You are showing performance gross of fees. How can I assess the impact of trading on my performance ?

As we have often communicated, we are presenting backtests/simulations of our performance on our NapBots platform that are gross of trading costs. This has been done on purpose to allow our users to have a universal benchmark that is not exchange-dependent. As our product is currently available on 8 different platforms, each with a dedicated fee scheme, it would be strange to present figures only for one exchange and confusing to present figures for all exchanges. As users can have several exchanges plugged in on their account, we cannot even tailor made the performance depending on the chosen exchange. 

Our strategies are theoretical strategies based on CryptoCompare prices that are averages of the representative crypto exchanges. Once again, the idea is to present figures that are not exchange dependent, even though we are fully aware of the execution component. Most of Index providers calculate theoretical indices that are not directly replicable. But these indices provide very good indications of potential (past) returns once we incorporate trading costs. That is why we have decided to launch such a simulator in order to assist you in getting a better understanding on past performance with accurate trading costs.

What trading costs should we incorporate in our analysis ? First, you have the trading fees that need to be paid to the exchange on which you are trading. Here we have deliberately chosen to execute market orders only even though they are more expensive than maker orders. The current fee levels for all our 8 platforms are as follows (and provided you do not reach certain 30-day volume levels that would open the possibility for a discount):

 

Exchange level Fee Level
Binance Futures 4 bps
Bitmex 7,5 bps
Phemex 7,5 bps
Bitfinex 20 bps
Kraken 26/24 bps*
Bitstamp 50/25 bps**
OKEx 15/13,5 bps***
Bitpanda Pro 15 bps


*          24 bps for 30-day volume above 50k USD

**        25 bps for 30-day volume above 10 k USD

***      13,5 bps for OKB holding above 500

The rationale of taker vs maker orders is as follows. First, you have to understand that every hour, we are potentially executing hundreds of orders for all our users. We have to prioritize them in order to avoid reaching the market with one big order that would come from various accounts and that would not be executed exactly at the same time as we would need to go quite deep in the order book. So, what we are doing is analysing the actual liquidity of the market when we need to send orders. We then determine an atomic optimal size. And for each user, we are dividing their target trading amount into n atomic orders.

These atomic orders are also not of the exact same size to be as discrete as possible. We then tag each of these individual orders for all our users and send them randomly at market price with a timer. Now imagine that we want to place maker orders to reduce fees and improve slippage. When the market is going up and we have a buy order, you might wait a long time before being executed if at all. We would then take the risk to not have a potential winning position. Also, as we would need to manage the same queuing process, one order not being executed means all the following ones would have to wait. There would be a lot of complaints.

We have seen some subtilities on trading fees as each platform has its own pricing policies that depend on its customer base and commercial positioning. Now we can go to the next level by looking at the slippage cost. Slippage cost is basically the difference between the actual execution of your order and the reference price (CryptoCompare) that is being used to calculate the theoretical performance. There are 2 components here. The first one is the fact that prices taken at the very same timing between on-exchange and CryptoCompare are not the same. And the second one is the fact that when your orders are effectively executed, market prices have moved away from your reference price. Once again, this is quite tricky but very important to assess the difference between actual execution and theoretical performance.

Unless there is a form of martingale, there is no obvious reason why there should be a regular discrepancy between our theoretical performance and a theoretical one that would be calculated using the exchange prices as reference prices. The impact should more likely come from the price are moving away from their reference levels when we delay the execution. But this should be of a much smaller order of magnitude than trading costs form a statistical approach as there are always noise in market prices. If we wanted to reduce this timing of execution, that means that we would need to concentrate our global orders into a shorter timeframe, and this would clearly increase slippage as order book would not have sufficient time to adapt to a big aggregated order. Overall, one should not expect slippage cost of more than 1-3bps over a long timeframe. This slippage cost is pretty much related to the liquidity of the exchange on which you are trading.
 

Allocation Performance Tool


Now you have all the tools to understand the regular differences that you will observe between our theoretical performance and the performance of your allocation with our NapBots service. Our new tool will allow you to do several things (you can find it on your exchange allocation page once logged into your account):
  1.  At the bottom of your exchange allocation page click on Advanced parameters. 
  2. Input the slippage costs of your choice
  3. Click on "Show my allocation performance"
  4. The tab displau will calculate a theoretical performance and KPIs of your own allocation. This will allow you to assess risk factors as well as the average trading gain. Remember the higher this trading gain, the more likely it will cover your trading costs (trading fees + slippage costs)

 

This way, we hope you will get a better feel for how a particular mix of strategies has behaved in the past. And you will see that sometimes, the best performing strategies on paper might not be the optimal allocation when you want to maximize your returns.
 
Was this article helpful?
0 out of 0 found this helpful

Comments

0 comments

Please sign in to leave a comment.