Right now has happened before.

Underspace searches market history for the moments that most resemble the one on screen, then shows you the distribution of what actually followed.

We can tell you how often, and what came next.

We can't tell you what happens this time.

Open the chart
Create a free accountFree account · no card

A live read

What the Statistics panel says

This is the panel a signed-in user sees on the chart. The values below were observed in the product; the page does not fetch live data.

BTCUSDT.P · 1hStatistics

124 past moments matched this technical state.

Within 6 hours

55.5% went up5.4% flat39.1% went down
Avg movement+0.42%
vs all bars+4.7pp
Largest up+3.85%
Largest down−2.90%

Within 12 hours

49.8% went up5.6% flat44.6% went down
Avg movement+0.05%
vs all bars−1.0pp
Largest up+5.12%
Largest down−4.31%

Statistical observation, not a recommendation.

Method

How a match is made

  1. 01

    Describe the moment

    The current bar is reduced to a small set of technical features: distance from its moving averages, trend strength, volatility. No news, no sentiment, no narrative.

  2. 02

    Search history

    Every bar in the lookback window is scored on the same features. Bars that land in the same bucket as the present are matches; consecutive matches collapse into one episode so a single stretch is not counted twice.

  3. 03

    Report what followed

    For each matched episode, the forward movement over fixed horizons is recorded. What you see is the distribution of those outcomes — not a forecast drawn from them.

Why the second number exists

Every read is shown against its base rate

50.8%
of matched moments went up. Looks like a coin flip. On its own, it is.
+4.7pp
The number that matters: the difference from how often any bar in the same period went up.

A percentage alone flatters almost any situation, because markets drift. So the panel always shows the same statistic computed over every bar in the lookback, and reports the gap between the two in percentage points.

The gap is allowed to be zero. In the read above, the same technical state at the 12-hour horizon shows −1.0pp — indistinguishable from the base rate — and the panel says so rather than hiding the horizon that didn't work.

Limits

What this does not do

It does not predict
It reports how often something followed a similar state in the past. Frequency is not a forecast.
It does not recommend
There are no buy or sell signals on this panel, and none are implied. Every screen that shows a distribution also says so.
Matching is not a mechanism
Two moments can share technical features for unrelated reasons. Resemblance says nothing about why either moved.
Small samples happen
Some states have few precedents. Below roughly ten matches the panel warns instead of asserting.

Who builds this

Two people, building carefully

Underspace is developed by two people. The public claims on this page are deliberately narrow, because the product's whole premise is that unaudited claims are worthless.

  • Founded 2025
  • Two people
  • Product live at underspace.ai
  • Real capital: not yet deployed