Trading Bitcoin for Fun and Profit: My Method, and Insight into Renaissance Technologies

For the past few months been making good money trading Bitcoin. What I do is I short Bitcoin on Sundays (using Coinbase, Deribit, and other brokers) and weekdays. After the stock market opens, at 6:30 AM PST (although I may do it 15-30 minutes before the open), I short Bitcoin and or Ethereum while also going long DJIA, NQ and ES futures (or ETF equivalents, those being SPY, DIA, or QQQ). This benefits from stocks outperforming BTC, but also stocks and BTC being highly correlated to the downside but not the upside. On Saturdays and after the market closes, at 1PM PST, I close all the positions. It’s surprisingly consistent and simple. This has been like a money printing machine that is as simple as just keeping a few windows on on my computer as I do other stuff, like write this blog, post online, watch YouTube, etc.

The way it works, in more detail, is I determine if BTC is weak or strong relative to movement of the S&P 500. If weak, I will short BTC and go long SPY/DIA in equal proportions (so, for example, $100k short BTC and $100k long SPY, DIA, or QQQ [0]). If no weakness/divergence is noted, I sit it out for the day. The position is closed at 4:00 PM PST. As evidence of superior pattern recognition abilities, I was able to figure this out. The reason why this works so well is when a large private investor or hedge fund decides to liquidate Bitcoin, they do so when the stock market is open (because that is when there is the most liquidity), and not all at once, but gradually throughout the day to try to get a better price. This liquidation is what causes the divergence, in which BTC begins to fall more than precited by just the the movement of the stock market and correlations. And then this fund may repeat it again the next day, etc., until the entire position is sold. You’re not actually trying to predict what will happen, but just riding or drafting in the ‘wake’ created by this hedge fund. Hedge funds try to be discrete when unloading large positions, but telltale signs, such as divergence, tend to emerge.

It’s possible Renaissance Technologies’ Medallion Fund does something similar to this, but scaled up by a factor of thousands and across many individual assets, not just Bitcoin. [How Renaissance makes so much money is a hotly debated topic online that has turned up no leads.] The S&P 500 has 500 individual stocks, but not all of them trade in tandem with the index; some will go up more , others will fall more, on even short timeframes. And then it’s possible, in theory, to extract from this noise the vector field or current underlying these individual stocks, like extracting individual speakers from a noisy auditorium of hundreds of people.

Here is an example from a month ago during an FOMC meeting, to get an idea of how this works:

Notice how the BITO ETF (a proxy for Bitcoin) closely tracks the movement of the Nasdaq from 11:30 AM to 12:30. Then from 12:30 to 1:30 BITO begins to lag a bit. I opened the position here, shorting Bitcoin and going ‘long’ SPY and DIA. After 1:30 the divergence becomes obvious, with the Nasdaq making new highs but Bitcoin still below where it was many hours ago at 11:30 AM. This was a profitable day for me.

Here it is again today, note the huge divergence:

Note how BITO again is much weaker. The Nasdaq keeps making new intraday highs but Bitcoin stalled out. I opened this 30 minutes before the market open and will close it before the market close. No algorithms, no data feeds, no coding languages, no APIs, etc. The pattern is so reliable there is no need to even check it. This goes against notions of market efficacy but I will not turn up my nose to free money.

In the past , pre April 2022, due to the tendency of Bitcoin to spike randomly during the day this method would not work as well, but those days are over given the extreme weakness and high correlation of Bitcoin to the downside with the S&P 500, and funds liquidating. Bitcoin is now used as a hedging mechanism against index positions, like SPY and QQQ.

[0] The way I choose which one is based on relative strength. So if QQQ is stronger relative to SPY or DIA, I buy QQQ. If DIA is stronger, I buy DIA, etc.