Adaptive Grid v2.1 Validation Study

Volatility-Adjusted Spacing vs Static 2.59% — Consultant Report

Date: 2026-02-23 Classification: Research — For Professional Review Version: 2.1

Executive Summary

This study validates the v2.1 Adaptive Grid hypothesis by running three spacing methods through the same 3 rolling windows with 13 cryptocurrencies: static (2.59%), full adaptive (Φ×ATR with 1.8% floor), and wider-only adaptive (Φ×ATR with 2.59% floor — never tighter than static). All parameters derived from Year 1 data only (blind forward test).

Key findings — Full Adaptive (1.8% floor):

Per-asset return: -2.5pp mean change (-17.4pp to +14.5pp)
Per-asset Calmar: +0.04 mean change

Key findings — Wider-Only Adaptive (2.59% floor):

Per-asset return: -1.3pp mean change (-11.0pp to +8.2pp)
Per-asset Calmar: -0.03 mean change

Portfolio-level (Calmar CDaR):

Static: 44.6% mean return
Full Adaptive: 39.7% mean return (-4.9pp)
Wider-Only: 43.6% mean return (-1.0pp)

1. Methodology

1.1 Adaptive Spacing Formula

The v2.1 framework replaces the static 2.59% grid spacing with:

  Full Adaptive:   Spacing_i = clamp(Φ_grade × ATR_pct_i, 1.8%, 6.5%)
  Wider-Only:      Spacing_i = clamp(Φ_grade × ATR_pct_i, 2.59%, 6.5%)

  Where:
    Φ_grade  = Grid Density Multiplier (varies by rating grade)
    ATR_pct  = Average True Range as % of price (from Year 1 data)
    1.8%     = Full Adaptive floor (fee coverage)
    2.59%    = Wider-Only floor (never tighter than proven static)
    6.5%     = Ceiling (too wide = missed micro-fluctuations)

The Wider-Only variant preserves the static 2.59% as a baseline, only widening for volatile assets. This tests whether the adaptive benefit comes from widening (protecting high-ATR assets) or tightening (capturing micro-moves on low-ATR assets).

1.2 Rating-Based Φ Multiplier

Grade	Φ	Rationale
A	0.85	Tighter grid — proven grid performer, capture more trades
B	0.95	Slightly tighter — strong candidate
C	1.05	Slightly wider — average, need more room
D	1.20	Wider grid — higher volatility, more safety margin

1.3 Blind Forward Test Design

ATR calculated from Year 1 data only (last 90 days of Year 1)
Spacing fixed at start of Year 2 — no mid-simulation recalculation
Same grid range, entry price, and capital ($10,000) for both methods
Only the spacing parameter differs between v1.0 and v2.1

2. Technical Appendix: Per-Asset ATR & Spacing

W1

Asset	ATR%	Grade	Φ	Static	Adaptive	Wider-Only	Adapt Δ	Wider Δ
BTC	1.29%	D-	1.20	2.59%	1.80%	2.59%	-30.5%	0.0%
ETH	1.59%	C	1.05	2.59%	1.80%	2.59%	-30.5%	0.0%
SOL	2.50%	C+	1.05	2.59%	2.63%	2.63%	1.5%	1.5%
BNB	1.68%	C-	1.05	2.59%	1.80%	2.59%	-30.5%	0.0%
ADA	2.10%	D+	1.20	2.59%	2.52%	2.59%	-2.7%	0.0%
XRP	2.01%	C	1.05	2.59%	2.11%	2.59%	-18.7%	0.0%
DOT	1.96%	C-	1.05	2.59%	2.06%	2.59%	-20.6%	0.0%
AVAX	2.47%	D	1.20	2.59%	2.96%	2.96%	14.4%	14.4%
POL	3.13%	C	1.05	2.59%	3.29%	3.29%	27.1%	27.1%
LINK	2.43%	D+	1.20	2.59%	2.92%	2.92%	12.6%	12.6%
UNI	2.73%	D+	1.20	2.59%	3.27%	3.27%	26.3%	26.3%
LTC	1.66%	A-	0.85	2.59%	1.80%	2.59%	-30.5%	0.0%
DOGE	2.39%	D-	1.20	2.59%	2.87%	2.87%	10.7%	10.7%

