Charting, signal backtesting and strategy validation.
A research tool for studying price behaviour in equities — built to test whether a technical signal actually held up, rather than whether it looks convincing on a chart. Every result is stored, so a question asked once never has to be re-run.
Daily+1D1W1M MA 20MA 50RSI 14
Daily, weekly and monthly series · MA, RSI, MACD, ADX, Ichimoku and Bollinger overlays
Every combination, ranked by edge — not by eye
The Backtest tab runs every single, pair and triple combination of the available signals against a ticker’s full history and ranks them by a Sortino-style edge score — average return × √trades ÷ downside deviation, so only losing trades are penalised.
- ▸Single, pair and triple signal combinations — correlated pairs are pruned before they inflate the ranking.
- ▸Sortino score, profit factor, Kelly %, hit ratio, payoff, exposure and drawdown per strategy.
- ▸Click any row to replay that strategy’s trades on the chart above.
- ▸Results are persisted, so revisiting a ticker loads instantly instead of re-running.
| Strategy | Sortino ★ | PF | Avg ret | Trades | Hit |
|---|---|---|---|---|---|
| MA20>MA50 + RSI<40 | 3.4 | 2.81 | +4.12% | 41 | 63% |
| Golden cross + ADX>25 | 2.9 | 2.44 | +3.96% | 28 | 61% |
| RSI oversold reversal | 2.2 | 1.97 | +2.71% | 66 | 56% |
| MACD cross + MA200 up | 1.8 | 1.62 | +2.04% | 37 | 54% |
| Ichimoku cloud break | 1.3 | 1.34 | +1.42% | 22 | 50% |
| Bollinger squeeze | 0.7 | 0.94 | -0.38% | 54 | 44% |
| Stochastic %K cross | 0.4 | 0.81 | -0.91% | 73 | 41% |
| Strategy | OOS OK | Consistency | Med. Sortino | Med. avg ret |
|---|---|---|---|---|
| MA20>MA50 + RSI<40 | 19/24 | 79% | 2.61 | +3.44% |
| Golden cross + ADX>25 | 17/24 | 71% | 2.28 | +2.91% |
| RSI oversold reversal | 14/24 | 58% | 1.94 | +2.12% |
| MACD cross + MA200 up | 11/24 | 46% | 1.61 | +1.38% |
| Ichimoku cloud break | 8/24 | 33% | 1.22 | +0.74% |
| Bollinger squeeze | 4/24 | 17% | 0.88 | -0.22% |
A strategy has to survive data it was never fitted on
The Universe tab repeats the whole search across a list of names, splitting each history 70/30 into in-sample and out-of-sample. Strategies are ranked by consistency — the share of stocks where the edge still held on the unseen 30% — and every per-company result stays inspectable.
- ▸70/30 in-sample / out-of-sample split per name, ranked by consistency across the universe.
- ▸Per-company breakdown: IS Sortino, IS hit %, OOS avg return, OOS hit % and exclusions.
- ▸Progress streams over SSE — partial results are usable while the scan is still running.
- ▸Filter by volatility bucket to see whether an edge is regime-specific.
Build the benchmark you actually care about
Compare overlays several names indexed to 100 in a chosen currency. Custom lines go further: a cap- or equal-weighted index built from any watch list, or a ratio of A ÷ B — a stock against its own peer group — with the weights editable name by name.
- ▸Any number of tickers indexed to 100 from the visible start, in EUR, USD or local currency.
- ▸Cap- or equal-weighted custom indices built from a watch list, weights editable name by name.
- ▸Ratio lines (A ÷ B) so relative strength is a first-class series, not a mental calculation.
- ▸Names without market-cap data are flagged and dropped rather than silently mis-weighted.
Smart Fit
Draw the zones where you wanted to be long or flat, and score every signal on how well it matched your intent alongside its own edge.
Pattern reliability
Detects double tops and bottoms, head & shoulders, triangles, wedges and support/resistance, each with a confidence score, then measures how often they resolved as the textbook claims.
Named signal rules
Save a combination as a named rule and evaluate it against any ticker in real time, with a quick backtest behind it.
Regime tagging
Each stock is characterised by momentum and volatility regime, so you can ask whether an edge decays as the peer group grows less similar.
Commodities & FX context
Related commodity and currency series alongside the equity, for names whose price is driven by an input cost.
Deflated results
Every search records how many strategies were tried, so a winner’s apparent alpha can be discounted for the attempts that produced it.
This is a private research tool, not a commercial service. It is used by invitation and carries no advertising, no tracking and no public data.
If you have been given access, sign in at app.technicalanalysis.nl.