Project Tradebot III: coin pairs and exchanges

Wednesday 06 April 2016

In this instalment, I’ll be looking at which coins the bot will trade, and the best place to do it.

To recap quickly, I’ve decided to make a grid-trading bot. This will place lots of little buys and sells at intervals. Whenever a bid is met it will open a sell for the same amount slightly above that price, and when an ask is met it will open the original sell again. That way, I hope that I can benefit from relatively small market movements, and lots of little profits will add up. I also don’t have to worry about predicting trends or the direction of the next move.

In terms of which pair I’m going to trade, what I’m looking for overall has several desirable properties:

  1. High volumes. I want there to be lots of trades in the given market, because the more churn there is, the better. There’s no limit to the number of times a trade can be opened and closed - this is the benefit of doing things automatically.
  2. At least some volatility. A completely static market won’t be profitable. Fortunately, most crypto markets move around a fair bit, but that might count out bitcoin.
  3. I want something that - within reason - trades within a given range, because if the price moves outside my grid I won’t profit (and may lose money long term). 
  4. The exchange I use should have low fees. Every tenth of a percent will count here.
  5. I need to be able to access the exchange easily. Most crypto exchanges have APIs to enable developers to build applications on them, but some are more complex than others. My guiding principle of KISS (Keep It Simple, Stupid) means I’ll be looking for something that makes my life as easy as possible.


R2D2, in his long sabbatical between episode VI and VII, had a profitable career as a tradebot

I’m deciding against trading bitcoin/USD for now. At the moment, bitcoin is relatively stable (for bitcoin), which I could deal with but it’s not something that’s likely to stay the case, so I might end up setting it up for low volatility but getting hit by a big move. Fees on bitcoin exchanges also tend to be higher, perhaps 0.5% vs as little as 0.2% or less for some altcoin exchanges. Although alts are more volatile, I can take that into account at the beginning by using a larger number of more spaced-out tranches (which wouldn’t be profitable for bitcoin).

In terms of exchanges, it’s going to be Poloniex. Polo is currently the go-to exchange for alts, and they haven’t had the security problems that others like Cryptsy and BTER have. Bittrex may work, but the volumes are on Polo, and that's critical for this project. Plus, it looks like their API is pretty well documented. So this is an easy decision to make.

Now, for the right crypto to trade. My initial instinct is to go for NXT, because I’ve been a part of the community for a while and know the market reasonably well, as much as anyone can. I’ve also got plenty of NXT I picked up at historically low prices, which puts me in a good starting position. But NXT is a low-volume coin, and that’s not much good here. NXT sees maybe 30-50 BTC of volume on an average day, which is nothing in the grand scheme of things.

Ethereum is the obvious choice in terms of volume - Poloniex sees massive amounts traded every day, currently around 50,000 BTC or $20 million. That means lots of churn and a nice fast turnover in orders being met, but… it’s Ethereum. It’s just spiked in price, the future is unclear, and I have no sense of what to expect from it. The grid trading method is pretty tolerant, but it won’t work if ETH’s price gets cut in half, or doubles, and neither would surprise me in the slightest. I’ll bear Ethereum in mind for the future, then, but right now it’s a no.

It looks like Bitshares is the right compromise. BTS does good volumes - up to 3,000 BTC a day on Polo - and although it’s somewhat unpredictable at the moment, it’s not so wild that I’m likely to run into insurmountable problems. In any case, during the testing phase I’ll only be using small amounts of money. If it works, I can add more, or let it compound slowly over time. I’ll see when I get there. If I get there.

comments powered by Disqus