The only astrology platform that grades itself.

Most cosmic apps want you to trust them on faith. They give you a reading, you nod or roll your eyes, and the app never asks how it landed for you. There's no feedback loop, so there's no learning.

imotionengine inverts that. From day one, every engine's predictions get checked against what actually happens in your life. The system measures itself out loud, in front of you. After thirty days you stop guessing which tradition is the one for you - you see it, drawn from your own data.

The user benefit. You stop wasting attention on engines that don't predict for your specific patterns, and you sharpen your focus on the ones that do. If Vedic dashas consistently land for you, the platform amplifies them. If Mayan day signs leave you cold, the platform dims them - but never deletes them, because you can always re-enable any engine the system marked Quiet.

The platform runs two grading systems side by side. They check different kinds of claims and together they shape each engine's trust state.

• Reflect grading (automatic). The biorhythm engine generates one to three falsifiable predictions every day - things like "sleep around 7.0h" or "energy peaks at 10am". Tomorrow, after your day closes, the engine compares each prediction against what you actually logged. Hits show ✓, misses show ✗. No user action needed - the loop runs itself.
• Did this hit? chips (manual). Beneath any reading on the platform - your Life Path number, your Saturn return window, your Tzolkin day sign, today's biorhythm forecast - there's a small chip asking "Did this hit?". Tap Yes, Maybe, or No. Your answer goes into the calibration ledger.

Both systems feed the same verdict: a per-engine trust state. Reflect grades the bio engine's falsifiable claims with tight numeric tolerances. The chips collect your subjective judgment on every other engine's symbolic readings. Together they measure two different things and converge on one honest signal: does this engine actually predict for your life?

The chip is small and quiet. It sits below an engine reading and shows three buttons: Yes / Maybe / No. Tap one. That's the whole interaction.

Where they appear. Below every interpretive reading the platform writes for you. Numerology Personal Year, Western Saturn return, Vedic Dasha period, Mayan day sign, Chinese animal interaction, biorhythm daily forecast, Brahman synthesis, Zumi archetype reading - every reading carries a chip.

Why your honest No matters as much as your Yes. The system learns from misses. If you tap No on a prediction the engine made, that's not a failure - that's a signal that helps the platform mark the engine as Quiet for you. Honest rejection is data. Polite-yes inflation is the failure mode that breaks calibration.

Maybe is real too. Use it when the reading partially landed or you genuinely cannot tell. The grading math weighs Maybe as half a Yes, so it neither inflates nor deflates the engine's standing.

Changing your mind. Every rated chip shows a small Change my mind link at the bottom. The system honors revisions. Your most recent answer counts.

The biorhythm Reflect AI is the platform's most rigorous grader. Once you've logged at least three daily inputs, the Reflect on today button appears. Tap it and the engine generates three cards plus one to three falsifiable predictions.

The predictions. Each prediction names a specific metric and an expected value with a tolerance: "sleep around 7.0h", "stress around 4", "today's energy around 72". Tomorrow, after your day closes, the engine compares each prediction against your actual logged value.

Tolerances. The engine grades against per-metric windows:

• Sleep hours: ±1.0h
• Stress, pain, caffeine, sugar, alcohol: ±1 point
• Energy and mood averages: ±0.7
• Today's Energy score: ±8 points

Inside the window = ✓ hit. Outside = ✗ miss. Mixed or unknown = ◐ neutral. Over 30 reflections you get a per-metric accuracy score - an honest signal of how well the bio engine predicts you specifically.

The day must be closed. Predictions only grade against a closed-out snapshot. If you forgot to close yesterday, the engine auto-closes it at midnight using whatever you logged. Auto-closed days still grade.

Each engine's trust state is calibrated to YOUR data, not some abstract truth. There are three states:

• Speaks for you. This engine's readings consistently land for your life. The label shows on chips at five or more ratings with a hit rate at or above 66%. The badge glows. The engine rises in any sort.
• Still listening. Either fewer than five ratings, or a hit rate between 33% and 66%. The engine is forming a picture but the platform is honest that the picture is incomplete. Neutral treatment.
• Quiet. Five or more ratings at a hit rate at or below 33%. The engine has had its chance and the readings did not match your life. The engine stays visible but soft, with a Re-enable link so you can keep it loud if you want.

Two threshold systems coexist. What you see on chips uses the simpler five-rating threshold above. There's also a deeper server-side calibration that uses Pearson r² and requires 30 or more outcomes - that shows on the engine page's trust-state badge. Both report the same three states; the server adapter is more conservative because it's based on stricter statistics. For day-to-day reading you can trust the chip-level state; for the deepest fit, read the engine page's state.

Once you've rated more than one signal of the same kind, every rated chip shows a small inline tally: "73% match · 22/30 ✓". That's your cumulative hit rate for that family of signals.

