Trend following has worked better in crypto than in almost any other asset class, and the reason is structural rather than mysterious. Understanding the structure also tells you exactly how the strategy fails.
Why crypto trends
Four properties combine to produce unusually persistent trends and unusually violent reversals.
No valuation anchor. An equity has earnings, a bond has coupons, a currency has a rate differential and a trade balance. Bitcoin has none of these. There is no fair value to mean-revert toward, so nothing arrests a move except the exhaustion of flow. Prices can trend far further than any fundamental-based investor would predict, in both directions.
Retail-dominated flow. Crypto participation skews heavily toward individual investors, and retail flow is more momentum-driven and more reflexive than institutional flow. Price attracts attention, attention attracts buyers, buyers move price.
Continuous trading with no circuit breakers. Equity markets close, and closing is a coordination device — it interrupts a cascade and gives participants time to reassess. Crypto trades 24/7 with no halts. A liquidation cascade can run for hours without interruption.
Reflexive leverage. Much of crypto trading happens on perpetual futures with high leverage. Falling prices trigger liquidations, liquidations are forced market sells, forced sells push prices lower and trigger more liquidations. This makes the downside moves faster and deeper than the fundamentals-free upside would suggest.
The result is an asset class with long persistent trends and drawdowns that have exceeded 70% multiple times. Both come from the same structure.
The two-asset problem
Our crypto universe is Bitcoin and Ethereum perpetuals, with daily history from August 2020. Two assets is a real constraint and it changes what kind of strategy is even coherent.
With 500 stocks, the interesting question is selection: which of these do I own? With two highly correlated assets, selection is nearly meaningless. BTC and ETH move together most of the time; picking the stronger one is a marginal decision.
The interesting questions with two assets are timing (am I in at all?) and sizing (how much?). Both templates in the crypto catalog are built around that, and the first of them turns on a distinction that is easy to state and easy to get wrong.
Absolute versus relative momentum
Relative momentum ranks the universe and holds the best. With two assets, that means always holding at least one of them.
Absolute momentum asks a prior question of each asset independently: is this trending up? An asset that fails is not held, regardless of how it ranks against the other.
In a 500-stock universe the difference is modest, because the top-ranked names in a falling market usually include something genuinely rising. In a two-asset universe the difference is the entire strategy. A relative screen on BTC and ETH is fully invested through every crypto winter, holding whichever fell less — which in 2022 meant holding through a drawdown exceeding 70%.
Gary Antonacci’s dual momentum framework makes the general case: the absolute filter, not the relative ranking, is what does the drawdown work. Cash is a position. In an asset class that draws down 70%+, it is frequently the correct one.
The template requires two conditions independently — price above its 100-day moving average and a positive 90-day trend — and holds nothing that fails both. There is no requirement to hold anything at all.
Sizing: BTC and ETH are not equal risks
ETH is structurally more volatile than BTC. Allocating 50/50 in rupee terms is therefore not 50/50 in risk terms — ETH contributes the larger share of portfolio variance.
Inverse-volatility weighting sizes each position by the reciprocal of its own recent volatility, so the calmer asset takes more capital and risk contribution is roughly equalised. In a two-asset book this typically produces something like a 60/40 or 65/35 tilt toward BTC, varying as relative volatility moves.
This is the same principle behind risk parity in a multi-asset portfolio, applied at the smallest possible scale.
The Indian tax reality
No honest discussion of crypto strategy for an Indian investor can skip this, because it changes which strategies are viable rather than merely reducing returns.
Under the regime introduced in the Finance Act 2022:
- Gains are taxed at 30%, plus applicable surcharge and cess.
- No deduction is allowed for any expenditure other than cost of acquisition.
- Losses cannot be set off against any other income, and cannot be carried forward.
- 1% TDS applies on the transfer of virtual digital assets above threshold values.
The loss set-off rule is the one that reshapes strategy design. In a normal tax regime, a strategy with a low win rate and large winners — which is what trend following is — benefits from offsetting losses against gains. Here, every winning trade is taxed at 30% and every losing trade is simply a loss. The effective tax on a trend-following strategy’s gross profit is therefore substantially higher than 30%.
The 1% TDS compounds it by scaling with turnover rather than profit. A strategy that turns over its book frequently pays 1% of notional each time regardless of whether the trade made money.
Both push in the same direction: lower turnover. A weekly-rebalanced strategy is far more exposed to TDS drag than a monthly one, and any high-frequency crypto strategy is essentially unviable for an Indian resident. If you deploy this, the first modification worth testing is a slower schedule.
Other honest caveats
The sample is short. Daily history from August 2020 covers roughly one full cycle. A strategy backtested over one cycle has been fitted to one cycle. Treat the result as a plausibility check, not evidence.
Perpetual futures are not spot. These instruments carry a funding rate paid between longs and shorts to keep the perpetual price anchored to spot. In sustained bull markets longs pay funding continuously, which is a real cost a price-only backtest does not capture. It is frequently large enough to matter.
Counterparty and venue risk are not in the price series. FTX’s failure in November 2022 made this concrete. No price-based backtest models the risk that the venue holding your collateral stops existing.
Try it
The Crypto Dual Momentum template requires both a 100-day moving average condition and a positive 90-day trend, holds whatever passes under inverse-volatility weighting, and goes to cash when nothing does. Weekly rebalance, fractional units, elevated slippage.
The experiment that demonstrates the point: remove the two absolute filters so the strategy always holds the stronger of the two assets, and re-run. Compare the maximum drawdowns across 2022. That difference is what “cash is a position” means in numbers rather than in principle.
Further reading
- Volatility targeting: sizing is a more reliable lever than timing goes further into the sizing half of this strategy
- Commodity trend following on MCX applies the same absolute-plus-relative structure where the trend has a physical mechanism behind it
- Country momentum: rotating across markets, not stocks for a third universe where the cross-asset version has held up
Glossary: momentum, volatility, inverse volatility weighting, max drawdown.