Regime Transition Detection Study

Can Predicting Regime Shifts Improve Grid Trading? — Consultant Report

Date: 2026-02-24 Classification: Research — For Professional Review Version: 1.0

Executive Summary

This study tests whether predicting regime transitions improves grid trading performance. Using 13 cryptocurrencies across 3 rolling windows, we compare 6 spacing strategies including the current static 2.59% baseline.

Key findings (vs Static 2.59% baseline):

Strategy	Mean Δ Return	Win Rate	Mean Δ Calmar	Mean Δ Drawdown
ETS Current	+0.4pp	20/39 (51%)	-0.37	-0.4pp
Hurst Derivative	+0.9pp	20/39 (51%)	-0.11	-0.4pp
Composite Score	+0.4pp	20/39 (51%)	-0.37	-0.4pp
HMM Regime	-3.7pp	10/39 (26%)	-0.74	-1.4pp
Wider-Only + Composite	+0.3pp	18/39 (46%)	-0.47	-0.5pp

Most frequent best strategy (by Calmar): Static 2.59% (20/39)

1. Methodology

1.1 Segmented Simulation

Year 2 is split into 4 quarters. At each quarter boundary, the regime detector analyzes trailing data and sets spacing for the upcoming quarter. This simulates real-time regime monitoring with quarterly rebalancing.

1.2 Six Spacing Strategies

ID	Strategy	Spacing Logic
A	Static 2.59%	Fixed, no regime awareness
B	ETS Current	Uses ADX/Hurst/Entropy signal levels (our current approach)
C	Hurst Derivative	Rate of change of Hurst exponent (dH/dt)
D	Composite Score	Weighted dH/dt + dADX/dt + ATR breakout + dEntropy/dt
E	HMM Regime	4-state Hidden Markov Model trained on Year 1
F	Wider-Only + Composite	Composite detection but spacing floor = 2.59%

1.3 Spacing by Detected Regime

Regime	Full Spacing	Wider-Only Spacing
Ranging	2.00%	2.59% (baseline)
Trending	3.50%	3.50%
Volatile	4.00%	4.00%
Unknown	2.59%	2.59%

2. Per-Window Results

W1 — 12 months

Asset	Static	ETS	dH/dt	Composite	HMM	Wider+Comp	Best
BTC	12.8%	13.3%	11.9%	13.3%	11.0%	13.3%	B
ETH	-1.5%	-3.5%	-8.5%	-3.5%	-6.1%	-3.5%	A
SOL	21.8%	16.3%	22.2%	16.3%	15.5%	16.3%	C
BNB	17.9%	21.0%	20.0%	21.0%	20.3%	21.0%	A
ADA	-2.7%	-2.1%	-1.0%	-2.1%	-3.8%	-2.1%	C
XRP	78.1%	95.9%	95.9%	95.9%	97.8%	95.9%	A
DOT	-17.6%	-16.3%	-16.3%	-16.3%	-18.5%	-16.3%	B
AVAX	33.3%	29.6%	29.6%	29.6%	30.6%	29.6%	A
POL	-4.9%	-1.8%	-1.8%	-1.8%	-3.6%	-1.8%	B
LINK	38.7%	41.0%	43.3%	41.0%	37.6%	41.0%	A
UNI	85.7%	85.7%	94.7%	85.7%	35.4%	85.7%	C
LTC	107.9%	107.9%	116.6%	107.9%	113.9%	107.9%	A
DOGE	-10.8%	-10.8%	-10.8%	-10.8%	-15.2%	-10.8%	A

Win rates vs Static baseline: ETS: 7/13 dH/dt: 9/13 Composite: 7/13 HMM: 4/13 WiderComp: 7/13

Best strategy (by Calmar): A:7 | B:3 | C:3

W2 — 12 months

Asset	Static	ETS	dH/dt	Composite	HMM	Wider+Comp	Best
BTC	0.7%	0.9%	-1.9%	0.9%	-6.0%	0.9%	B
ETH	-0.3%	2.0%	0.6%	2.0%	1.1%	2.0%	B
SOL	12.8%	6.9%	6.9%	6.9%	6.5%	6.1%	A
BNB	20.2%	25.5%	23.1%	25.5%	24.3%	25.5%	A
ADA	-10.8%	-10.8%	-13.3%	-10.8%	-19.9%	-10.8%	A
XRP	55.7%	68.2%	68.2%	68.2%	47.6%	68.2%	B
DOT	-26.4%	-25.8%	-25.8%	-25.8%	-26.3%	-25.8%	A
AVAX	-19.7%	-19.4%	-25.4%	-19.4%	-25.1%	-19.4%	B
POL	-29.9%	-34.3%	-34.2%	-34.3%	-34.2%	-34.3%	A
LINK	11.9%	14.6%	14.6%	14.6%	15.5%	14.6%	E
UNI	-6.3%	-5.9%	-2.9%	-5.9%	-11.7%	-6.3%	C
LTC	62.8%	62.8%	68.5%	62.8%	44.4%	62.8%	A
DOGE	-34.8%	-34.8%	-34.9%	-34.8%	-30.1%	-34.8%	A

