How Prexora works, in plain English

A short manual for traders who want to use Prexora signals on Polymarket. No jargon you don't need. Methodology depth lives at /methodology.

Getting started

1. Open Live signals. You'll see today's open Polymarket weather brackets where we have edge.
2. Each card shows the city, the bracket (e.g. "≥ 0.5 inches of rain"), our model's probability vs the market's, and the edge in percentage points.
3. Click COPY on any card to open Polymarket with the bet pre-staged. You'll complete the trade in your own wallet — we never touch your money.
4. (Optional) Subscribe to Pro for Telegram alerts and personalized stake sizing.

How Prexora works

For every weather market on Polymarket, Prexora pulls forecasts from 5 numerical weather models (Wunderground, ECMWF, GFS, ICON, NWS) plus METAR observations. We compute the probability the actual high temperature lands in each bracket and compare it to the market's implied probability. If our number diverges from the market by enough — and a stack of gates pass — we publish a signal.

The system also auto-learns: every resolved bet feeds a per-city σ calibration and a per-detector confidence multiplier. Detectors that lose money get auto-down-weighted. (city, direction) pairs with poor track records get auto-blacklisted. The system gets stricter over time, not looser.

Anatomy of a signal

Each card shows:

• City + bracket — e.g. "Boston ≥ 0.5"
• Side — YES or NO based on edge direction
• Edge — model probability minus market price, in percentage points
• Strength — Strong / Moderate / Soft, derived from edge × model agreement × sigma
• Models agreeing — out of 5 forecast models, how many point the same direction
• Bracket type — floor (≥ X), ceiling (≤ X), or range/single — affects how the math integrates

Tier-gating (post-token launch)

Once $PREXORA launches, holding ≥ 1,000 PRXR unlocks the Member tier: pre-publication signals (you see them before the public dashboard), full LLM rationale per signal, custom alert thresholds. Pro subscribers who stake PRXR get a 50% subscription discount and a 5% Auto-trade performance fee (down from 10%).

Until the token launches, all current features are open to everyone. You're not losing anything by waiting.

Copy trading

We support two flavours:

1-click copy (free for all users). Each signal card has a COPY button. Click it → confirmation modal → open Polymarket with the bet pre-staged → complete in your wallet. ~15 seconds vs ~2 minutes manual. We never touch your funds.

Auto-trade (Pro tier, ships in weeks 5-6). You paste your Polymarket API credentials into Prexora settings (Polymarket itself restricts these to trading-only — no withdrawals possible). You configure risk rules: max % per bet, daily loss cap, strength filter, allowed cities. When a signal fires, our bot places the bet on your Polymarket account within your rules. You can revoke anytime by deleting the API key on Polymarket.

Telegram alerts

Join the public channel @prexora to get every new pick, win, daily summary, and weekly recap. No signup required. The bot DM (Pro tier) sends you filtered, personalized notifications — only the signals matching your strength / city / bracket-type preferences.

One rule: we don't broadcast losses on Telegram. Losses still appear on the public Lesson Stream where we publish the AI's reflection on every resolved bet. Telegram is for momentum; lessons.html is for honesty. Both surfaces serve different purposes.

FAQ

Are these guaranteed picks?

No. Nothing is guaranteed. Our backtested all-time win rate is ~71.7% on resolved bets, but past performance is not a promise. The methodology page shows our limitations honestly. You are responsible for every bet you place.

Why are some markets shown as YES and others as NO?

For each bracket, we compute the model probability. If the model probability is higher than the market's YES price, we publish YES (the market is under-pricing). If the model probability is lower, we publish NO (the market is over-pricing). The "edge" is the difference in percentage points.

What's the difference between Strong, Moderate, and Soft?

Strong = edge ≥ 15pp AND ≥ 3 models agree AND σ ≤ 1.6. Moderate = edge ≥ 8pp AND ≥ 2 models. Soft = edge ≥ 5pp. Strong picks have historically had the best WR (~74% on backtest).

Why "weather" specifically?

Three reasons: (1) prediction-market weather brackets are public, liquid, and resolve daily, (2) numerical weather models give us five independent forecasts to blend, which is hard to replicate in elections or sports, (3) the markets are large enough to matter ($K-$M volume) but small enough that we can have edge.

Will Prexora ever expand beyond weather?

Yes — sports, elections, and crypto verticals are roadmapped for late 2026 / early 2027. They become available to Trader-tier subscribers ($129/mo) and PRXR token holders.

Is this US-regulated?

Prexora is a research and analytics service. We don't accept funds, take custody, or execute trades on your behalf without explicit per-trade authorization (Auto-trade tier). Polymarket is the regulated venue you trade on; consult your local laws.

Glossary

Risk disclosure