W2

Asset	ATR%	Grade	Φ	Static	Adaptive	Wider-Only	Adapt Δ	Wider Δ
BTC	1.37%	D+	1.20	2.59%	1.80%	2.59%	-30.5%	0.0%
ETH	1.74%	D	1.20	2.59%	2.09%	2.59%	-19.3%	0.0%
SOL	1.91%	C	1.05	2.59%	2.00%	2.59%	-22.8%	0.0%
BNB	1.43%	D+	1.20	2.59%	1.80%	2.59%	-30.5%	0.0%
ADA	3.02%	C	1.05	2.59%	3.17%	3.17%	22.5%	22.5%
XRP	1.52%	C	1.05	2.59%	1.80%	2.59%	-30.5%	0.0%
DOT	1.60%	D-	1.20	2.59%	1.92%	2.59%	-26.0%	0.0%
AVAX	2.07%	D-	1.20	2.59%	2.48%	2.59%	-4.3%	0.0%
POL	2.78%	D-	1.20	2.59%	3.34%	3.34%	28.8%	28.8%
LINK	1.86%	C	1.05	2.59%	1.95%	2.59%	-24.8%	0.0%
UNI	3.40%	C	1.05	2.59%	3.57%	3.57%	38.0%	38.0%
LTC	1.83%	A-	0.85	2.59%	1.80%	2.59%	-30.5%	0.0%
DOGE	4.14%	D-	1.20	2.59%	4.96%	4.96%	91.7%	91.7%

W3 (Partial)

Asset	ATR%	Grade	Φ	Static	Adaptive	Wider-Only	Adapt Δ	Wider Δ
BTC	1.31%	C	1.05	2.59%	1.80%	2.59%	-30.5%	0.0%
ETH	1.73%	C+	1.05	2.59%	1.82%	2.59%	-29.8%	0.0%
SOL	2.61%	C	1.05	2.59%	2.74%	2.74%	6.0%	6.0%
BNB	2.13%	C+	1.05	2.59%	2.24%	2.59%	-13.5%	0.0%
ADA	3.16%	D-	1.20	2.59%	3.79%	3.79%	46.2%	46.2%
XRP	4.01%	C	1.05	2.59%	4.21%	4.21%	62.4%	62.4%
DOT	3.37%	C	1.05	2.59%	3.54%	3.54%	36.7%	36.7%
AVAX	3.18%	D-	1.20	2.59%	3.82%	3.82%	47.4%	47.4%
POL	3.74%	C	1.05	2.59%	3.93%	3.93%	51.8%	51.8%
LINK	3.64%	C+	1.05	2.59%	3.83%	3.83%	47.7%	47.7%
UNI	3.67%	D-	1.20	2.59%	4.40%	4.40%	69.9%	69.9%
LTC	3.91%	B+	0.95	2.59%	3.71%	3.71%	43.4%	43.4%
DOGE	3.27%	D	1.20	2.59%	3.93%	3.93%	51.6%	51.6%

3. Per-Window Asset-Level Results

W1

Asset	Static	Adaptive	Wider-Only	Adapt Δ	Wider Δ	Static Cal	Adapt Cal	Wider Cal
BTC	4.7%	3.7%	4.7%	-1.0pp	0.0pp	1.30	1.06	1.30
ETH	20.2%	18.6%	20.2%	-1.5pp	0.0pp	0.82	1.01	0.82
SOL	13.5%	13.5%	13.5%	-0.0pp	-0.0pp	3.16	3.50	3.50
BNB	16.7%	9.6%	16.7%	-7.0pp	0.0pp	3.18	4.80	3.18
ADA	70.0%	84.5%	70.0%	+14.5pp	0.0pp	1.65	2.05	1.65
XRP	98.8%	82.0%	98.8%	-16.8pp	0.0pp	5.34	4.21	5.34
DOT	6.1%	10.6%	6.1%	+4.5pp	0.0pp	0.30	0.70	0.30
AVAX	7.7%	2.5%	2.5%	-5.2pp	-5.2pp	0.15	0.05	0.05
POL	-8.1%	-10.4%	-10.4%	-2.3pp	-2.3pp	-0.22	-0.25	-0.25
LINK	57.6%	63.3%	63.3%	+5.7pp	+5.7pp	1.45	1.58	1.58
UNI	37.5%	38.6%	38.6%	+1.0pp	+1.0pp	0.58	0.61	0.61
LTC	184.6%	196.9%	184.6%	+12.3pp	0.0pp	6.48	7.58	6.48
DOGE	4.4%	12.6%	12.6%	+8.2pp	+8.2pp	0.07	0.21	0.21

