Skip to content
alphr

Humans vs AI

Who reads crypto markets best?

Human analysts, three simulated AI profiles and the crowd forecast the same Rounds. Everything below is derived from settled results. How scoring works →

Humans

58%

129/223 correct · 18 Rounds won

vs

AI analysts

28%

19/69 correct · 5 Rounds won

Humans · 7d

55% (67)

AI · 7d

29% (21)

Humans · 30d

58% (223)

AI · 30d

28% (69)

Ties: 0. Same Rounds, same rules, same clock.Methodology

Rounds decided

23

Humans 18 · AI 5 · Ties 0

Best-performing signal

Market Trend

63% accurate across 92 settled human forecasts

Human vs AI gap (30d)

+30%

Positive favours humans

Performance trend

Share of correct forecasts per settled Round, humans vs the three AI profiles.

AI analysts

Three profiles, three styles

These are deterministic rule-based simulations built for Alphr. They are not commercial AI models and their forecasts are not financial advice.

ATLAS

Weighs trend, momentum and volume

Patient and systematic. ATLAS only commits when the multi-day trend, short-term momentum and volume agree, and sits Neutral when they don't.

Rule-based simulation · atlas-v1.0.0

Accuracy
35%
Settled
23
Rating
40.0

PULSE

Rides short-term momentum

Quick and decisive. PULSE follows the last one to three days of price action and is rarely Neutral.

Rule-based simulation · pulse-v1.0.0

Accuracy
26%
Settled
23
Rating
34.8

DRIFT

Reads sentiment, fades crowded moves

Contrarian by temperament. DRIFT follows the narrative until price runs too far from its weekly mean, then calls the reversal.

Rule-based simulation · drift-v1.0.0

Accuracy
22%
Settled
23
Rating
32.2

Recent Rounds

Biggest human/AI disagreement

BTC Daily Round

Crowd majority (89%): Neutral

AI majority: Bearish

Outcome: Neutral — Correct the crowd was right.

Methodology, void rules and data limitations are documented on the methodology page.