Three Layers to Better Crypto Returns
How Rating Selection, Portfolio Optimization, and Dynamic Grid Trading Each Independently Improve Performance
The Big Question
Does CoinRoc's three-layer methodology — asset selection, portfolio optimization, and grid trading — actually produce better results than just buying and holding? And how much does each layer independently contribute?
We tested this across 3 rolling time windows spanning 30 months with 13 cryptocurrencies, using blind forward testing. Each layer was isolated and measured independently.
The Three-Layer Design
Layer 1: Pick the Right Cryptos
Our composite rating scores every crypto across multiple weighted factors. Only B- rated and above make the cut. This filters out underperformers before you invest a single dollar.
Layer 2: Allocate Smartly
Efficient Frontier optimization puts more money where the risk/return math is strongest. Three methods available: MVO Sharpe, Calmar CDaR, and HRP Risk Parity.
Layer 3: Trade With a Grid
Dynamic Grid trading places automatic buy and sell orders at regular intervals, profiting from every price oscillation. Reduces drawdowns by ~9pp consistently.
The Most Recent Period — Primary Showcase
What would $10,000 have become in the most recent test period (early 2025 – February 2026)?
Improvement: +59.8 percentage points. From losing $4,343 to gaining $1,640 — a swing of nearly $6,000 on a $10,000 investment.
Consistency Across All Test Periods
Each layer was tested across 3 rolling windows — bull, mixed, and bear markets. Selection and optimization were positive in every single window.
Layer 1: B- Rating Selection
Layer 2: EF Optimization (MVO)
Layer 3: Dynamic Grid Strategy
Grid trading reduces raw return in bull markets (it sells as prices rise) but consistently reduces drawdowns by ~9pp and provides critical protection in bear markets.
Which Optimization Method Works Best?
CoinRoc offers three portfolio optimization methods. Here's how they compared across all test windows:
Which should you choose?
Six-Portfolio Comparison — All Windows
Each window compares six portfolios to isolate each layer's contribution:
Window 1 (Bull Market)
| Portfolio | Return | Max DD | Sharpe | Assets |
|---|---|---|---|---|
| A Naive Buy & Hold | +88.6% | 40.0% | 4.35 | 13 |
| B All Assets / Grid | +39.5% | 30.8% | 2.72 | 13 |
| C B- Rated / EqWt | +61.8% | 22.6% | 3.54 | 5 |
| D B- / MVO Sharpe | +83.8% | 18.0% | 5.39 | 5 |
| E B- / Calmar CDaR | +106.9% | 18.2% | 4.94 | 5 |
| F B- / HRP | +46.5% | 18.5% | 3.51 | 5 |
Window 2 (Mixed Market)
| Portfolio | Return | Max DD | Sharpe | Assets |
|---|---|---|---|---|
| A Naive Buy & Hold | +15.8% | 42.8% | 0.54 | 13 |
| B All Assets / Grid | +7.5% | 32.9% | 0.21 | 13 |
| C B- Rated / EqWt | +32.0% | 27.9% | 1.63 | 6 |
| D B- / MVO Sharpe | +53.1% | 32.8% | 1.82 | 6 |
| E B- / Calmar CDaR | +64.0% | 19.2% | 2.99 | 6 |
| F B- / HRP | +24.0% | 23.7% | 1.58 | 6 |
Window 3 (Bear Market — Most Recent)
| Portfolio | Return | Max DD | Sharpe | Assets |
|---|---|---|---|---|
| A Naive Buy & Hold | -43.4% | 39.8% | -2.71 | 13 |
| B All Assets / Grid | -9.8% | 30.6% | -1.21 | 13 |
| C B- Rated / EqWt | +2.2% | 18.9% | -0.32 | 9 |
| D B- / MVO Sharpe | +16.4% | 12.4% | 1.46 | 9 |
| E B- / Calmar CDaR | +12.3% | 9.5% | 1.24 | 9 |
| F B- / HRP | +1.4% | 11.6% | -0.65 | 9 |
What Each Layer Does Best
| What You Want | Which Layer Delivers |
|---|---|
| Consistent alpha | Selection — always positive, +19.6pp mean |
| Higher risk-adjusted returns | Optimization — always positive, +19.1pp mean |
| Downside protection | Grid Trading — reduces drawdown 9pp, prevents crashes |
| Best overall | All three together — each solves a different problem |
Honest Limitations
~30 months of data
Enough to see a pattern, not enough for statistical certainty. More data will strengthen these findings.
Overlapping test periods
Three related experiments, not three independent ones.
Included a major bull run
A multi-year bear market would produce different absolute numbers.
Simulations ≠ reality
Real trading involves fees, slippage, exchange outages, and emotional decisions.
Past performance ≠ future results
The most important caveat in all of finance.
The Bottom Line
Each layer of the CoinRoc methodology independently improves results. Selection picks the winners. Optimization concentrates your capital smartly. Grid trading protects you when markets turn ugly.
Together, they turned a -43% loss into a +16% gain in the toughest market we tested — a 60 percentage point improvement. Not from one lucky trade, but from a systematic, repeatable process.
Related Research
Rolling Backtest Validation
54.9% mean return across 3 rolling windows with blind forward testing*
Rating System Validation
100% accuracy for B-rated and above across 60 cryptocurrencies*
Grid Profitability by Exchange
480 backtests: Binance.US 85% vs Coinbase 45%
HRP vs MVO Construction
Comparing portfolio optimization methods for crypto
* Note on this figure, added 2026-08-05: The 0.886 correlation (and the return/win-rate figures on this page) were measured under an earlier version of the rating formula (dated 2026-01-29). The formula has since been re-optimized three times; internal walk-forward testing as of 2026-05-13 indicates a higher correlation under the current formula. We are re-running this study under the live formula and will republish updated figures. The qualitative finding — that higher-rated assets have historically outperformed lower-rated ones — is not affected by this note; the exact numbers above should not be cited as describing today's system until the re-run is complete.
See All Three Layers in Action
Start with Discovery to find B- rated cryptos, build an optimized portfolio with the Efficient Frontier, and deploy Dynamic Grid bots — all in one platform.