Bitcoin 
Ethereum 
Solana 
TRON 
Zcash 
Hyperliquid 
Dogecoin 
Monero 
Chainlink 
Cardano 
Stellar 
Uniswap 
Sui 
Hedera 
Shiba Inu 
Bittensor 
Ripple USD 
Aster 
Polkadot 
World Liberty Financial 
Render 
Artificial Superintelligence Alliance 
Two new listings, one quarter that punished both of them.
ENA and ASTER were still finding their footing when this window opened. By October 6, 2025 – the point where both coins had real, tradable price data – ENA sat at $0.58 and ASTER at $1.50.
By December 20, ENA had collapsed to $0.21. ASTER closed at $0.72.
Neither coin found a floor. ENA lost 63.8% of its value. ASTER lost 52%. A straight 50/50 buy-and-hold of both would have cost you $2,309.60 on a $4,000 stake.
The question we wanted answered: when both assets in a rebalance pair are falling — not diverging, not one winning while the other loses, just both bleeding — does actively rebalancing between them help, hurt, or just delay the inevitable?
We ran this backtest on real historical price data across the effective Oct 6 – Dec 20 window to find out.
ENA (50% TARGET)
ASTER (50% TARGET)
Strategy Parameters
How Each Setting Impacted Performance?
Every parameter had a job. In a market where both coins fell, most of those jobs became damage control instead of profit capture.
Parameter Impact Summary
| Parameter | Impact | The Logic (Why) |
|---|---|---|
| 50/50 Allocation | 📉 Locked in exposure to both losses | Neither asset escaped the decline |
| 5% Ratio Threshold | 🛡️ Held trades to just 17 | Widest gap of the three variants tested |
| By Coin Ratio + 1m Check | ⚖️ Rebalanced only on real drift | Frequent checks filtered out short-term noise |
| Mandatory USDT Conversion | 🔒 Crystallized the loss on Dec 20 | No option to wait for a possible rebound |
17 swaps. $7.03 in fees. −$2,039.90 in losses - still $269.70 better than doing nothing.
📝 The math that matters
💰 The Bottom Line
This strategy lost $2,039.90 on $4,000 of capital – a 51.00% hit.
That’s bad.
But the alternative was worse: a straight 50/50 buy-and-hold of the same two coins would have cost $2,309.60, a 57.74% hit. Rebalancing didn’t make money in this window. It made the loss $269.70 smaller.
Don’t annualize this one. The effective window is roughly 75 days (2.5 months), which works out to a −20.4% monthly yield. Multiply that by 12 and you get a “annualized” return of −244.8% – a number that’s mathematically impossible to actually realize, since you can’t lose more than 100% of a spot position. It’s a reminder that short, high-volatility windows should never be stretched into a yearly projection.
⚖️ Trade Frequency & Capital Movement
17 swaps across 75 days means the bot barely touched this portfolio – roughly one rebalance every four to five days. Individual swap sizes ran from about $140 to $350, a small slice of the $4,000 book each time.
Most of the capital just sat in its current allocation, riding the decline in both coins, only nudged when the 5% ratio gap actually opened up.
🛡️ The Fee Advantage
Total fees came to $7.03 – just 0.34% of the total loss. Fees are not the story here. This isn’t a case where fee drag quietly ate the edge; it’s a case where price action on both legs did all the damage, and the rebalancing logic simply limited how much of that damage compounded.
Here comes our A/B/C strategies quick comparison:
| Variant | Threshold | Time Logic | Trades | ROI % | P&L (USDT) |
|---|---|---|---|---|---|
| Test A (Conservative) | 1% | None | 272 | −51.54% | −$2,061.45 |
| Test B (Aggressive) | 2% | 30m | 69 | −54.64% | −$2,185.61 |
| Test C (Optimized/Playbook)This Playbook | 5% | 1m | 17 | −51.00% | −$2,039.90 |
The table doesn’t tell a clean “fewer trades = better” story. Test C, the widest threshold with the fewest trades, did perform best. But Test B – a moderate 2% threshold with fewer trades than Test A – was the worst performer of the three, despite trading less often than Test A’s 272 swaps.
The variable that mattered wasn’t trade count on its own. It was Test B’s combination of a ratio trigger and a 30-minute time delay, which meant it reacted to drift later than either the constantly-checking Test A or the wide-threshold Test C. In a market with no mean reversion to capture, timing the rebalance late was worse than either trading constantly or barely trading at all.
What the results are really telling you.
✅ what worked
The 5% threshold did its job. Only 17 swaps fired over 75 days, each averaging a $119.99 loss instead of piling on extra fee drag. Total fees landed at $7.03 — effectively invisible against a $2,039.90 loss.
