We tested the same $10,000 starting portfolio from 1 January 2018 to 19 September 2026. Every model began with 50% BTC and 50% ETH. The comparison changes only the rebalancing rule.
Rebalancing is a risk rule, not a price forecast
When BTC outperforms, its weight rises above 50%. Rebalancing sells part of that overweight position and buys ETH; after ETH outperforms, the process works in reverse.
The method can harvest relative volatility, but it cannot remove market risk. BTC and ETH can fall together, so a perfectly balanced portfolio may still suffer a deep drawdown.
Four models under the same conditions
Buy and hold
Allocate $5,000 to BTC and $5,000 to ETH on the first day, then make no further trades. This is the benchmark for natural weight drift.
Monthly reset
At each month-end, return both assets to 50%. A 0.10% fee is charged on each side of the transferred amount.
Quarterly reset
Rebalance at the end of each calendar quarter. The rule trades less often and allows wider drift between review dates.
10-point threshold
Trade only when BTC moves outside the 40%–60% band. This responds to meaningful drift instead of an arbitrary date.
Three views of the same portfolio
The first chart compares portfolio value. The second shows how BTC weight drifts and where the threshold model resets it. The third compares peak-to-trough declines.
BTC + ETH · $10,000
2018-01-01 — 2026-09-19 · Study period
The full series contains 3,184 consecutive daily observations. The logarithmic scale keeps the early years readable instead of compressing them into an apparently empty strip.
BTC · BTC weight
40% / 50% / 60% · 10-point threshold
Sharp returns toward 50% are actual threshold rebalances, not missing data.
Max drawdown
BTC + ETH · %
The winner was not the whole lesson
Monthly rebalancing finished highest in this sample, while the threshold model reached a similar result with far fewer interventions. That ranking belongs to this exact period, asset pair and fee assumption; it is not a universal law.
All four variants suffered an approximately 87% maximum drawdown because both assets collapsed during the same crypto bear markets. Rebalancing changed the path and final value, but it did not create diversification between truly independent risks.
| Model | Final value | Return | Max drawdown | Rebalances | Modelled fees |
|---|---|---|---|---|---|
| Buy and hold | $47,798 | +378.0% | -87.9% | 0 | $0 |
| Monthly | $65,467 | +554.7% | -87.4% | 104 | $193 |
| Quarterly | $60,735 | +507.3% | -87.8% | 34 | $96 |
| 10-point threshold | $63,581 | +535.8% | -87.6% | 14 | $85 |
What can break the result
A different starting date
A test beginning near another market cycle can reverse the ranking. Always repeat the calculation across several start dates.
Real execution and taxes
Trading fees, spread, slippage and taxable disposals can make frequent rebalancing materially less attractive.
Hidden concentration
Two cryptoassets are still one concentrated risk bucket. A 50/50 label does not mean half the portfolio risk comes from each asset.
A beginner-friendly implementation plan
- Write down the target weights before buying.
- Choose one trigger: calendar or threshold, never both opportunistically.
- Estimate the full cost and tax effect before every reset.
- Use limit or staged execution when the required transfer is large.
- Record the pre-trade weights, amount transferred and post-trade weights.
The useful rule is the one you can follow
A threshold rule was the most efficient compromise in this study: it captured much of the rebalancing effect with only 14 resets. The correct choice still depends on costs, taxes and whether BTC/ETH is only one sleeve of a wider portfolio.
Data and methodology
Binance Spot daily OHLCV · BTCUSDT & ETHUSDT · UTC · 2018-01-01 — 2026-09-19
Transparent rules · fees included where specified