Read the Tape gives players the same 5 S&P500 stock charts per day to predict. You select low, medium or high confidence and then call the chart UP or DOWN. It's a 1d chart which then resolves over 5 days. Alpha is scored against the Monkey Index, a basket of 11 random coin flips at low confidence which provides a tangible win/lose condition.
We're two weeks in and some interesting data is being kicked up. Players like to call tops even though stonks go up- 60% of the 70 charts so far resolve higher, players' down calls have only been right 31% of the time. There's a full stats dive at https://readthetape.cc/notes/tape-report-1
Your feedback and thoughts is most welcome.
This is fascinating.
That said, if it's possible to do better than random guessing, then does this reflect the fact that the five charts are presumably hand-selected to be "interesting"?
My naive guess, and I'd be very curious to learn if this were wrong, is that something very close to the efficient market hypothesis is true; that, if it were possible to beat the monkeys on randomly chosen stocks on random dates, then someone would have figured this out already and deployed bots to capture whatever profits are available.
I ended up generating alpha so this is clearly a great game and I'm a genius investor (compared to random monkeys). When I lose it all tomorrow, my opinion may of course change.
Seriously though, this is a clever idea and I'm interested to see if I can consistently beat the monkeys. Time will tell! Thank you for sharing!
> Read the Tape gives players the same 5 S&P500 stock charts per day to predict. You select low, medium or high confidence and then call the chart UP or DOWN. It's a 1d chart which then resolves over 5 days. Alpha is scored against the Monkey Index, a basket of 11 random coin flips at low confidence which provides a tangible win/lose condition.
Your description here and on the website is not clear to me about what I am predicting. Am I predicting whether the stock is going to continue to go up the same day? Is the chart of one day and I am predicting whether it will go up or down the next day? In my opinion, you need to get the explanation of what I'm looking at and the directions of what I am predicting down to one clear sentence.
Yes I think you're right, it's unclear. I'm hopeful the first reveal explains the dynamic as the player goes. Maybe something like- each chart shows 60 daily candles of a real S&P 500 stock, a past window with the ticker and dates hidden, you call if the stock closes higher or lower 5 days later?
It would find it clearer if:
- “up” and “down” was explained. If I bet “down” I’m betting my stock will underperform the monkey index? And if I am correct, I will make money (short the asset) or simply loose less than if I’d invested in monkey?
- loosing less money was green (win) when monkey is down even more.
Thanks for surfacing. Up or down is a call on the stock price alone, nothing to do with the monkeys. Down means you think its close is lower in 5 days. The monkeys make a call up or down on the stock price too, and then their pnl is subtracted from yours.
So you can lose money on the trade and still go green because the monkeys lost more, or make money and go red because a monkey made more. The idea is it's your edge over random. Hope that makes sense and thanks for raising, will strive to make it clearer in-game.
Technical analysis is astrology for boys.
Thinking there are straightforward, simple algorithmic strategies to beat the market requires one to also believe that the billions of dollars behind large investments for some reason doesn’t care about using these same strategies to make money.
“They have different goals”, I hear. Literally nobody’s investment strategy involves passing up above-average returns for low risk.
There are plenty of known effects that are relatively simple to automate
Holding overnight risk yields superior risk-adjusted returns. Buying end of month and dumping few days into new month. There are dozens of such effects
These have mechanical reasons for their outperformance - overnight risk has to do with how borrow interest rates for equity markets are calculated and firms unwilling to hold unhedged exposure overnight. Longing EOM has to do with people getting their paycheck EOM and auto buying index funds, and so on and so forth.
Systematic trading isn't magical, it's identifying these kinds of (often simple) effects and building a portfolio of them.
As an individual investor you actually have a huge advantage over large institutions in that your portfolio is nimble and easy to get out of.
As a relatively simple exercise - consider a hypothetical portfolio that's simply long SPY for the year. Could you identify _one_ day in which you'd rather be flat? The answer is probably yes, and the reason you can do this (and not a billion $ AUM fund) is that rotating in and out of positions is cheap for you.
When news about Iran hits the tape, who do you think can exit their positions faster? Joe Schmoe with $30k in his brokerage account, or Citadel with a $100bn position?
I call it money-astrology.
It’s a fun game and it throws funny shade for getting the answers wrong, but who would have guessed and sized these answers right based on this sparse data without any additional context? There’s barely enough data in the charts to do TA properly.
If I was going to add 3 more indicators, what should they be? And/or more history?
Well the map is not the territory. Just because there’s been a big gap down on a chart that wasn’t the chart doing that, the chart displays that as some kind of causal event like an earnings release or something the company has done to make the market react. So I don’t know what kind of Quant can look at that sparse TA data knowing nothing else about the company and make those calls accurately or know how to size the positions. One random money in a crowd will randomly do better than trying to sensibly do TA with this data.
Maybe Volume would be useful though.
Volume is on there but not shown until toggled on. Yeah making it blind on company and broader macro/market movements is what produces some of the the jeopardy, and outperforming the random monkeys turns out to be a bit tougher than some players might be expecting.
To me the description was very clear, the game is fun to play. Would love to see the "prediction" of a LLM in the stats after playing
Nice idea, maybe when there's enough traction for weighted scoreboards I could enter some models
you should probably rename it since reading the tape is the opposite of charts :p
Oh boy, I know more than a few guys who will get a kick out of this, hah
Appreciate you spreading the word, thanks
Needs a way to copy and paste text to share with friends. I feel like this is what made wordle viral
The config is currently if you're on desktop, you get text copied to the clipboard. On mobile, you get text + an image to share. Sharing points are 'Challenge a friend' and 'Share result' on the Performance Review, or you can share your career card from the Career Record. I added the image to the mobile shares as early users were screenshotting for x eg https://x.com/investingidiocy/status/2075471303809659212
you should include volume
It's under the chart, toggled off by default. 3rd comment so very obviously needs to be made more prominent, on it.
would love to see Volume on the chart
On there as a toggle off by default, under the chart. You're the second to not notice it so design feedback duly noted
The interesting answer by the author, will_asouka, has been marked 'dead' for some inscrutable reason.
Thanks for flagging, summary is: Charts are from a seeded random draw. And the instinct is pretty spot on players state ~74% average confidence but hit ~54%. Accuracy is basically flat across confidence brackets. Up only strategy quietly beats the coin flipping monkeys (stocks generally go up and to the right) but players call down 43% of the time.
They're from a seeded random draw. Random S&P 500 stock, random 60-day window that's at least 12 months old, every player gets identical charts. Gemini writes the narrative vignette.
Your instinct is pretty spot on. 6,000+ calls in: players state ~74% average confidence but hit ~54%, and accuracy is basically flat across confidence brackets. Up only strategy quietly beats the coin flipping monkeys but players call down 43% of the time.