Reading the orderbook
After you’ve spent quite as many hours staring at futures orderbooks as I have, you learn a thing or two. Firstly, it makes a big difference what asset you’re trading. Whether equities on NYSE and you’re having to keep an eye on 25 levels of price each side or futures on the merc and you’re looking at maybe 10 levels.
You can also predict a lot about which way the price is about to go, just looking at the orderbook. Hell, this used to be an interview question at DE Shaw & Co., maybe it still is.
Here’s how it works:
- When there are more buyers than sellers showing real interest in the market, then the price is likely to go up
- Makes sense right? Lots of people want to buy, the market realizes they do, so those setting prices either move them up, or the buyers get impatient and push them up by taking from the book
On the CME, this doesn’t happen often and won’t last long — there are a lot of very smart and very fast players there who catch the move and trade appropriately. This is less so on some crypto exchanges.
As you may have guessed by now, there’s a lot more to the story than just imbalance and our friends over at Globe Research have crunched some numbers, used some actual artificial intelligence and have some street knowledge to drop:
- Worry about crypto orderbooks up to a depth of about 10 levels
- Only the last few minutes really matters
- Even now on spot exchanges like Coinbase you can correctly predict the price one minute in the future far more often than most wizards
The market as a whole learns pretty quickly, but there are often big, slow moving players you can outpace and out-trade.
For those of you with more active twitter accounts than GPUs (and didn’t get scammed hard buying the 2080 TI just before the 3000 came out), you can just follow a twitter bot that is retraining daily and does what’s in Globe Research’s paper. If you don’t like it then just download the code and do it your way.
Trade smart, and whatever you do, don’t trade against the house