Win rates vs Static:

Full Adaptive: Return 6/13 | Calmar 9/13
Wider-Only: Return 3/13 | Calmar 4/13

W2

Asset	Static	Adaptive	Wider-Only	Adapt Δ	Wider Δ	Static Cal	Adapt Cal	Wider Cal
BTC	9.7%	8.8%	9.7%	-0.9pp	0.0pp	2.38	2.37	2.38
ETH	30.6%	25.7%	30.6%	-4.9pp	0.0pp	0.80	0.79	0.80
SOL	14.1%	10.9%	14.1%	-3.1pp	0.0pp	0.34	0.28	0.34
BNB	12.8%	8.6%	12.8%	-4.2pp	0.0pp	2.70	2.25	2.70
ADA	-3.3%	-3.6%	-3.6%	-0.4pp	-0.4pp	-0.16	-0.17	-0.17
XRP	62.4%	59.0%	62.4%	-3.4pp	0.0pp	3.77	3.83	3.77
DOT	-50.6%	-50.3%	-50.6%	+0.3pp	0.0pp	-0.71	-0.72	-0.71
AVAX	-22.6%	-24.0%	-22.6%	-1.4pp	0.0pp	-0.48	-0.51	-0.48
POL	-23.0%	-25.7%	-25.7%	-2.7pp	-2.7pp	-0.61	-0.64	-0.64
LINK	26.4%	16.8%	26.4%	-9.5pp	0.0pp	0.66	0.49	0.66
UNI	-7.0%	-11.6%	-11.6%	-4.6pp	-4.6pp	-0.25	-0.35	-0.35
LTC	99.3%	81.9%	99.3%	-17.4pp	0.0pp	4.69	5.93	4.69
DOGE	-51.2%	-59.2%	-59.2%	-8.1pp	-8.1pp	-0.88	-0.94	-0.94

Win rates vs Static:

Full Adaptive: Return 1/13 | Calmar 2/13
Wider-Only: Return 0/13 | Calmar 0/13

W3 (Partial)

Asset	Static	Adaptive	Wider-Only	Adapt Δ	Wider Δ	Static Cal	Adapt Cal	Wider Cal
BTC	8.6%	7.4%	8.6%	-1.3pp	0.0pp	1.60	1.87	1.60
ETH	25.9%	21.9%	25.9%	-4.0pp	0.0pp	0.86	0.89	0.86
SOL	-15.9%	-15.8%	-15.8%	+0.0pp	+0.0pp	-0.41	-0.40	-0.40
BNB	7.6%	5.6%	7.6%	-2.0pp	0.0pp	2.04	1.64	2.04
ADA	-37.9%	-45.8%	-45.8%	-7.9pp	-7.9pp	-0.75	-0.79	-0.79
XRP	-5.2%	-10.0%	-10.0%	-4.8pp	-4.8pp	-0.27	-0.36	-0.36
DOT	-10.6%	-10.4%	-10.4%	+0.1pp	+0.1pp	-0.55	-0.59	-0.59
AVAX	-44.5%	-53.8%	-53.8%	-9.3pp	-9.3pp	-0.72	-0.78	-0.78
POL	-2.1%	-5.2%	-5.2%	-3.1pp	-3.1pp	-0.24	-0.35	-0.35
LINK	-8.8%	-19.8%	-19.8%	-11.0pp	-11.0pp	-0.27	-0.45	-0.45
UNI	-42.5%	-42.3%	-42.3%	+0.2pp	+0.2pp	-0.67	-0.69	-0.69
LTC	20.4%	14.1%	14.1%	-6.3pp	-6.3pp	1.67	0.71	0.71
DOGE	-22.9%	-24.5%	-24.5%	-1.6pp	-1.6pp	-0.45	-0.47	-0.47

