https://www.ftportfolios.com/Commentary/EconomicResearch/2025/1/8/the-sp-500-index-in-2024-a-market-driven-once-again-by-the-mag-7

In most years, most stocks underperform the S&P 500. So a good base rate is probably around 40% (which feels close to the average given above). However, the Magnificent 7, particularly Nvidia, is an exception. Mag 7 stocks, due to their size, to a large extent determine whether the stock market goes up or down.
Can this exceptionality bring Nvidia's rate up from 40% to 60%, even though it is trailing the S&P 500 by 3% so far? I don't know much about the stock market, but I don't think so. Most AI developments were factored into its stock before 2025. Additionally, the asymmetric upside vs. downside (i.e. it can gain unlimited value, but it can only lose its whole price) means the probability of it going down must be greater, to offset the chance of huge gains.
If I were estimating the odds of this market by myself, I would put it at 40%. However, given the community prediction, I will update my subjective probability to 58%.

Most AI developments were factored into its stock before 2025.

I suspect that is not true

@Bayesian This is basically off topic but I find it interesting:

I think of Rationalists and Economists (Hayekians specifically) as being on 2 ends of an epistemological spectrum. This market is a great indicator for that.

In traditional Economics (and modern Finance, no arbitrage principle etc.) all Stocks have the same EV basically, market is efficient and everything is a risk/reward tradeoff. There's a certain "humility"[0].

Among Rationalists I get the sense; after thinking through something from first principles, there's a great confidence in the result of those conclusions. They maybe don't weigh market indicators,[1] accumulated knowledge, and illegible signals as heavily.

I'm not sure if I'm being clear...

There's a lot to say for both approaches but I see this market as being a little more on the rationalist side

[0] Not that the ppl are more humble, obv

[1] I know, prediction markets are a thing. But I think there's still a point here

@FergusArgyll Two points here:

1. All stocks having the same EV doesn't imply all stocks having the same probability of outperforming the S&P 500. As an oversimplified gesture: if a stock has a 60% chance of doing 10% better than the S&P 500—returning 10 more cents on the dollar per cent-on-the-dollar returned by the S&P 500—and a 40% chance of doing 15% worse—returning 15 fewer cents on the dollar per cent-on-the-dollar returned by the S&P 500—then its expected yield in dollars would end up equal to the S&P 500's. So the stock market being efficient shouldn't have strong implications about the probability on this market; they're measuring different things.

2. My understanding is that large numbers of finance jobs are in fact predicated pretty precisely on the stock market not being perfectly efficient. On there being alpha to be grabbed by seeing patterns in it which the market as a whole has so far missed. That's what hedge funds are doing, that's what HFT companies are doing, et cetera. The market being hard to exploit—hard enough that most people who try will predictably fail—doesn't imply it being impossible to exploit, and I'd be surprised if this weren't generally understood within the finance world.

@Tulip None of what I wrote was meant to come across as "lol you guys are idiots" and I hope it didn't....

1) Is obviously a good point

2) Sure, but is that what you think is going on here? The work a hedge fund etc. puts into finding inefficiencies pays for finding them. HFT stuff especially, I doubt anyone here is running sophisticated models to bet on this question. As you say "hard to exploit but not impossible" I would think that "Hard" work is not being done here. I was responding to Bayesian who said he doubts ai developments are factored in to the price. That's more the mindset I mean.

To be extra clear, I was planning on writing pros and cons for each approach but decided that this isn't substack. There's clearly a lot of value for both to learn from each other if they haven't done much of it yet.

I myself haven't really made up my mind on this market. I had a limit NO at 70% when it was above that but I'm not much more confident then that and haven't bet since. I also get the feeling the Markets are not fully pricing in AI developments.

Anyway, I think my understanding is directionally right. First Principles vs conservative (lower case c)

Are we tracking the S&P500 index or an S&P500 ETF, and if so, which one? Small difference but it would be important to clarify

@SimoneRomeo please check the Metaculus question for fine print - since this resolution is based on theirs, you should stay up to date on their details.

@WalterMartin we use spy then

Currently trailing by about 16%: (+3.5 vs. -12.5)

https://finance.yahoo.com/quote/SPY/performance/
https://finance.yahoo.com/quote/NVDA/performance/

One under-appreciated thing even in this article (https://youtubetranscriptoptimizer.com/blog/05_the_short_case_for_nvda) is how DeepSeek makes interconnect between chips less important. This is something everyone except for Google has had a hard time matching Nvidia on.

With this as less of an issue (and the other things mentioned in the above article!), I expect relatively more of the profits from hardware to go to TSM/ASML and the hyperscalers.

YTD:

NVDA 🔻10.5%

S&P 500 🔻0.5%

@MaxGhenis Manifold Markets: this is good for NVDA

Deepseek R1 made Nvidia tank 15%. Don't see any correction in price here so I'm buying NO

@Moneyfolded Still got 11 months left, let’s see what happens

Level of Tariffs on TSMC seems relevant given Nvidia's heavy use of their fabs.