Most traders approach crypto the same way they approach stock picking – chasing last year’s winners, over-concentrating in one or two coins, and backtesting on data riddled with survivorship bias. The result is strategies that look brilliant in hindsight and fail in live trading.
Pavel Kycek has taken a fundamentally different path. A systematic trader with a background in equities and forex, Pavel now trades a portfolio of uncorrelated algorithmic strategies across crypto futures – with a sharp focus on robustness, diversification, and keeping individual strategies deliberately simple. In this episode, he breaks down how he approaches crypto differently from the crowd, and why he thinks the inefficiencies window is still open, but closing.
Watch the full episode below, then read on for the complete breakdown.
Pavel’s Path to Algorithmic Crypto Trading
Pavel’s trading journey started in 2007 as a typical discretionary trader – studying charts, developing a feel for price action, trying to outsmart the markets. Like most traders, this approach wasn’t profitable for years.
The shift began when he moved into semi-systematic trading: downloading end-of-day data, running scanners on mean reversion setups on e-mini equities using level two data. This was the first period he became “somehow profitable” – about four to five years in.
From there, he moved into fully systematic rotational strategies (gold and SPY ETF), then built broader systematic portfolios across equities, forex, and commodities. About two years before this episode, he committed fully to crypto – initially as an investor, then systematically when he recognised the volatility and infrastructure challenges required a fully automated approach.
“For us traders, volatility is the reason, number one, why we want to trade the asset,” Pavel said. “That’s why it took me like another two years before I started trading crypto systematically.”
Wrong Expectations: What Trips Up Most Crypto Traders
Andrew opened the topic by referencing a private Twitter exchange where Pavel had written: “I’m trying to give back some of my knowledge on Twitter, because especially in crypto trading, people have such wrong expectations.”
Pavel’s diagnosis comes down to two root causes:
- Hindsight bias: Many coins have returned 10,000% or 100,000% in the past. Traders chase the last best opportunity, not realising they’re looking at outcomes, not edges. Everyone wants to buy the next Solana after it has already run 100,000%.
- Gambler mindset: “Buy this coin because it will go to the moon overnight.” Once that mindset takes hold, proper risk management thinking becomes very hard to restore. “Once you get into this gambler mindset, I don’t think you can make any money in crypto.”
Pavel’s view is that proper position sizing and risk management – the same discipline serious systematic traders apply to any asset – is a prerequisite for trading crypto profitably, not an optional extra.
Strategy Types That Work in Crypto
When asked what strategies work best in crypto, Pavel’s answer was consistent with how he approaches any asset: don’t bet on one approach.
Crypto currently favours trend and breakout strategies because of its strong directional moves. Simple moving average crossover strategies still outperform benchmarks, though the edge has narrowed compared to three or four years ago. At the same time, mean reversion strategies – particularly to the short side – have become more effective as the market matures and overreactions from less experienced participants create exploitable dislocations.
“To me, it is really more about building broader portfolios and taking advantage of trends through trend strategies and breakout strategies, and inefficiencies through mean reversion strategies.”
He uses the analogy of mature futures markets to frame where crypto may be heading. In the 1980s, commodity traders could run simple moving average strategies and print money. Over time, as more sophisticated participants entered, those simple edges compressed. He expects a similar evolution in crypto – which is why he’s focused on building robust, multi-strategy portfolios now, while the window is still open.
For more on how strategies evolve as markets mature, see our episode with Larry Williams on 50 years of trading.
Survivorship Bias: Crypto’s Most Underappreciated Danger
Pavel described survivorship bias as having a “huge impact” on crypto backtesting – far greater than in more mature markets. The reason is structural: crypto is an exponential asset where a very small number of coins become massive winners, and the rest stagnate or go to zero.
He shared a clear example: take today’s top 10 coins and run a trend strategy backtest on them, and you get an impressive equity curve. But if you backtest using the coins that were actually in the top 10 at each point in history (not the ones we now know survived), results drop by 5 to 10 times. The difference is purely survivorship bias.
“The more volatile the asset and the more immature the asset, the more survivorship bias can affect your trading and your backtesting – especially your backtesting.”
In traditional markets, survivorship bias-free data is commercially available (Norgate for equities, for example). In crypto, Pavel couldn’t find any equivalent service, so he and his partner built their own database pulling from Binance and Kucoin. If you’re backtesting crypto strategies without accounting for this, your results are likely significantly overstated.
Robustness Testing When You Don’t Have Enough Data
Robustness testing in crypto presents a unique challenge: there simply isn’t much historical data. Bitcoin is roughly 15 years old; most altcoins have far less. You can’t run the same style of walk-forward testing you’d apply to a 40-year futures dataset.
Pavel’s response to this constraint is a multi-part framework:
- Idea-first strategy building: Every strategy starts with a clear market logic – why should this edge exist? Fitting a strategy to historical data without a reason is the fastest path to overfitting.
