Invest Without Emotions
Binance is one of the longest surviving crypto CEX (Centralized EXchange). At last count, they had around 3000 tokens listed. Just like how it is in tradfi exchanges, most of the liquidity is concentrated in the top 50% of tokens.
We use the bid/ask spread as a short-hand for liquidity.
If you want to keep your trading costs in check, then play in the top 5 deciles.
Data is the lifeblood of quantitative research and trading. The first step is to understand the benefits and shortcomings of different data sources and mapping out their use for the tasks at hand.
For example, Tiingo does a fantastic job of consolidating prices from different exchanges and presenting it through an easy to use API. While the consolidated tape is a decent starting point for developing trading strategies, you can’t trade the consolidated tape – you can trade only at a handful of venues, mostly just one.
How do Hyperliquid quotes compare with the Tiingo consolidated tape? Most of the time, the differences are within a tight range (zero mean and median). However, there are certain times when the quotes are way off even for the most liquid coins.
There have been instances where the quoted mid was off more than 10% from what Tiingo reported.
Given that there are dozens of crypto exchanges and the volatile nature of the coins themselves, some of these differences are inevitable. However, the data highlights an inefficiency and the need to have multiple exchange feeds so that you don’t shoot yourself in the foot while trading.
Code and charts are up on github.
Our introduction to trend-following posts on Zerodha Varsity (Part I, II, III, IV and V) used tradfi instruments to build a basic model. What if we applied the same principles on crypto assets?
To keep things simple, we’ll pick only two assets: BTC and ETH. The portfolio is evenly split between the two. Since crypto markets are 24/7/365, we’ll divide each day into 24 hour slots and construct a daily series based on the closing prices at each hour. The portfolio is further split into 24 parts each. Each position is an average of a binary trend signal.
Individually, trend-following boosts the Sharpe ratio of each asset.
You may not have captured the absolute highs but you would have avoided the steep drawdowns.
They are stronger together than individually.
Needless to say, leverage in this scenario would be fatal.
Code on github.
Hyperliquid is probably the most hyped DEX (decentralized exchange) right now. The advantage the CEXs (centralized exchanges) had was in the “ease-of-use” area, especially in the on-boarding and trading APIs front. Hyperliquid tries to marry the convenience of a CEX and the benefits of a DEX.
We downloaded their historical order book to check the spreads and depth of popular coins.
BTC, ETH and SOL are a world apart – tight spreads and deep books.
However, trading the less popular coins needs some care.
Zooming into the spreads brings out the difference between the top-tier coins (BTC) and the rest (PENDLE).
The ranges look decent if you are trading a retail-sized book.
Vojtko, Radovan and Pauchlyová, Margaréta, How to Improve Commodity Momentum Using Intra-Market Correlation (SSRN) discusses using short-term and long-term correlations between constituents to bet on momentum and reversal.
Since we are always on the lookout for strategies for reducing momentum drawdowns, we did a quick check to find out if a similar strategy can be used for long-only momentum.
The rule is fairly basic. Using the momentum portfolio already formed, if 20-day average correlation between them is greater than 200-day average correlation, then go long, else, go to cash.
Ignoring transaction costs, it looked like it avoided the brutal 2018-2019 drawdown. So, we dived a bit deeper to see if it was materially better than our 50-day SMA idea.
Here, COR_RET represents using only correlations to go long/cash, SMA_RET represents using only SMA, EITHER_RET is correlation or SMA and COMBINED_RET is correlations and SMA.
Going long if either correlations or SMA (EITHER_RET) seemed to be a winning strategy. However, high transaction costs in India can turn any decent strategy into a loser in a heartbeat.
25bps in transaction costs negates most of the advantages of considering the correlation signal. However, the post-COVID data does point towards EITHER_RET outperforming SMA_RET.
The biggest disappointment for us was that there was no improvement in drawdown metrics but 5% of outperformance might be worth the additional complexity.
Code and charts are on github.