049 – David Bush – Build a High-Performance Quant Crypto Portfolio
Without Blowing Yourself Up!
NOTE: For a one-hour Crypto Masterclass with David, where he reveals one of his strategies, simply sign up to the Collective, you’ll be on a 3-day free trial (yes, you’ll need a credit card), watch the masterclass and if it’s not your thing, cancel your trial. Easy.
Crypto might attract the same kind of attention a chainsaw gets at a dinner party, frequently handled by people who should not be touching it.
That is exactly why it remains so attractive to systematic traders.
The usual retail approach to crypto is painfully predictable: find a coin, fall in love with a narrative, survive one big run-up, then give it all back when the market decides to impersonate gravity. The smarter (systematic) approach is very different. You do not build a crypto portfolio by finding “the next Solana.” You build it by creating a process that can survive crypto’s violence while still harvesting its upside, and volatility.
Many quant traders have shied away from Crypto: it’s hard, immature, structurally messy, and has limited data history. But if you think like a trader, you’ll want to go to where the easy money is. The question is, can it be done without over-fitting, given the history is short, and full of insane rides that will likely never happen again. Can we navigate the pump and dumps, the meme coins, the counterparty risk, the 24/7/365 trading without crazy risks? David Bush (like Pavel on the pod before him) answer with a resounding: ‘of course’.
One of the best rules anybody can learn about investing is to do nothing, absolutely nothing, unless there is something to do… I just wait until there is money lying in the corner, and all I have to do is go over there and pick it up… I wait for a situation that is like the proverbial “shooting fish in a barrel.” Jim Rogers (in ‘Market Wizards’)
New Market, New Thinking
David highlights the obvious but not so obvious: crypto trades 24/7/365. That means far more time for price discovery, more volatility, more dislocations, and more opportunities for systematic traders willing to treat it as a live laboratory rather than a religion. The volatility is not outsized compared to equities if adjusted for this enormous leap in time that the market’s are open for price discovery.
The second mindset shift: simple logic beats fancy nonsense
Crypto’s complexity and modernity don’t mean the solution is to become even more complex. It’s the opposite. New markets have the same profile they always did – speculation dominates; rules and regulations are lax; institutional adoption is less. This means less efficiency and more edges, if you’re willing to participate on the path to ‘full maturity’.
Both David and Pavel remind us that Crypto has limited reliable history. Bitcoin data is short. Altcoin data is much shorter. That means you do not have the luxury of building overengineered models with twenty moving parts and a machine-learning cherry on top. If you do, you are not finding truth. You are decorating noise.
David’s preference is classic logic: trend following, momentum, mean reversion, breakouts, volatility-aware exits, and simple filters that generalize well. He repeatedly stresses the need to use principles proven in other markets, then adapt them carefully to crypto.
Start with ideas that already work in other asset classes, then make as few changes as possible when bringing them into crypto. The main adjustment tends to be in the exits, because crypto is faster, noisier, and more explosive than stocks or commodities.
So the working rule is simple:
- Start with robust ideas, not clever code.
- Use as few conditions as possible.
- Make crypto-specific adjustments where the market genuinely demands them.
- Treat complexity like a tax, because that is exactly what it is.
Then there’s the bonus data:
Crypto has new data, additional opportunities. There’s ‘on-chain’ data for starters, but there are indexes and metrics that are only available to crypto traders and can definitely be used for filters and regime switches. Add this to your OHLCV data for extra risk management, and extra spice.
And we like a clean build-process that is ‘research driven’
Build edges first, strategies second, portfolios third.
Most traders want a “strategy” immediately. That is understandable. It is also backwards. First identify statistically meaningful edges or alpha engines, then derive strategies from them, and only then combine those strategies into a portfolio.
We’ve been talking about this in the Collective. How to build a ‘research first’ pipeline for ‘robust strategy creation’. Truly fascinating discussions with some legendary traders. It’s all Members only content, but super value:
Sign up here: collective.algoadvantage.io
That means looking under the hood. Trade-level analysis matters. You want to know what each edge actually does. When does it win? When does it fail? Does it work because it captures one or two freak trades, or because there is something real under the surface?
David makes a particularly useful point here for crypto: in trend-following systems, average trade can be misleading because a few outliers can drag the average upward. YES, this is normal for trend following, BUT with Crypto these moves can be so over-sized you don’t want that level of optimism going forward! David looks at median trade data. If the median trade is pathetic, the edge may not be robust beneath the glamorous surface. That is a nasty but necessary test.
Trend following and mean reversion: use both
Trend following
Trend following is the obvious fit for crypto because crypto occasionally goes berserk in your favour. These are the escape-velocity trades David talks about: lower hit rate, bigger payoff, positive skew, and the need to let winners run. That means breakouts, momentum, and trailing exits matter more than pretty win rates.
Pavel agrees, but adds an important twist: in crypto, long-term momentum exists, but often in shorter windows than traders expect. Crypto trends fast, overshoots fast, and then turns into a drunken crab. So even the “long-term” momentum trader in crypto may be operating on much shorter holding periods than an equities trader would assume.
