How it works
Make your own call first. Then learn from the AI.
Every day you commit to a view first — before the crowd or the AI. Only then do they appear, next to your own read, so you can see where you differ and learn from it. Here's the whole system — skim it in 90 seconds, or read it fully in about four minutes.
Judgment has many parts. We measure the most measurable one — how well you call uncertain outcomes — and help you get better at it.
Step 1
Every day, real questions arrive
A fresh set of questions about real-world events — macro, geopolitics, elections, crypto, AI & tech, sports, biotech — and the occasional wildcard.
You engage with what's actually happening in the world, curated into what matters.
Step 2
You make your call — first
You predict each one with a confidence level, while the crowd and the AI are still sealed.
The honest test — your real judgment, uninfluenced. No following the herd.
Step 3
The reveal: your call beside the crowd's and the AI's
Once you've predicted, see the crowd lean, the AI's probability and its one-line take. Members also get the AI's full analysis.
See how your read compares — and study where you and the AI diverge. That gap is where you learn most.
Step 4
The world decides who was right
The real event happens. The question resolves against a named source. You find out.
Real feedback on your judgment — not opinions, outcomes. This is how you actually get sharper.
Step 5
Your track record builds
Every result moves your Presage Rating. Over time, a verified history of your judgment forms.
You build calibrated, provable judgment — the rarest skill in the age of AI.
Most learning is textbook learning. This is judgment, learned the only way it can be — by doing.
Sharpen it. Prove it. Trust it.
The details
How you forecast
Instead of a vague guess, you pick one of four confidence levels — a direction and a strength:
Why there's no 50%
Forecasting is about taking a side. “50%” is just another way of saying “I don't know” — so we don't offer it. Committing to a direction and a strength is what turns a hunch into a real, measurable test of judgment.
How scoring works
Good forecasting isn't just being right — it's being right and appropriately confident. Being confidently right scores best. Being confidently wrong is penalized most. Playing it cautious lands in between.
We measure this with the Brier score — the standard, respected way to measure forecasting accuracy, the same method used in serious forecasting research. The math is simple in spirit: it measures the gap between your forecast and what actually happened, so smaller is better.
Worked example
You forecast Very Likely (90%) that the Fed holds rates.
The Fed holds
Confidently right
You committed and nailed it.
+24
The Fed cuts
Confidently wrong
You committed the wrong way.
−32
You'd hedged at 70%
Right, but cautious
Correct — yet you hedged.
+16
The confident, correct call gains the most; the same confidence on the wrong side loses the most; hedging trims both the upside and the sting.
Moves are capped at ±32 per question, and no rating falls below 700.
Only professional-lane questions — macro, geopolitics, elections, crypto, AI & tech, biotech — count toward your rating. Recreational questions (sports and the occasional wildcard) are just for fun and never touch your credential.
Your Presage Rating
Your Presage Rating is your headline number — a rating that works like a chess rating, rising and falling with the quality of your forecasts over time. Everyone starts at the same baseline of 1000 and earns their way up.
Each resolved professional question nudges it: a confident correct call moves it up meaningfully, a confident wrong call moves it down, and hedged calls move it less.
Verification tiers
You also earn a tier based on how many forecasts you've made — not how good they are — because a credential needs a track record to be trustworthy.
Newcomer
under 100 forecasts
Verified
100+ forecasts
Expert
500+ forecasts
Elite
2000+ forecasts
Everyone earns a tier; Members can display the verification seal on their public profile.
By category
Presage also tracks your accuracy by category— macro, geopolitics, elections, crypto, AI & tech, biotech and more — so you can see, and prove, exactly where you're sharpest.
Every forecast is timestamped before the event, scored by fixed math, and permanent — which is why the rating is a verified, permanent record of your judgment.
Rating vs. activity
Two numbers, two different jobs:
Activity
Participation & streaks
Tracks participation, streaks, and correct directional calls. It only ever grows — a measure of how much you show up, not how calibrated you are.
Presage Rating
Calibration vs. outcomes
Measures calibration against real outcomes — how right you are, and how appropriately confident. It rises and falls with the quality of your forecasts.
Rankings use your Presage Rating — calibration against real outcomes. You need 10 resolved professional forecasts to be ranked; until then you're listed as Unranked. Activity shows alongside as a secondary measure.
Why you can trust it
Forecasts lock the moment you submit — the crowd and AI aren't revealed until then, so there's nothing to copy.
Every call is timestamped and permanent — the record can't be quietly edited.
Scoring uses fixed, transparent math — identical for everyone.
Points have no monetary value. This is a measure of judgment — not a betting or trading product.
The Presage is a research and education platform. It is not investment, legal, or financial advice.
Ready to make your first call?
Start sharpening — free
Measured, not asserted. That's the whole idea.