Backtesting
Saved portfolios
Upload
Library
RESET
← BACK TO BUILDER
TEST — STAGE 2a
TEST — STAGE 2b ↓
AI
General
Details
Builder
Planner
Forward
← HOME
STRATIQ
Strategy Analysis Platform · IMPULSE
DRAG YOUR .XLSX FILES HERE
or use the button below · parsing runs in your browser, nothing is uploaded
+ ADD XLSX FILES
▶ ANALYZE
QUEUED STRATEGIES
clear queue
TradingView → Strategy Tester → Performance Summary → ↓ → Save as .xlsx English TradingView exports only — other export languages will not parse
build 2026-10-15 14:50 · 87c3f7
Stage 1 — Builder: compose the book
Basket analysis — recomputed on every change
Candidates — admitted, not in the book
NOT ADMITTED ▾
Forward — what happened after the save
UPDATE PORTFOLIO
SEND UPDATE
CLOSE
Upload · codes and reports
loading…
Library
loading…
Frozen portfolios — the only test that cannot be gamed
A freeze records composition, weights and a fingerprint of every strategy's trades.
Upload fresh xlsx exports of the same strategies at any time — everything after the
freeze date is forward data nobody selected. Verdict in 3–6 months, not one.
loading…
Clear
Add files
▶ Run
Holdout 12m: ON
Save portfolio
0 strategiesbuild 2026-10-15 14:50 · 87c3f7
Equity curve
FUNDING ON
FUNDING OFF
Strategies
What the book is exposed to
Activity heatmap — when each strategy earns
green = profitable month · red = losing month · hover for value
Monthly PnL — all strategies
look for months where strategies profit at different times
Quarterly result — click a bar for the breakdown
Analysis — what your library is telling you
Generated from your data by fixed rules. Every line states the number behind it.
Holdout — the last 12 months were never used
Score calibration — does the number predict anything?
Walk-forward — does a good quarter predict the next one?
Is the selection procedure itself overfitting?
Planner — turning the book into position sizes
EQUAL
25% KELLY
50% KELLY
100% KELLY
TOTAL USDC — SUBACCOUNTS
FLOOR FOR ZEROED
none — trust Kelly
2% each
5% each
SL
5%
10%
TARGET EXPOSURE
auto — buffer rule
1× of total
2× of total
3× of total
5× — extreme
10× — extreme
BUFFER
2 bad stretches
1 bad stretch
margin only
SUBACCOUNT LEV
Allocation per strategy
Setting heatmap — return per unit of drawdown
columns: weight tilt from equal to pure
Kelly proportions · rows: total exposure · cell: median CAGR ÷ p95 drawdown across 120 resampled
paths — resampling breaks the lucky day-ordering that selection baked into the single historical
curve. ✝ = the book goes below −90% equity in that share of paths; there is no number there
because there is no coming back from it. Click a cell to apply its setting. History after
selection, not a promise.
Position sizing side by side
What a bad run looks like
Distribution of outcomes — one path is not evidence
Stress test — is the result an edge or a few lucky events?
Remove the best trades and recount
Behaviour on the worst 10% of portfolio days
Data integrity — what could quietly be wrong
Duplicate trade sets
Backtest settings audit
Shared mechanisms across the book
Correlation of daily returns
Pair behaviour
Best pair synergies — % of months where at least one earns
Pairs losing simultaneously — portfolio risk
TradingView