Mean reversion
Mean reversion matters too, especially on the long side. Crypto is often a “mean-reversion-long” market in the sense that sharp selloffs frequently bounce. That makes dip-buying logic, quick profit-taking, and higher win-rate systems highly relevant. But mean reversion brings its own ugliness: negative skew, left-tail risk, and the constant temptation to keep averaging into stupidity.
Why the mix matters
The real magic is not “trend following versus mean reversion.” It is the combination.
Short-term mean reversion short strategies can help reduce the open exposure created by momentum long systems when markets become overheated. Meanwhile, mean reversion long strategies can help offset the drag from momentum short models in a market that tends to snap back upward. In other words, the styles can hedge one another if combined intelligently.
That is a serious portfolio insight. A good crypto portfolio is not just a bunch of profitable strategies dumped into a bucket. It is a set of return streams that interact in useful ways when the market gets ugly.
Universe selection: careful what you test
Crypto has a survivorship bias problem the size of a small planet.
If you only test today’s winners, you are effectively asking history to flatter you. David is explicit on this point. In his altcoin research, he did not simply take the current top names and pretend that was a historically honest universe. Instead, he wanted coins with enough data from around 2020 onward, then ranked or grouped them with market cap as a secondary filter, not the only one.
The additional solution is to trade a full portfolio with survivorship bias free data (include dead coins) and trade the ‘top 20’ or ‘top 50’ by, say, turnover. Even then, different exchanges have different coins and different liquidity. The serious trader might revert to coin market cap historical data, and I’ll soon share how to do that using their API & Python.
Pavel makes the same point from the institutional side. He warns against building models on a few coins that happened to perform spectacularly in hindsight. His preference is always a portfolio approach across a tradeable universe, not a hero trade approach based on yesterday’s poster child.
Robustness in crypto is not optional
If there was one theme louder than the rest, it was robustness.
David repeatedly argues that out-of-sample data must be treated like a synthetic live environment. If a model fails there, you do not tweak it and pretend nothing happened. You accept that the development process failed and move on. Painful? Yes. Necessary? Absolutely.
This can be extended into a much broader framework. Robustness is not just about the model. It is about the portfolio, the data, and the infrastructure:
- using simple, causal logic
- testing alternative daily closes and offsets
- making sure models work across a broad universe
- monitoring whether each model behaves as it should in the regime it was built for
The main point is that ‘robustness testing’ in crypto can’t rely on the same tools one might use in markets with decades of mature history. Hence the need to lean on simple, causal logic for time-tested models and principles.
Risk management: the portfolio is the real stop-loss
Hard stops are often a poor primary risk tool in crypto because intraday volatility is savage and false breakouts are common. The preferred approach is broader: small position sizing, many models, many positions, regime-aware exposures, and explicit hedging structures inside the portfolio.
David mentioned using Monte Carlo testing as a precursor to sizing because it shows alternative trade histories and helps define what kind of drawdown you actually might have to live through.
A high-performance crypto portfolio manages risk through structure more than prediction.
That means:
- many small positions rather than a few oversized bets
- multiple strategy types rather than one style
- liquid, tradeable coins rather than random carnival tokens
- regime filters and hedges rather than faith
- exchange diversification rather than custodial complacency
And yes, exchange risk matters. Both David and Pavel discuss it directly. Spread capital across multiple exchanges. Keep only what you need on-exchange where possible. Understand custody risk. Crypto is one of the few places where your strategy can be right and your venue can still mug you in the parking lot.
The all-weather portfolio is the end game
The ‘all-weather’ blueprint in a world of ‘loose pants’ (not over-fitting in the slightest) is combining momentum and mean reversion, long and short, aggressive and regime-dependent models, in ways that keep the portfolio exposed to opportunity while muting left-tail damage.
Build a library of functional components and then combine them at the portfolio level, where correlation, diversification, sizing, and risk-adjusted returns really start to matter. Individual strategies are not the final product. They are ingredients.
That is how you should think about your portfolio too. As a collection of behaviours.
Some models catch breakouts. Some fade overstretched moves. Some hedge. Some only wake up in hostile regimes. Some exist mainly to improve the portfolio rather than to look brilliant on their own.
Conclusion: build something that deserves to survive
If you want a practical takeaway, here it is.
A high-performance crypto trading portfolio is built on a few blunt truths:
- Crypto is still rife with inefficiency – retail trading creates opportunities.
- Simplicity beats cleverness when the data history is short.
- Classic logic usually travels better than bespoke wizardry.
- Trend following and mean reversion both matter; together they matter more.
- Robustness must come from elegant simplicity and not from complexity.
- Risk management lives at the portfolio level, not just at the stop-loss line.
- Diversification across models, coins, regimes, and exchanges is what keeps you alive.
Again, there’s a free Masterclass with David in the Collective.
Trade well and prosper!
Connect with David: