Stingray Academy

Stingray Academy is a free, practical course for moving from a market idea to an inspectable strategy workflow. Each lesson produces an artifact you can review: a typed rule, a backtest card, a trigger log, or an alert policy.

The exercises use markets where venue, funding, order-book, prediction-market, and onchain data are unusually observable. The method is simple: test market theses with evidence before deciding which ideas deserve more resources.

Course path

Lesson	Question	Artifact
1. Write a falsifiable thesis	What would have to be true for the idea to matter?	A one-sentence thesis with market, signal, and horizon
2. Turn it into a typed rule	What exactly counts as a trigger?	An inspectable condition with sources and thresholds
3. Run a reproducible replay	How did that rule behave in a stated window?	A backtest card with coverage and fire counts
4. Read the evidence	Is the result broad, clustered, or fragile?	A review of forward returns, sample size, and regimes
5. Monitor out of sample	Does the live signal behave like the replay?	An alert with cooldown and delivery policy
6. Decide what comes next	Does the live evidence still support the thesis?	A documented continue, refine, or stop decision

Lesson 1: write a thesis that can fail

“BTC looks strong” is an observation, not a testable strategy. A useful thesis identifies a market, measurable conditions, and an evaluation horizon.

For example:

Negative BTC funding combined with positive 60-minute price momentum may identify crowded shorts. Test what happened 1, 4, and 24 hours after the condition appeared.

The point is not to make the idea sound quantitative. It is to make disagreement possible. If a threshold, source, or horizon is missing, add it before reading a result.

Lesson 2: inspect the rule

A natural-language interface should not hide the rule it creates. Confirm:

• The venue and instrument are correct.
• Every threshold has a unit and lookback.
• Event and market timestamps use information available at that moment.
• The evaluation interval is fine enough for the condition.
• A cooldown prevents one persistent episode from becoming dozens of apparent signals.

Read How Stingray Works for the full thesis-to-rule workflow.

Lesson 3: run the replay

Freeze the rule before the replay begins. Record the source, interval, window, warmup, cooldown, and output horizons. If a source is unavailable or the history is too coarse, reject the replay instead of silently approximating it.

The No-Code Backtesting Tutorial walks through one complete exercise. The Backtesting Methodology documents the replay semantics in detail.

Lesson 4: read evidence without turning it into a promise

Start with coverage and sample size, not the largest return. Compare raw matches with distinct events and cooldown-filtered fires. Look for clustering, regime dependence, and whether a few observations dominate the average.

The SEC reminds investors that backtested performance is hypothetical and does not reflect actual performance. That is why Stingray publishes trigger history and limitations rather than treating one headline statistic as a verdict. See the SEC performance-claims bulletin.

Lesson 5: monitor out of sample

The first live version should normally be an alert. Compare live fires with the frozen backtest:

• Did the source arrive on time?
• Did the rule fire when expected?
• Did market structure or liquidity change?
• Did the alert create a useful decision, or only noise?

Monitoring turns a historical artifact into an operational test without giving the workflow permission to trade.

Lesson 6: decide what comes next

Use the live evidence to make a clear decision:

• Continue: keep the alert when live behavior matches the evidence.
• Refine: change one assumption, then run a new backtest.
• Stop: retire the rule when the evidence no longer supports it.

Stingray does not place trades. The trader decides what to do elsewhere.

Graduation checklist

You have completed the course when you can show:

• A falsifiable thesis and typed rule
• Named sources, interval, window, warmup, and cooldown
• Every historical fire and the declared forward horizons
• Coverage and cost limitations
• A decision to reject, refine, or monitor
• A documented next-step decision

Continue with the Strategy Examples, compare published evidence in the Backtesting Benchmark Report, or browse the Trading Platform Comparison.