What counts as a family. Daily readings (like the Personal Day, today's Tzolkin energy, the biorhythm forecast) form one family per engine. Each rating you give a daily reading rolls into that engine's daily hit rate. Identity readings (Life Path, birth sign, Mahadasha) are singletons - one signal, one rating, no cumulative needed.

Why it shows on the chip directly. So you don't need to leave the page to see your standing with an engine. The chip is the smallest possible window into your fit; the Your-fit dashboard is the wider one.

The trust label appears alongside the tally once you cross five ratings on a family. Below five, the tally still shows but the label is hidden - the platform won't make a claim about an engine until there's enough data to mean something.

The Your-fit tab is the panoramic view. Open your dashboard and tap the 📊 Your fit tab to see, at a glance, how every engine is doing for you specifically.

What it shows.

• Per-engine hit-rate bars. One bar per engine, sorted by hit rate. Visible immediately - no clicks needed.
• Trust state badges. Each bar carries its current state (Speaks for you / Still listening / Quiet) for quick scanning.
• Time-series chart. Below the bars, a line showing how each engine's hit rate has shifted week by week. Honest engines stabilize. Noisy engines wobble.
• Re-enable controls. Quick toggles to force any Quiet engine back to full strength.

The tab gets populated automatically as you rate chips. Around your fifth rating, the first trust labels start surfacing. By thirty ratings the picture is meaningful. By a hundred you can see exactly which traditions speak truth for your specific life.

A Quiet engine still renders. The platform never deletes a tradition from your reading - it just visually dims the engines that don't predict for you so the loud ones earn your attention.

Why re-enable an engine? Maybe you find Mayan day signs personally meaningful even though the system marked Mayan as Quiet for you. Maybe you want to keep an engine visible while you give it more time to prove itself. Maybe you disagree with the platform's reading of your patterns. All valid.

How to do it. Tap the Re-enable link on any Quiet engine's card. The engine returns to full strength on your dashboard, your spiral, and your daily readings. The override persists across devices via cloud sync.

The reverse is also true. If you want to silence an engine the platform marked "Speaks for you" because you don't personally find it meaningful, you can force it Quiet the same way. The override is per-engine, reversible at any time, and never affects how the underlying math runs - just what you see.

Co-Star, AstroSage, Nebula, Tarot.com, Astrotalk, Prokerala - every major astrology and biorhythm platform on the market gives you readings and asks for nothing in return. No grading, no calibration, no trust state. They want you to trust them on faith.

imotionengine is built on the opposite assumption: the only way you can trust a reading is if the platform has been honest about its misses. Honest misses build credibility that polite-only successes never can.

This is harder to ship. It means the platform sometimes has to tell you that your favorite tradition isn't predicting for you, or that the engine you don't care for is actually nailing it. That's the trade. We think the trade is worth it. You decide.

This section is for users who want to understand the statistics. You don't need any of this to use the platform.

The chip-level system uses hit rate over a sample. With n ratings on a family and k Yes answers (Maybe counts as 0.5), hit rate = k/n. The thresholds:

• ≥ 5 ratings AND hit rate ≥ 66% → Speaks for you
• ≥ 5 ratings AND hit rate ≤ 33% → Quiet
• ≥ 5 ratings AND hit rate between 33% and 66% → Still listening
• < 5 ratings → trust label hidden, tally still shown

The deeper system uses Pearson r². A quick explainer first. Pearson r is the correlation coefficient between two variables, on a scale from -1 to +1. Zero means no relationship. Plus one means they move in perfect lockstep. Minus one means perfect inverse. Square it and you get r², which has a beautifully concrete meaning: the fraction of variation in one variable that's explained by the other.

Here, the two variables are (a) what the engine predicted for you and (b) what actually happened in your life. If an engine reaches r² = 0.20 against your outcomes, that means twenty percent of the day-to-day swings in how your life felt are accounted for by what that engine forecast. Twenty percent might sound modest, but for behavioral data it's a genuinely strong signal - most clinical psychology studies celebrate r² values of 0.05 to 0.10.

The thresholds:

• r² ≥ 0.20 over ≥ 30 outcomes → Speaks for you
• r² between 0.05 and 0.20 OR < 30 outcomes → Still listening
• r² < 0.05 over ≥ 60 outcomes → Quiet

Why two systems? The chip-level one fires fast and gives you a quick sense of fit. The Pearson-based one needs more data to call an engine, so it's slower to move but more conservative. Both report the same three states and they generally agree once you have enough ratings. Reading the chip is enough for day-to-day use; the deeper state is reachable via the engine page's How well does this engine know me? expander.

Tolerances and sample sizes vary per engine. Biorhythm uses tight numeric tolerances because its predictions are quantitative. Vedic and Western use looser tolerances on symbolic readings because the predictions are qualitative. The exact tolerance per metric lives in the engine page's deep accordion.