Clarity - Q1 Accuracy and More Free Money
And a smidge of CDLX/CVNA
If you’d like to skip reading and go straight to supporting me by taking my money, go here:
https://www.clarity-markets.com/sponsoredPredict
Otherwise I have an article for you.
As you may know from previous notes here and on Twitter I’m working on a project called Clarity Markets. The goal is to pay users to forecast information relating to the economy and stock market. Instead of paying real money like you would on Kalshi or buying stocks directly, you pay in time and hopefully get decent enough monetary rewards for it.
We ran an experiment asking for revenue predictions for the Mag-7 and CVNA/CDLX in Q1. This post is sharing those results which took awhile since NVDA likes releasing way later than everyone else as I leave for a conference and get bogged down with other projects.
Before we begin two quick updates:
I added CVNA shares/calls a bit below $100 as mentioned on Twitter. Not a huge add, but they’ve been doing quite well QTD per
on the growth side and looking for them to continue that momentum into Q3. I’ll give some pre-earnings updates on them as appropriate when we have more color on Q3 trajectoryEither CDLX is one of the most mispriced asset in public markets or I’m a moron. We’ll leave this here to look back on in 12 months once AmEx is launched (in theory). I haven’t added more since trimming CVNA at $120 for CDLX around the current price. The podcast I did with ChitChatStocks will be available I believe next week, so hopefully you enjoy it and feel free to DM with any follow ups.
Clarity Q1 Results:
Drumroll please:
User predictions and sell side predictions for revenue versus actual results were as follows:
AMZN 0.00%↑ -0.7% users vs -0.5% sell side
CVNA 0.00%↑ -11% users vs -13% sell side
MSFT 0.00%↑ -1.8% users vs -1.6% sell side
CDLX 0.00%↑ +7.6% users vs +6% sell side
TSLA 0.00%↑ +4.7% users vs +9.5% sell side
AAPL 0.00%↑ -1.1% users vs +0.3% sell side
GOOGL 0.00%↑ -1.6% users vs -2.3% sell side
META 0.00%↑ +0.6% users vs -1% sell side
NVDA 0.00%↑ -2.2% users vs -7.4% sell side
Overall: 3.5% average error for users vs 4.7% average error for sell side.
The majority of the outperformance of course was TSLA and NVDA, with CVNA also favoring you, the lovely reader, as well. The CVNA performance perhaps indicates you guys are doing something right as it’s pretty neatly tied to transparent KPI’s I follow here and on Twitter since ASP’s and volumes are provided via
Overall a great result! The scale is still small with predictions on any individual stock being 10-20 users. The best user had an overall error of 2.1%, significantly better than everyone else and sell side as a whole, so kudos to them.
Winners have emails by the way asking how they wish to be paid (two have not responded)
I consider it a great first quarter in that experiment and it hit the hypothesis quite nicely that sell side estimates aren’t super accurate, especially for companies harder to cover. As a result, the current plan is to expand the experiment in a couple ways.
1. Sponsored Markets:
I’m launching a sponsored market section at Clarity that you can find by clicking this lovely preview:
The goal of these campaigns are to cover items that don’t fall under the core “predictions” umbrella which for now I plan to continue just having Mag-7 and CVNA/CDLX/whatever I’m researching. These campaigns can be sponsored by 3rd parties (such as the current inflation and UK market) or just my own tests (the equity 1 shot experiment).
I highly encourage any readers to go submit estimates on these campaigns, as it takes seconds to upload an estimate, supports the project, and you get paid for it if you do well. Current prize pools per quarter are something in the realm of $1500 which I hope to expand over time (either CDLX does well and I free up personal liquidity, or people sponsor markets).
If you are interested in sponsoring a market feel free to reach out! You can sponsor existing markets (add $ to the user reward pool) or create your own custom questions. As a sponsor you would receive a CSV/email/API with response data and hopefully learn something about the question you’re asking. Curious how a stock you follow is doing? Sponsor a market and get other people to give you an update.
Still very early days here, but hoping to scale up the operation meaningfully with college investing clubs and the like. Would love to have you as a forecaster or sponsor partner.
2. Mag-7 Runback:
While not currently live the plan is to again ask for forecasts on Mag-7 and CVNA/CDLX plus one or two other companies I’m dipping my toes into.
This time around instead of just revenue, it will include some core KPI’s to forecast as well. Examples would be CVNA unit sales, CDLX billings, AMZN AWS growth rate, etc.
If you think (or know) you have an edge with forecasting those KPI’s, you will be able to submit estimates for them to hone in your overall accuracy.
Again, if you’d like to sponsor that market (I currently fully fund the rewards) you can do so and receive user estimates on those KPI’s, paying them for their hard work.
And if you’re a user, they should be live sometime in the next week or so as I work with my engineers on cleaning everything up. I encourage telling your friends as while they may add competition, more forecasts is easier to translate into more sponsor dollars.
I’m quite bullish on this idea for quite a few reasons I’ll go into in future articles where it could be interesting for both retail traders and funds to participate as sponsors and forecasters. If you build up a history of doing well, I can also connect you to any sponsors that may be interested in chatting.
Overall:
While we had good success in Q1 with user predictions, I look forward to ramping that up in Q2 and beyond. If you’d like to chat on this initiative feel free to reach out. I will also try not to inundate you with asinine content and make the initiative interesting even if you don’t feel like directly participating, examples being perhaps crafting and posting public trades on user estimates similar to Dalibali on Twitter and his Twitter shorting polls. I will also continue to cover stocks and work through some ideas I’ve been sent in the later half of this month, however again I will likely be biased towards CDLX given I’m quite bullish on that opportunity.
Cheers and thanks for reading. The next article might be about steak, debt, and other such fun topics.