March 2019 Performance

jason ย ยทย  2019 Apr 3

March 2019 Performance

  • Market performance: +8.9% from 3799.12 to 4137.01
  • Trades made in March V5: 10
  • Trades made in March V4: 9 (1 pending into April)
  • V5 performance: -0.4%
  • V4 performance: -1,59% (1 pending into April)
  • V4 pending performance April: ~11.5% (more details below)
  • V5 pending performance April: ~7.5% (more details below)

Explanation

Very slow moving market with almost no volatility during all March having difficulty breaking through the $4K resistance. As I'm testing the supporting strategy now it's obvious that the sensitivity needs to be tweaked. More details below.

V5 trades V5 2019 march trades V5's last position was opened in April.

V4 trades V4 2019 trades march V4's last position was opened in March.

What can be improved?

There's nothing much to say about March, it's just very challenging if there's little growth spread out over the entire month. So we will make some losses here and there.

However with the recent volatility there are a couple things going on. I launched the supporting strategy last month with anti-crash monitoring. I also updated both V4 and V5 to not buy in during a single large pump happening within an hour. This prevents us from fomoing and buying high. This works natively within V4 and V5 and has been backtested and performs well over long periods of time.

However when combined, this created unintended situations. During the recent pump, what happened was that we bought in before the pump, then when the price stabilized and the anti-crash was triggered, it sold. Because of the parameters that we wouldn't buy in on recent large pumps, until this value settles down again, it was not able to buy in again.

Normally V4 and V5 would not have sold in this situation and would still have held the position. V5 might have sold inbetween (will share backtests with the April performance review) but would always buy in quickly again if the market keeps moving.

Because of these two implementations running together, anti-crash + no buy if we are late to the party, it prevented from profiting massively from the latest pump. We made gains, but V4 and V5 are by default designed to have open ended upsides. So the unexpected circumstances prevented it from doing what it was supposed to be doing.

Solution

I've said before the anti-crash is less suitable for V4 than for V5, as V4 needs more room to trade. I've also said that the anti-crash shouldn't be running 100% of the time but only in uncertain situations. But instead I've enabled it on both and ran it 100% of the time. Expecting that it would buy in again anyway. Sometimes I just need to listen to myself and stick to the plan. But I'm pretty sure I'm not all alone here ๐Ÿ™ƒ

As this stands in the way now how V4 and V5 should be performing, I've disabled it for now and will manually enable it for high risk situations. This will be implemented in the Master strategy in phase 2.

As for the current situation. It will buy in again once it settles down. I don't want to buy mid-trade.

The supporting strategy will keep running with its stop loss and I will add a trailing All Time High stop loss. This gives us protection for now until we can make the anti-crash more intelligent and dynamic based on market situations.

As I'm moving into multiple strategies and over different timeframes, things will get more complicated and harder to test. That's just the downside of keeping moving forward.

CoinBase Fees

CoinBase changed their fee structure and limit orders will not be free anymore. Limit orders go from 0% to 0.15% fee. Market orders go from 0.3% to 0.25%.

There's still an incentive to using limit orders but much less so. For the entire fee structure check: https://support.pro.coinbase.com/customer/en/portal/articles/2945310-fees

After this update there were some bugs with orders not being filled on Coinbase. This has been fixed.

Development Update

I'm running the phase 1 Master strategy in production now on my own accounts. Still needs some work but we'll get there.

Jason