FILTRIX Logo FILTRIX Find Your Edge →

FILTRIX Tutorial

How to Validate a Trade-Ideas Alert

Turn any Trade-Ideas alert into measurable filters and check it against 400K+ historical alerts before you trade. Read the alert as data, translate it to FILTRIX filters, and decide whether the setup deserves real risk.

Intermediate 8 min read Updated 2026-04-29

What You Will Learn

  • Read a Trade-Ideas alert and pick the measurable signals out of the headline.
  • Translate alert text into FILTRIX Context filters that mirror the original conditions.
  • Judge whether the result has enough sample size to act on.
  • Decide between trade, skip, and Execution Analysis as the next step.

Before You Start

  • A FILTRIX account. Free accounts can run the full Context Analysis workflow with basic limits.
  • A specific Trade-Ideas alert you want to test. Recent or hypothetical both work.
  • A clear definition of what 'good' means for the trade — green close, target hit, bounce out of a range, or a specific intraday window.
01

Read the alert as data, not as a story

A Trade-Ideas alert like 'GapUp30 — XYZ — Float 12M — RVol 5x — Premarket +35%' packs facts inside a headline. Each fact is a filter. The alert title ('GapUp30') tells you the alert type. The float, RVol, and premarket gap are measurable conditions you can replay against history. Strip the narrative — keep the numbers.

Action: Copy the alert text and split it into separate facts: alert type, ticker, float, premarket volume, gap size, time of day. Write them down before opening Context Analysis.
02

Translate facts into Context filters

Open Context Analysis. For each fact you wrote down, add a filter that matches it as closely as possible. A 35% premarket gap becomes a gap-size filter set to ≥30%. A 12M float becomes a float-range filter under 20M. RVol 5x becomes a relative-volume filter ≥3x. The alert type itself maps to a Trade-Ideas alert filter. Aim for four to six filters total — too few and you backtest the wrong setup, too many and you have no historical matches.

Action: Add one filter per fact from step 1. Use ranges, not exact values, so similar historical alerts can match.
03

Pick the outcome you actually care about

FILTRIX's default outcome is whether the day closed green, but that may not be your real question. If you trade morning gap-and-go, you care about the move from open to a specific time of day, not the close. If you fade gaps, you care about retracement to VWAP. If you swing-trade earnings reactions, you care about overnight moves. Pick the outcome that matches your strategy before you click run — otherwise the win rate number tells you nothing useful.

Action: Set the outcome and threshold in Context Analysis. For directional intraday trades, that is typically close vs. open with a percent threshold. For fade strategies, set a retracement or time-of-day band.
04

Run the analysis and check sample size first

When the result loads, your first eye should go to the matching example count, not to the win rate. If FILTRIX found 5 historical matches, every percent on the win rate is noise — even a 100% win rate on 5 trades is meaningless. Useful results start around 30 matches. Confident results need 100 or more. Below 30, your filters are too narrow or the setup is too rare to backtest reliably.

Action: Look at the matching count first. If it is below 30, broaden the most restrictive filter (usually float or RVol range) and re-run. If it is above 500, tighten the filter that most defines the setup so the sample stays specific.
05

Read the close-position split honestly

With a usable sample, FILTRIX shows how the matching days closed: above the target band, inside it, or below. This is the actual edge signal. A 60 / 20 / 20 split (60% closed above target) is a real signal. A 35 / 30 / 35 split is coin-flip behavior with extra steps, even with 1,000 matching alerts. The Trade-Ideas alert that brought you here does not change those probabilities — the alert is just a way to find the setup, not proof that it works.

Action: Compare the above-band percentage to your minimum win-rate threshold (typical: 55% for tight intraday setups, 60%+ for swing setups). If the result is below your threshold, the setup has no edge regardless of how clean the alert looked.
06

Decide: trade, skip, or test execution

A green Context result is necessary but not sufficient. It says 'this setup has historical edge', not 'this entry and exit will be profitable'. Move to Execution Analysis when you want to test how a specific entry rule (break of premarket high, retest of VWAP, etc.), stop loss, and exit time would have behaved on the matching historical alerts. If the Context win rate is borderline, skip — Trade-Ideas surfaces hundreds of alerts a day, and better setups exist.

Action: Save the context if you will trade it. If the result is borderline, run Execution Analysis to test exact entry and exit rules. If the result is below your threshold, close it and find a better alert.

Common Mistakes to Avoid

  • Don't backtest the alert headline. Backtest the underlying conditions — the alert title is just a label.
  • Don't trust a win rate without a sample-size check. Below 30 matches, every number is noise.
  • Don't change the outcome to chase a better-looking percentage. Pick the outcome before the run.
  • Don't stop at Context Analysis for tight intraday setups. Entry timing and slippage need Execution Analysis to be honest about whether the edge survives real trading.

Next Tutorials

Ready to test the setup?

Open FILTRIX, build the context, and let the data decide before you chase the next alert.

Open FILTRIX