historicaldocuments comments on OP created a stock market game where you compete with a couch. Guess who wins.

Best Content of Lemmy

176 readers

885 active

2 users here now

A place to celebrate the best content of Lemmy. Reminder: This is a celebration of even the small things of the Best of Lemmy. It doesn't have to be the best you've ever seen, just the best you've seen in the last hour or day.

Rules:

Be Excellent to each other
Provide screenshots if possible and a link if possible
Please provide date (not required) of original content your sharing. If you hover over the date, it gives the exact time.

Here is a link to the award icon if you desire to let the OP know they've been added: https://sh.itjust.works/pictrs/image/37cb0ef5-a25c-4ac5-bcba-2048d2738f87.png

founded 1 month ago

moderators

[–] historicaldocuments@lemmy.world 4 points 1 week ago (1 child)

Buffet can sling around more money than most and tends to buy huge chunks of single companies rather than move a lot of chess pieces at once. I thought the article on Renaissance was interesting with regards to how periods of high volatility jammed up their computers/models. I wonder if all these NVIDIA chips that work well for LLMs would work well for investment style calculations.

permalink

fedilink

source

parent

hideshow 1 child comment

[–] ooli3@sopuli.xyz 3 points 1 week ago* (last edited 1 week ago)

no llm are bad at predicting market : https://www.zerve.ai/blog/llms-in-quantitative-research And I'm pretty sure it is not a human problem, a lot of fund, and intelligent people inside those, must have tried to make very advanced llm predict stock market for them. And if some succeeded (even remotely) the llm makers would have loved to publicize those.

Quantitative model, usually use average and long term strategy, using data in the marked to anticipate how the marked would move next. High volatility move.. are just things outside of the market moving it weirdly. Models can not predict nor anticipate those. It would be like predicting a terrorist attack, or something unusual.

Aside: High intensity trading is not a strategy, it is just a computer market maker : using known bid and ask and matching them fast, and catching a few fraction of cents every time, but doing it all the time. No prediction in that.

permalink

fedilink

source

parent