Recently one of our users said on our Telegram channel that "the bot lost me money". The situation of this user, Ann, is that her dollar amount had gone up by about 10%, while her bitcoin amount had gone down by about 10%, simply because Bitcoin had gone up 20% during the same period.
So let's discuss what an investment managed by a trading bot like Airbag is, and what "winning" or "losing" really means.
The best way to understand Airbag is as a derivative of Bitcoin. This derivative has lower volatility than Bitcoin, and lower expected return.
- When Bitcoin goes up, Airbag will go up in $, but less (so you lose a bit in BTC value)
- When Bitcoin goes down, Airbag will go down, but less (so you win a bit in BTC value)
Over time, Airbag aims to make more wins than loses, and so the return should be higher per unit of risk.
One way to think of Airbag is as an altcoin. When you look at your portfolio and your Bitcoin has gone up 10%, and your Monero has gone up 5%, you don't automatically think badly of Monero. You are happy for all your investments, and hope that at some other point Monery will be the one winning over Bitcoin. The same with Airbag, with the main and fundamental difference that it will have a lower risk than most (if not all) altcoins.
Airbag should not be your only investment. It should be part of a healthy portfolio of assets, where you don't want to be overexposed to any single one. Airbag provides a hedge against the bad times of crypto. If you are HODLing all your other investments in cold wallets, you might be overexposed to a hard time (just like during 2018), so having a derivative that prevents counter-market is a very reasonable thing to do in case things go south at any time.
Unreasonable Expectations and Measuring performance effectively:
Many people only speak about returns. In the case of Ann: "It lost me money"(when she meant "Bitcoin increased in value faster than Airbag"), it is not taking into consideration the second variable, risk:
- Airbag has lower volatility or risk than Bitcoin. It is an unreasonable expectation to think that you can have both lower risk and higher return. You can have more risk and more return if you use margin trading, for example. In the case of Airbag it is the opposite, we have lower risk. The same way you don't compare the performance of Apple's stock with German Bonds, because they have different risk levels, it is not reasonable to compare the two directly on return. You need to find assets with similar risk to measure performance, or modifiers that take risk into consideration.
- Performance needs to be measured over a long period. This user had used Airbag for about 1 week, during which Bitcoin went up. Had she used the product 2 weeks earlier, when Bitcoin went down, her experience would have been different. For any investment, it is important to consider long term performance.
This leads to the topic of fair comparisons. If people are going to compare Bitcoin's performance with Airbag's performance, how do we create a fair comparison between them? Adjusting for Risk.
We will soon release this graph on the Airbag App, which would be a better measure of performance over time, adjusted for risk, using Sharpe. Each Sharpe unit will be the average of the previous 21 days, making it a fairly stable metric.
To understand why Sharpe is a better metric even for those who care about just "more return", we need to go back to fundamentals of investment theory and the Capital Market Line:
Let's say Bitcoin as an Asset has a 6% return, and a volatility of 6% too. While Airbag has a lower return, 4%, with a lower volatility of 3%.
If you only care about Return, you could argue that you would rather invest in Bitcoin. However, finance theory suggests that if you care about return and you are willing to accept the 6% risk, what you should do is to invest double the money with the Airbag Asset (turning 3% risk into an equivalent 6% risk), which would give you a higher return of 8%.
You can read a long explanation of our deep learning technology here.
We are still early in our development but we hope we can over time help people understand better that performance needs to always be measured in terms of risk. If you have any questions, ideas or suggestions, please contact us or join us.