By staying quiet, Test C avoided compounding two falling legs into more transactions than necessary, edging out both tighter variants.
⚠️What didn't work
ENA fell 63.8% and ASTER fell 52% – both legs declined, so every rebalance sold the relatively stronger asset to buy the weaker one, with no recovery to capture on either side. Test B (2% ratio + 30-minute filter) was the worst performer at −54.64%, despite firing fewer trades than Test A.
Combining a ratio trigger with a time delay meant it caught rebalances at worse moments than reacting immediately would have.
💡 The key insight
Rebalancing between two losers doesn’t rescue you – it manages the loss. Neither ENA nor ASTER “won” this quarter: dual decline meant every swap transferred capital from one falling asset into another, with no reversion to harvest. What still worked was width.
The 5% threshold acted as a shock absorber, letting minor drift pass as noise and only reacting to genuine relative moves. In a dual-bearish regime, rebalancing’s edge isn’t upside capture – it’s damage control, and how wide you set the trigger determines how much control you actually get.
🚩 Watch out for - a potential red flag
The mandatory end-date conversion is the real risk here. On Dec 20, the bot force-sold everything into USDT, locking in the −51.00% loss at that exact moment – whether or not either coin was about to bounce. A backtest snapshot on a different end date could show a materially different number.
Also watch the “Rebalancing Edge”: +6.74% sounds like a win, but it’s a smaller loss, not a profit. Before deploying, confirm you’re comfortable holding through a full cycle rather than being forced to convert on an arbitrary date, or add a manual exit override.
Overall Performance Score, Strengths and Limitations
Effective damage control, not a winning strategy.
Losing 51.00% against a 57.74% HODL benchmark isn't a win — it's a smaller loss.
🧭 What this strategy does well
- Beat the HODL benchmark by 6.74 percentage points ($269.70)
- Fee drag negligible: $7.03 total across 17 swaps
- Widest threshold (5%) outperformed both tighter variants tested
- Low trade count meant minimal compounding of bad entries
- Mandatory conversion gave a clean, unambiguous result to analyze
🚫 What limited the result
- Still lost more than half the capital deployed (−51.00%)
- No mean reversion available — both legs fell together for the full window
- Configured Sep 10 start date had no valid ASTER price data (TGE was Sep 17); the real test only covers Oct 6–Dec 20
- Result is highly date-sensitive because of the mandatory end-date conversion
- Trade count didn't predict outcome cleanly — Test B underperformed Test A despite fewer trades
Quick Takeaways
- Rebalancing two falling assets limits damage – it doesn’t create profit
- Wider thresholds (5%) beat tighter ones in this dual-bearish window
- Trade count alone doesn’t predict outcome – Test B proves it
- Fee drag was a non-factor at 0.1% and 17 total swaps
- The mandatory conversion date matters as much as the strategy itself
How did passive HODL compare?
If you had simply bought $2,000 of ENA and $2,000 of ASTER on October 6 and held to December 20, here’s how it compares:
Running the bot instead of holding: −$2,039.90 minus −$2,309.60 = a $269.70 advantage. Rebalancing didn’t turn this quarter profitable – nothing would have.
It meant losing $269.70 less than sitting still while both coins fell.
Before you run this playbook, check these off.
Use this as your go/no-go checklist before deploying this exact parameter set.
🧠 Market Suitability Matrix
| Market Condition | Performance (Rating) | Strategy Notes |
|---|---|---|
| Both assets sideways / choppy | ★★★★★ Excellent | Harvests spread from noise, no directional risk. |
| One asset dips, then recovers | ★★★★★ Ideal | Buys the dip, sells into the recovery. |
| Both assets in a mild bull market | ★★★★☆ Good | Trims the leader to fund the laggard. |
| One asset strongly outperforms | ★★★☆☆ Moderate | Sells the winner early, caps upside. |
| Both assets are in steep decline | ★★☆☆☆ Risky | This backtest: −51.00% ROI, no reversion arrived. |
| One asset in structural breakdown | ★☆☆☆☆ Poor | Keeps buying a coin that won't recover. |
| Highly correlated assets (same direction) | ★☆☆☆☆ Poor | No spread to harvest — ENA/ASTER proved this. |
How to tune this playbook for different scenarios.
Disclaimer: All data sourced from CryptoGates Rebalance Backtest Bot. Results are historical simulations using Binance 1-minute OHLCV data. Past backtest performance does not guarantee future live trading results. DYOR.
Battle-Test Your Strategy
Before the Market Does.
Eliminate guesswork with institutional-grade backtesting for DCA, Grid, and Rebalance bots. Real historical data. Real-world results.