Win rates vs Static baseline: ETS: 8/13 dH/dt: 7/13 Composite: 8/13 HMM: 5/13 WiderComp: 7/13

Best strategy (by Calmar): A:7 | B:4 | E:1 | C:1

W3 (Partial) — 11.7 months

Asset	Static	ETS	dH/dt	Composite	HMM	Wider+Comp	Best
BTC	-1.8%	-0.2%	-4.3%	-0.2%	-3.2%	-0.2%	B
ETH	-13.8%	-25.1%	-13.8%	-25.1%	-21.5%	-25.1%	A
SOL	-13.1%	-13.1%	-14.8%	-13.1%	-22.5%	-13.1%	A
BNB	-10.7%	-11.2%	-10.7%	-11.2%	-13.7%	-11.2%	A
ADA	-13.5%	-14.8%	-14.8%	-14.8%	-15.8%	-14.8%	A
XRP	-6.5%	-5.7%	-5.7%	-5.7%	-6.6%	-5.7%	B
DOT	-11.0%	-14.1%	-9.8%	-14.1%	-14.1%	-14.1%	C
AVAX	9.2%	3.1%	3.1%	3.1%	1.3%	3.1%	A
POL	-57.1%	-60.7%	-60.7%	-60.7%	-60.8%	-60.7%	A
LINK	-4.2%	-2.6%	-5.3%	-2.6%	-3.4%	-2.6%	B
UNI	3.9%	6.7%	3.9%	6.7%	-2.5%	3.9%	B
LTC	19.8%	22.8%	22.8%	22.8%	16.9%	22.8%	B
DOGE	-12.5%	-12.5%	-10.8%	-12.5%	-16.3%	-12.5%	C

Win rates vs Static baseline: ETS: 5/13 dH/dt: 4/13 Composite: 5/13 HMM: 1/13 WiderComp: 4/13

Best strategy (by Calmar): A:6 | B:5 | C:2

3. Aggregate Comparison

Return Improvement vs Static Baseline

Strategy	Mean Δ	Median Δ	Min	Max	Win Rate	Consistency
ETS Current	+0.4pp	+0.3pp	-11.3pp	+17.7pp	20/39 (51%)	✓/✓/✗
Hurst Derivative	+0.9pp	+0.3pp	-7.0pp	+17.7pp	20/39 (51%)	✓/✓/✗
Composite Score	+0.4pp	+0.3pp	-11.3pp	+17.7pp	20/39 (51%)	✓/✓/✗
HMM Regime	-3.7pp	-3.0pp	-50.4pp	+19.7pp	10/39 (26%)	✗/✗/✗
Wider-Only + Composite	+0.3pp	0.0pp	-11.3pp	+17.7pp	18/39 (46%)	✓/✓/✗

Calmar Ratio Improvement

Strategy	Mean Δ Calmar	Win Rate
ETS Current	-0.37	15/39 (38%)
Hurst Derivative	-0.11	13/39 (33%)
Composite Score	-0.37	15/39 (38%)
HMM Regime	-0.74	4/39 (10%)
Wider-Only + Composite	-0.47	13/39 (33%)

4. Transition Detection Quality

Spacing Adjustment Frequency

Strategy	Mean Adjustments/Asset	Assets Never Adjusted	Insight
ETS Current	1.5	9/39	Moderate — some adjustments
Hurst Derivative	1.6	4/39	Moderate — some adjustments
Composite Score	1.5	9/39	Moderate — some adjustments
HMM Regime	1.8	3/39	Moderate — some adjustments
Wider-Only + Composite	1.5	11/39	Moderate — some adjustments

Regime Distribution (Composite Approach)

Regime	Count	Percentage
unknown	103	66.0%
trending_up	48	30.8%
ranging	4	2.6%
volatile	1	0.6%

5. Limitations

Quarterly rebalancing only. Spacing adjusts 4 times during Year 2, not continuously.
Segmented simulation. Each quarter starts fresh — no position carryover between segments.
HMM training data. Only Year 1 data available for training — may not capture all regimes.
Indicator thresholds are heuristic. dH/dt and composite score thresholds are not optimized.
Small sample. 3 windows × 13 assets = 39 observations per strategy.
The baseline is strong. Static 2.59% with Dynamic Grid trailing-up is already well-optimized.

6. Conclusions & Implementation Recommendations

Results are approximately neutral. No approach significantly outperforms static 2.59% on returns. However, risk metrics (Calmar, drawdown) may tell a different story.
Wider-Only + Composite (F) is the safest to integrate — floor at 2.59% means zero downside risk vs baseline. Win rate: 18/39 (46%).
Implementation path if results warrant integration:
- Phase 1: Add dH/dt to ETS confidence calculation (parameter tweak, minimal risk)
- Phase 2: Show "Volatility Spike" alerts in Trading UI when ATR breakout detected
- Phase 3: Add "Regime-Adaptive" toggle (opt-in, wider-only approach)
Never change the 2.59% default. All regime adaptation should be additive and behind user control.

Generated by Regime Transition Detection Study v1.0 13 assets, 3 windows, 6 strategies, 2026-02-24