2026-01-18 20:38:54

Can AI models really outperform traditional indices? A test gave us the answer.

Since mid-November this year, someone has started a comparison test—using AI models like Grok to make investment decisions and see if they can beat the S&P 500. Nearly two months have passed, and the results are quite interesting:

Grok's return is 5.2%📈, while the S&P 500 has only increased by 2.5%📈.

Although the gap isn't particularly large, it highlights a phenomenon—under market volatility, AI indeed has advantages in data analysis and decision-making speed. It can quickly process vast amounts of information and is unaffected by emotions, which is often a plus in trading.

Of course, a two-month test doesn't reveal any major issues. But for those interested in AI applications in finance, this at least is an intriguing signal—AI tools are not just theoretical in actual investing.

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

12 Likes

Reward
12
10
Repost
Share

Comment

0/400

RiddleMaster

· 01-21 15:03

Doubling in just two months? It depends on how things develop later. Don't rush to get off now.

View OriginalReply0

MEVHunterNoLoss

· 01-20 09:15

Did it beat the market in just two months? I feel like this is just survivor bias in the data. What if we look at a different time period? Don't talk about being unaffected by emotions; AI can also be "confused" by training data, okay? The biggest risk in this kind of benchmarking test is cherry-picking the time window... Wait, what logic does Grok use to select coins? Can the strategy be made public? Honestly, if it were really that strong, someone would have gone all in already. How could it still be in the testing phase? Making 5% in two months and then losing 20% in the next two months wouldn't be surprising... It feels more like hype. I still trust index funds to be more stable.

View OriginalReply0

probably_nothing_anon

· 01-19 19:20

You want to explain the issue in just two months? Wait until the bear market comes and then boast.

View OriginalReply0

notSatoshi1971

· 01-18 21:07

What can be said in two months? Let's wait and see when the bear market comes.

View OriginalReply0

GasFeeTherapist

· 01-18 21:03

Jumping to conclusions in just two months? I think it's just good luck; there's a high chance you'll be proven wrong later.

View OriginalReply0

OnlyOnMainnet

· 01-18 21:03

Double the difference in two months? That's not convincing; honestly, the sample size is too small.

View OriginalReply0

WhaleMistaker

· 01-18 20:55

Think you can explain the problem in just two months? Give me a break, anyone can make money in a bear market --- Grok was lucky this time, next month 888 will directly break even --- Really? Let me try using AI to trade, feels like easy money is not far away --- Stop bragging, after a bull market cycle, let’s see who lasts longer --- 5.2% vs 2.5%, this gap is indeed significant, but two months of data is too shallow --- No emotional interference? AI has never seen the fear during a K-line plunge, haha --- Finally someone is speaking with facts, traditional investing is indeed outdated --- Wait, is this another marketing trick? I feel it’s a bit shallow --- Damn, should I go all-in on Grok? But that’s too much of a gamble --- AI can make quick money, but try letting it endure a black swan event

View OriginalReply0

OnChainDetective

· 01-18 20:54

nah wait... 2 months of data? that's basically noise. historical data suggests these outperformance windows collapse pretty fast once they become public knowledge. suspicious activity detected in the methodology tho—how'd they avoid survivorship bias here? traced through the transaction patterns and smells like cherry-picked timeframe tbh

Reply0

TokenomicsDetective

· 01-18 20:46

What can you say in two months? Wait half a year before bragging --- Grok's luck is pretty good too, what if there's a pullback later --- Data analysis isn't the same as steady profit; I still trust myself --- Another new trick of AI to cut leeks --- This thing relies on having no emotions to make money; humans can't do that --- But 5.2%, even this is my annualized rate --- Why does it feel like every new tool claims to beat the index? If it really could, it would have flown away long ago --- Wait, does this test have no trading fees? --- It's interesting, but I don't dare to go all-in --- It's the AI era. Is sticking to index funds too conservative?

View OriginalReply0