Thread by @ltrd_

[1/n] Friday’s night crash was very interesting because there was no big news at this moment and some of the instruments fell about 50% within two hours. I would like to break down the microstructure behaviors and do an overall analysis from that day. 🧵

#trading #cryptocrash

Firstly, let’s take a look at the macroscopic view of the market. On the histogram, we can see the changes from the highest price to the lowest price within the first couple of hours of June 10th (UTC). Big moves around the whole crypto space.

Some of the instruments fell about 50% and GTC fell over 73% due to the fact that during such crashes, liquidity is poor. This example can show you how liquidity is important when there is huge volatility in the market. Unfortunately, this move happened during the weekend.

Before we move to the market microstructure, let’s look at the performance of the coins during and after the crash. You can see here by the slope of the regression line that the more the coin lost in the crash, the more it rose after the crash.

In this kind of analysis, it is important to see what are the outliers here. Is this some basket that performed better than the overall market or it is a coin that performed very well in the last days? It is a moment when you have to dig deep down into particular examples.

Let’s move to the worst coin during the crash - GTC On the plot you have liquidity for the first 25 bids and asks with CVD (cumulative volume delta) and the price of GTC. You can see that before the big crash, liquidity vanished on both the bid and ask sides.

Here is the important lesson - in those moments your algorithms can have the best predictions but simultaneously the worst performance due to the lack of liquidity. If your algorithm has to do a rebalancing here, it is worse than a nightmare to do it in those market conditions.

Let’s look at the data. You can see here the impact of the market order for $5k. Before the crash, the cost is marginal in comparison to the cost during the crash and even an hour after the crash. If you do not consider it during your analysis, then you do a big mistake.

Here, you can see the impact for BNXUSDT. Doing a $5k market order can rise your costs by almost 5% in the worst cases.

The last thing that I would like to show you is the net volume (buy - sell) on a couple of exchanges for BTC. Binance perpetuals still dominates the market, but OKX and Bybit are much more important than were before.

I hope that you enjoyed the analysis. If you have any questions or you would like for me to do some next steps related to this crash - please comment before this thread.