Take-Profit 6% to 8%: The Full Decision Process

The Question

Should we change BB Squeeze SHORT’s take-profit from 6% to 8%?

This seems simple. It’s not. A wrong decision here means leaving money on the table (if TP is too tight) or never hitting targets (if TP is too wide). The strategy was already verified across 549+ coins. We needed to get this right.

The Data

We ran the full backtest: 2,898 trades across 535 coins over 2+ years (Dec 2023 — Feb 2026).

Metric	TP 6%	TP 8%	Change
Total PnL	$617	$794	+28.6%
Profit Factor	2.01	2.22	+10.4%
Win Rate	70.4%	68.6%	-1.8%p
Break-even WR	62.5%	55.6%	-6.9%p (more margin)
TP Hit Rate	51.8%	42.0%	-9.8%p
SL Hit Rate	7.5%	7.5%	0%
TIMEOUT Rate	40.7%	50.5%	+9.8%p
TIMEOUT Avg PnL	-1.20%	-0.39%	67% better

Key insight: TP 8% reduces TP hits but the ones that hit are worth more. And the timeouts become less painful because positions have more room to recover.

The Expert Panel

We consulted 10 specialized agents. Each used a different methodology:

Expert	Verdict	Reasoning
Backtester	GO	+28.6% PnL, 535 coins, 2yr+
Strategy Analyst	GO	R:R improvement, 78/100 score
Risk Manager	GO	Break-even margin 2.4x wider
Execution Compliance	GO	Code reads config dynamically
Slippage Auditor	GO	MFE shows 42% reach TP8
Validation Analyst	WAIT	p=0.36 (not statistically significant)
Market Regime Monitor	Limited	Only 238 BTC-correlated trades
Portfolio Correlation	INVALID	Methodology error in analysis
Anomaly Detector	GO	No anomalies in TP8 data
Data Quality Engineer	GO	Data integrity confirmed

Final tally: 7 GO / 1 WAIT / 2 INVALID

The Dissent

The Validation Analyst raised a valid concern: the improvement’s p-value was 0.36, meaning there’s a 36% chance the difference is random noise.

However, Monte Carlo simulation (10,000 runs) showed TP 8% outperformed TP 6% in 78% of simulations. The improvement is consistent even if not “statistically significant” by strict academic standards.

We accepted this risk because:

The downside is small (win rate drops 1.8%p)
The upside is large (+28.6% PnL)
Break-even margin widens from 5.1%p to 12.2%p (more safety)

The Most Important Lesson

Before this analysis, we also tested BTC Regime Filters — the idea that you should only trade when Bitcoin is trending a certain way.

6 out of 6 expert analysts recommended this filter as their #1 priority.

We backtested 4 different regime filter designs across the same 535 coins and 2+ years. Every single one made performance worse.

Expert intuition was unanimous and wrong. The data was clear.

This is why PRUVIQ exists. Don’t believe the experts. Don’t believe the signals. Don’t believe the backtest screenshots.

Pruv it.

What Changed

v1.7.0 deployed February 15, 2026
Take-Profit: 6% → 8%
GUNUSDT blacklisted (single coin responsible for -$378 loss due to implementation bug)
All other parameters unchanged (SL 10%, 535 coins, $60 positions, 5x leverage)