So you've come up with some indicator or bot algorithm to make money reliably over time on some asset. You've built an algorithm to calculate gains and losses. You plot the asset price and your wallet development over time (in finplot). Then you tweak it until it can cope with futures instead of just spot trading to make you faster gains. Finally you pour your wisdom into it to make it stable over time.
Now you're pondering what margin to use. So you plot your wallet with different, but static, leverage. Here's what that might look like if you find something really awesome.
First a word of non-financial advice
Always use isolated margin, i.e. don't use leverage on all of your capital. Prefer something stable over something that jumps and plateaus. This something must show exponential growth in your backtesting — which means it forms a straight line when you plot it in a log scale as above. Be prepared to take a loss, as your chance of losing money within a day is something like 96% and at any point in time in the next three months is 99.7%. That certain loss has, of course, next to nothing to do with your long term gains.
Leverage is an amplifier. If you're doing it wrong, it's going to take you quicker in the wrong direction.
So start your bot with low leverage, then raise it in a couple of steps until you're sure it follows your backtesting. Even in promising cases, your bets are going to be off sometimes. The black spots in fig. 1 are 10x liquidations in my Twerk algorithm. Liquidations are when your whole bet is going to be eaten by the exchange, so never bet more than 60% of your wallet.
Setting your leverage
1x is too low, 20x is too high. Most long-term algorithms are not going to be stable with more than 3x–7x. Also know where you are in the cycle. Sometimes you're going to win, but 47% of the time you're going to lose. Don't set your highest leverage until you are in a good place in your cycle. My cycle varies between one and three months. Right now I'm at a high, which means I should lower my leverage so as to dampen the upcoming crash. (And yes, "crash" is the right word for a 90%+ drop in value.)
Alright. Now focus on the pink circle in fig. 1. That's when 2017 turned to 2018 and many people's BTC dreams were shattered by whales. The price was fluctuating wildly, and trend-following, short-term trading was near-impossible. (Scalping was however excellent.) Notice that 10x and 7x in Twerk would have obliterated 93% of my wallet. Sure it's possible to get that back with 14x, but take it from me: the feeling of losing everything is baaad.
But when zooming in you see that 4x and 2x are not that different from one another, and they didn't do so horribly bad. At least in comparison!
The reason is that they've not been liquified. So if you don't fall for fomo you can always go for a low leverage, and soon enough you'll be rich anyway. I, however, am in the try-to-get-super-rich-super-fast camp.
Which means I've lost hundreds of thousands of dollars. Several times. And is still down 70% from my initial investment after three years. But this time it's different! :-} No more playing around. Tomorrow I'll reduce my leverage to 2x.
