Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

ChatGPT can help yield more accurate stock trading predictions – study

papers.ssrn.com

Post reply on HN
Print view View original

41–50 of 58 posts

Re: ChatGPT can help yield more accurate stock trading predictions – study

#41

Post by andreygrehov »

ckardat123 wrote: post #31
paulpauper wrote: post #26

Earlier quoted context omitted.

What sort of market prediction is decided by coin flip though? This makes more sense in the context of a binary outcome like sports betting in which you are betting on a team either winning or losing.

The scheme definitely could get foiled if people were just told that a certain stock would go up (without hedging against the market), because if it's a red day for the market the whole e-mail list could get wiped out as all the stocks go down. I would think that there is approximately a 50% chance of most stocks beating the market on any given day though. If you send a bunch of e-mails saying "${ticker} will go up m…

I'm not sure if that is truth (today). The majority of stocks have been doing a massive nose dive within the past year. I own stocks of approximately 10 different companies. All, except AAPL, under water. So - 90%. In contrast, defense and healthcare stock market has seen a different trend. I own zero of these.

Re: ChatGPT can help yield more accurate stock trading predictions – study

#42

Post by jwhiles »

fabiospampinato wrote: post #36
jwhiles wrote: post #29

Earlier quoted context omitted.

Buy stock X - it will definitely go up in the next week!

Though it's not like 50% of the days any given stock goes up and the other 50% of the days the stock goes down.

I've never looked into this, but I'd assume that a stock's price being the same as it started on any given day is one of the rarest results.

Re: ChatGPT can help yield more accurate stock trading predictions – study

#43

Post by andreygrehov »

zevv wrote: post #37

If any AI - or any other process or algorithm - would be able to provide significant better predictions than what is currently used, wouldn't this affect the stock prices as soon as this gets deployed by brokers at scale, effectively making the predictions useless again?

Check out Renaissance Technologies founded by Jim Simons [1]. They've been doing that for years.

[1] - https://en.wikipedia.org/wiki/Jim_Simons_(mathematician)

Re: ChatGPT can help yield more accurate stock trading predictions – study

#44

Post by OzzyB »

gooseus wrote: post #19

I feel compelled to write this though it's only tangentially relevant to the article and is an adaptation of Taleb's story of stock-picking monkeys[1] from either Fooled by Randomness or The Black Swan: Say you have an email list of 100000 investors, you ask ChatGPT to produce 100000 predictions and send these out to each investor. Assuming ChatGPT is as good as a coin flip, 50000 investors receive good predictions.…

This was dramatized in Darren Brown's "The System"[1] some years ago, but instead of stock picks he used horse race betting.

Its amazing to see the people that won on the previous Nth rounds believe that their next tip was a "sure thing".

[1] https://www.youtube.com/watch?v=zv-3EfC17Rc

Re: ChatGPT can help yield more accurate stock trading predictions – study

#45

Post by jrpt »

kljlkjfdsjlk wrote: post #40
jrpt wrote: post #18

I've found it useful using LLMs to help understand company 10-K reports like these: https://docalysis.com/library/10-k-reports For example, I used it to understand how much Meta is investing in its Reality Labs initiatives ($13.7 billion last year) as well as how much revenue Reality Labs brought in ($2.15 billion). I'm certain financial professionals are using it to better understand the world. The paper says it use…

I tried your website thought of a question to ask and then your bot said something about scammers and needing to sign up. Some meddling advice 1) think about giving an easy-win to users so they actually want to sign up. You didn't give one, and in fact surprise users in an annoying way because you don't tell users they need to sign up until after they've already submitited their question. Not even a teaser, just comp…

Sure, but the counterargument is that it costs real money to run queries, so it limits the amount of queries per day by your user account. Signing up is free, I just don't want someone racking up huge bills and abusing it.

Re: ChatGPT can help yield more accurate stock trading predictions – study

#46

Post by oliwary »

fabiospampinato wrote: post #36
jwhiles wrote: post #29

Earlier quoted context omitted.

Buy stock X - it will definitely go up in the next week!

Though it's not like 50% of the days any given stock goes up and the other 50% of the days the stock goes down.

One way to overcome this is to calculate the percentage change of a stock compared to the percentage change of the median performing stock in the market. 50% will be perform better than the median stock while 50% will perform worse.

Re: ChatGPT can help yield more accurate stock trading predictions – study

#47

Post by ckardat123 »

KeplerBoy wrote: post #35
ckardat123 wrote: post #31

Earlier quoted context omitted.

The scheme definitely could get foiled if people were just told that a certain stock would go up (without hedging against the market), because if it's a red day for the market the whole e-mail list could get wiped out as all the stocks go down. I would think that there is approximately a 50% chance of most stocks beating the market on any given day though. If you send a bunch of e-mails saying "${ticker} will go up m…

Beating the market also means just falling less than the index on some days. Not sure if those customers would be happy. The reverse also applies a.k.a. everyone could make money in the past few years.

That's why I included the part about also hedging against the market.

Not that it's particularly easy to do, but if you took out an equal weighted short position on the $SPY while going long on some ticker, it would pay off if that ticker outperformed the market.

Re: ChatGPT can help yield more accurate stock trading predictions – study

#48

Post by ckardat123 »

andreygrehov wrote: post #41
ckardat123 wrote: post #31

Earlier quoted context omitted.

The scheme definitely could get foiled if people were just told that a certain stock would go up (without hedging against the market), because if it's a red day for the market the whole e-mail list could get wiped out as all the stocks go down. I would think that there is approximately a 50% chance of most stocks beating the market on any given day though. If you send a bunch of e-mails saying "${ticker} will go up m…

I'm not sure if that is truth (today). The majority of stocks have been doing a massive nose dive within the past year. I own stocks of approximately 10 different companies. All, except AAPL, under water. So - 90%. In contrast, defense and healthcare stock market has seen a different trend. I own zero of these.

That's why I included the part about hedging against the market. There may be days where almost every stock goes down, but generally only around half will every go down more than the market.

Re: ChatGPT can help yield more accurate stock trading predictions – study

#49

Post by electroly »

fabiospampinato wrote: post #36
jwhiles wrote: post #29

Earlier quoted context omitted.

Buy stock X - it will definitely go up in the next week!

Though it's not like 50% of the days any given stock goes up and the other 50% of the days the stock goes down.

It doesn't matter for this hypothetical scheme. It'll always be one or the other outcome, and you will thus look prescient to one half or the other half of the victims. Each day your pool of victims cuts in half but you don't care which half it is. The scheme works regardless of how often the stock actually goes up or down. At the end, there will always be one victim who was given exactly the correct series of predictions by you ahead of time.
Post reply on HN