Automated Bitcoin trading increasingly difficult

Since the beginning of the year, with the highest price spike in Bitcoin so far, the market has been in decline, a bear market. This is December to date:

Bear markets, in themselves, are not a big concern for automated trading. Trading bots seek to do what people (especially me) can not: strip emotions and buy low, sell high and not chicken out until it makes statistical sense to do so. Big drops means big gains, so it doesn't matter if the market is bearish.

What matters is that the market is volatile though. Here are some sample months of 2018 in log scale.

Even in logarithmic scale, the change in volatility is striking! This makes automated trading exceptionally difficult.

Bitmex is an exchange without trading fees according to their documentation. However, when placing an order there is still a fee involved somehow, I don't know enough to tell is this is due to effective interest, deviation from their .BXBT index or some other shit.

Totally free trading would be an effective way forward, as one dollar gained is exactly that. Binance, for example, now has a 0.1% commission per trade for small-timers such as myself. This currently means that if you buy one Bitcoin on Binance, price needs to go up by 14.2 dollars to make you a single dollar. Anyway, Bitmex does not look to be free, so that route can be skipped over. Plus, there is more to gain in alt-coins normally, as they have a lot higher volatility. Binance is currently the best platform for automated trade of alt-coins, afaict.

Everything points to a crash shortly though. We're in a hard squeeze and sentiment is going down.

When fear becomes too high in a week or so, many will start selling. The Bitcoin price will settle for a few weeks 15-20% lower than it currently is and then we might start seeing some big uptrends and high volatility again. It's either that, or even lower volatility over the next coming months. I'm guessing odds are about 60-40 between these two scenarios; while others are pretty negligible in my estimation.

Any way, crash or huge upturn, it doesn't matter for by robot. As long as there is high volatility, it will do fine.

My most recent trading bot prototype just placed it's first order, buying a little Stellar. Too early to say, but at least that buy order was groovy:

Apart from implementing a new prototype, I've also spread my risks by moving into three other exchanges. I've worked 24/7 every day for the last couple of months. I deserve a little vacation, so I'm taking a couple of days off from programming. Ahhh...