Win rates vs Static:

Full Adaptive: Return 3/13 | Calmar 3/13
Wider-Only: Return 3/13 | Calmar 1/13

4. Portfolio-Level Comparison

Calmar CDaR Allocation

Window	Static	Adaptive	Wider-Only	Adapt Δ	Wider Δ
W1	77.2%	70.9%	75.1%	-6.2pp	-2.1pp
W2	45.8%	39.8%	46.1%	-6.0pp	+0.2pp
W3	10.8%	8.3%	9.7%	-2.5pp	-1.1pp

MVO Sharpe Allocation

Window	Static	Adaptive	Wider-Only	Adapt Δ	Wider Δ
W1	71.6%	95.7%	72.9%	+24.1pp	+1.3pp
W2	38.5%	46.8%	38.5%	+8.4pp	-0.0pp
W3	13.9%	13.4%	11.6%	-0.5pp	-2.3pp

5. Aggregate Comparison: v1.0 vs v2.1

Per-Asset Metrics (All Windows Combined)

Metric	v1.0 Static Mean	v2.1 Adaptive Mean	Δ	Improvement
Metric	Static	Full Adaptive	Wider-Only	Adapt Δ
:-------	:------:	:------------:	:----------:	:-------:
Mean Return	12.4%	9.9%	11.1%	-2.5pp
Mean Max DD	31.4%	31.9%	33.1%	-0.4pp
Mean Calmar	0.98	1.02	0.95	+0.04
Mean Sharpe	1.18	1.15	1.19	-0.04

Return Variance Compression

Metric	Static	Full Adaptive	Wider-Only
Range (Max-Min)	235.7%	256.1%	243.8%
Std Dev	45.3%	46.8%	46.8%
Mean-to-Median Gap	4.8%	2.5%	3.4%

6. Volatility Regime Analysis

Which assets benefit most from adaptive spacing?

Regime	Assets	Mean Δ Return	Mean Δ Calmar	Insight
High ATR (≥2.4%)	20 obs	-2.6pp	-0.06	Wider spacing prevents over-buying in crashes
Low ATR (<2.4%)	19 obs	-2.5pp	+0.15	Tighter spacing captures micro-fluctuations

7. Limitations

Static Φ values. The rating-based Φ multipliers (0.85–1.20) are heuristic, not optimized. Future work should test sensitivity to Φ choices.
ATR calculated once. ATR is fixed from Year 1 — no mid-Year-2 adaptation. A continuously adaptive version (Phase 2) may perform differently.
Same grid range. Both methods use the same Year 1-derived upper/lower limits. Adaptive spacing changes level density, not range.
Small sample. 3 windows × 13 assets = 39 observations. Directional but not statistically definitive.
Floor/ceiling clamps. Some assets hit the 1.8% floor or 6.5% ceiling, limiting the adaptive benefit.

8. Conclusions

Full Adaptive (1.8% floor) reduces per-asset returns by -2.5pp on average. Win rate: 10/39 (26%).
Wider-Only (2.59% floor) reduces per-asset returns by -1.3pp on average. Win rate: 6/39 (15%).
Calmar win rate (Full): 14/39 (36%) — the primary target metric.
Regime sensitivity: Low-ATR assets benefited more (-2.5pp vs -2.6pp), suggesting tighter spacing captures more in calm markets.
Critical finding: Tightening spacing below 2.59% for low-ATR assets (BTC, ETH, BNB) generally hurts performance. The static 2.59% appears near-optimal for our Dynamic Grid trailing-up mechanism.
Recommendation: Neither adaptive variant consistently outperforms static spacing in this test. The Dynamic Grid trailing-up mechanism may already provide sufficient volatility adaptation.

Generated by Adaptive Grid v2.1 Study 13 assets, 3 windows, 2026-02-23