- Parameter stability testing: If a strategy uses a 10-period moving average, does it still perform well with 7, 8, 12, or 15? Results should be stable across a roughly 50% parameter range in either direction.
- Cross-timeframe testing: Strategies are tested on 4-hour and 12-hour charts as well as daily, to increase the number of valid data points.
- Artificial daily closes: Because crypto trades 24/7, Pavel creates 24 “daily closes” per day – the official UTC close plus 23 offset versions – and tests performance stability across all of them.
- Logic consistency check: A long trend strategy should make money during uptrends. If a mean reversion strategy to the short side is losing money during strong bull moves, that’s expected and acceptable. Understanding when a strategy should lose is as important as understanding when it should win.
He uses RealTest as his primary backtesting platform (recommended for non-programmers due to its simple scripting language), cross-checked against a proprietary backend he built with his partner. Two independent systems is a non-negotiable for serious systematic trading.
He also draws on classic strategy literature as a source of robustness: Larry Williams, Linda Raschke, traders who were active when commodity markets were similarly young and volatile. “I’m looking at commodities in 1980s. I’m looking at stocks at the start of 2000 before the tech bubble. You have to find other assets to make your robustness testing and see what was working in those times.”
Building an All-Weather Crypto Portfolio
The centrepiece of Pavel’s approach is what he calls an “all-weather portfolio” – a mix of trend, mean reversion, and breakout strategies that collectively perform across different market regimes.
No single strategy type works in all conditions. Trend strategies need directional movement. Mean reversion strategies need volatility and overreaction. The only regime that genuinely hurts across the board is sideways, low-volatility consolidation – and even then, losses are manageable because the strategies are uncorrelated.
He makes a sharp distinction between asset diversification and strategy diversification. Many traders think diversification means running one trend strategy on Bitcoin, one on Ethereum, and one on Dogecoin. “The correlation would be pretty high.” True diversification means every strategy itself trades a portfolio of coins – each strategy is a sub-portfolio – and those sub-portfolios have different drawdown timing when combined at the top level.
His portfolio construction principles:
- Maximum exposure to any single coin is capped at a low threshold – a coin going to zero overnight is a real risk in crypto
- A new strategy only enters the portfolio if it contributes genuine diversification, not just additional performance
- He looks for strategies that draw down at different times, not strategies that just have different names or different assets
- “The only thing you can basically control is drawdown, not performance, because performance is more a function of the trendiness and volatility of the market. But drawdown is a function of your portfolio building.”
He keeps individual strategies intentionally simple – often one or two entry conditions and one exit – and puts the sophistication into portfolio construction. “You can build your very simple crypto strategy in one day. But the portfolio, the proper portfolio, you will be building many months, maybe even a year.”
The Biggest Mistakes Algo Crypto Traders Make
Pavel’s list of common mistakes reads like a checklist for anyone building crypto strategies from scratch:
- Overfitting: Number one problem. Too many conditions equals a strategy that fits the past perfectly and fails going forward. Simple is more robust.
- Single-asset strategies: Building and optimising a strategy on just Bitcoin or Solana, especially when that coin has had an extraordinary run. “If you have such an exponential coin, then everything will be making money on it.”
- Survivorship bias in data: Using today’s top coins to backtest a strategy that would have required holding those coins years earlier, before they were winners.
- Not accounting for transaction costs: Fees, slippage, and exchange risk all eat into returns. Shorter-term strategies are particularly vulnerable – edges that look good on paper can disappear when real costs are applied.
- Code bugs going to live: Pavel admitted he’s traded live with buggy code himself. Testing thoroughly and cross-checking results between two independent platforms is essential.
- Falling in love with one or two strategies: This leads to over-optimisation rather than building a properly diversified portfolio. “You have trend strategies, mean reversion strategies, and breakout strategies. Every trading approach makes money just in some market phase.”
Where Crypto Is Headed
Pavel closed with a candid long-term view. Regulatory clarity is coming – Europe’s MiCA framework, increasing institutional interest, and the sheer profitability of crypto fees for banks means smart money will continue entering the space. As it does, the inefficiencies that systematic traders currently exploit will narrow.
“What I would expect is that over time, like in five to ten years, crypto will be just another asset that will be like a diversifier for your portfolio, but not this huge opportunity that is right now.”
His advice: get the infrastructure, data, and portfolio-level thinking right now – while the edge is still there.
To follow Pavel’s work, find him on Twitter at @PKYCEK or visit robuxio.com.
Get the show notes & transcript
Related episodes
- How to Build Mean Reversion Trading Strategies with Stefan Friedrichowski
- Combining Algos Using State-Based Market Design with Richard Metzger
- How to Avoid Trading Strategies That Degrade Quickly with Timothy Masters
Want more on building robust systematic trading strategies? Subscribe to the Better System Trader podcast for weekly interviews with the world’s top systematic traders and quantitative researchers.
