - Finals streamed on Twitch + YouTube - $10,000 prize, winner-take-all https://luma.com/9y3zbnry https://x.com/battle_codes https://time.is/compare/1200PM_7_Oct_2025_in_New_York I'm not ready to compete, but it might be interesting to watch.

>>42966 This event was cancelled. Nevermind.

>>42760 >>42865 >>42866 Come on Llama Lain, please respond. All we want to know is how to never work another day in our lives - is that so much to ask?

Anyone here trading on prediction markets? Have a strong feeling these will blow up in popularity soon.

>>43289 >Have a strong feeling these will blow up in popularity soon. In response to what?

>>43307 VCs have made large investments in the top two prediction market platforms recently. But beyond that, these platforms have seen largely increased volume and liquidity over the past few months. As they keep adding new markets to bet on, they'll only get more popular

Do you think this is doing well for a grid bot? I simply set it up in like 5 minutes and leave it running and check it once a week or two, as you can see this is where it is after 43 days. If I had nothing and simply hold my ETH as is, I would have 96.40 USDT less at this point, right now I'm at 22.58 USDT profit and if rate goes back to the same as it was when I started the bot, I will be at more than 118.94 USDT profit (as going back to that level would mean selling half the open orders bot already has). Do you think there is an equally easy way to do better than this?

>>43309 Forgot the screenshot of bot itself kek.

This is how I'm doing right now. Still interested if you think it's fine for basically checking the thing once every week or two, or is there an easy way to do way better.

>>43386 >>43310 Share some starting-out study suggestions, since llama-Lain has left us?

Can we get any pdfs to get started with learning the maths and the theory behind trading?

>>43719 The mathematics you would use would depend entirely on what strategy you are going to do so you have to be more specific about "math" and "theory" behind trading

>>43758 I think its safe to assume no strategy involves geometry.

>>43719 Bottom line, get a textbook on statistics and one on differential equations

Can you point me to a single person who: 1. claims retail traders can out-perform the index 2. is not selling a book or course 3. has a published track record of profits?

>>40425 It's no better than random chance. You have trending algorithms and [mean-]reversion algorithms. One regime's will do poorly in the other's regime. If you use algo trading, you need to know when to turn them on or off, and on what underlying. This means you need external data, not just price/volume chart data. If you just let a trending algo rip on pure price data alone, you won't do better than coin-flip; and because of spread/tx fees, you are in effect playing a game like Roulette and will likely pay the market over time. Pirate Institute of Trading and Portfolio Management (ITPM) materials on bittorrent, specifically Professional Trading Masterclass v2 (PTM v2) and Professional Forex Trading Masterclass (PFTM). You don't need Options (POTM) until you've first learned top-down global macroeconomic portfolio management with equities.

>>44499 you won't usually find those because they're enjoying their life. there's a couple exceptions, I've been with a couple ex-bankers and traders that at some point made a "private group" that was paid. I would never ever join such a group if it was publicly advertised in any sort though. those groups were usually made up of skillful traders and they all shared knowledge and trade setups as they saw them, no strings attached to be honest im quite jaded nowadays so I dont even have the networking to make new connections like this, so I keep the groups I have dear to me. but yeah best piece of I advice I can give is to read, read, read and make connections with people that ACTUALLY work in the related industries

>>44540 >read, read, read So the books are for real? It's just hard to believe that anyone with the ability to basically pull money out of thin air would take the time to put together instructions for potential competitors. What do they get out of it? Are they just bored?

>>44540 >make connections with people that ACTUALLY work in the related industries Do you mean in trading/finance or in the industries of the companies you want to trade?