Feed fresh · 115 signalsGitHub View signals

CurrenSystematic crypto signals

Every signal.Measured in R.

Entry. Stop. Peak. Every published signal on one record.

115 signals · 36% reached +1R · median peak +0.76R

Open the record Signals on Telegram
Public record · fresh
Newest publisher snapshot ≤ 90s old · synced 18:16 UTC
115
Signals
36%
Reached +1R
+0.76R
Median peak R
37
Stopped at −1R
1
Active · last reported

Outcomes of 89 signals recomputed under the 0.5R rule from 1-minute exchange candles, as of 2026-10-06 13:13 UTC. 14 historical no-fill signals independently replayed as hypothetical market entries at the first full Bybit 1-minute candle; original risk distance rebased; +0.5R to break-even. Their Peak R and closures are retrospective simulations, not actual fills or recorded Telegram outcomes.

01Mechanics

One entry. One stop. One number.

Every signal begins with an entry and an initial stop. Peak R shows its best recorded move.

+3R+2R+1R0R entry−1RStop −1R · live until +0.5R+0.5R · stop off
Entry · 0RTimestamped, hashed
ExcursionMarked to Bybit candles
Peak R+2.84RMax favourable excursion.
Peak holdsRatchets up, never down
Back at entryTracking ends · no loss
  1. 01

    Entry

    Signal entry and initial risk are recorded.

  2. 02

    Risk

    Initial risk is −1R. Stops are recorded.

  3. 03

    Peak R

    Best observed movement in units of initial risk.

  4. 04

    Ratchet

    The highest observed R is retained.

  5. 05

    End of tracking

    A return to entry ends tracking under the current rule.

Illustrative path — not a recorded signal.Historical OHLC · no fills
02The public record

The full population.

Every published signal. Peaks and stop-outs, in order.

Feed · freshSynced 2026-10-30 18:16 UTC
Σ Peak R+179.7RSum of separate observed peaks · not earned.

Σ Peak R+179.7R · observed peaks, not earned.

Outcomes of 89 signals recomputed under the 0.5R rule from 1-minute exchange candles, as of 2026-10-06 13:13 UTC. 14 historical no-fill signals independently replayed as hypothetical market entries at the first full Bybit 1-minute candle; original risk distance rebased; +0.5R to break-even. Their Peak R and closures are retrospective simulations, not actual fills or recorded Telegram outcomes.

/
Showing 115 of 115 · chronological
Record landscape · complete population · publication order

One mark per signal. Hover for exact R; stop-outs included.

Peak R stemPeak R · best case reachedStop loss · −1R
MAR 2026APRMAYJUNJULAUGSEPOCT
Selected#115 · OPUSDT
Peaks are recorded maxima reached at different, unpublished times. They are not simultaneously realizable and do not add up to a return.Order: publication. Never sorted by outcome.
All signals · scroll, then select
115 signals · scroll the listOpen the full record
#115 · sig_746b4049a205826dda65beb1

OP

Long
Tracking ended
Peak R · best case reached
+2.13R
+5.99% max favorable move

Best price excursion. No fills.

Entry · 0R0R28 OCT 2026 · 17:30 UTC
Peak+2.13RBest case reached · +2.13R
TrackingEndedBack at entry after the peak · no longer tracked
Signal price history
PRICE ACTIONLoading candles
Loading exchange candles…
Published28 OCT 2026 · 17:30 UTC
TRACKING ENDED29 OCT 2026 · 01:58 UTC
Peak candle estimated · UTCHistorical OHLC · no fills
Record sig_746b4049a205826dda65beb1Inspect signal
03Alpha OS

Hypothesis in. Signal out.

Hypotheses are tested on point-in-time data. Only validated ideas move forward.

01

Hypothesis

A market idea, stated so that it can fail.

02

Point-in-time research

Tested only on data that existed at the time.

03

Falsification · validation

Weak ideas are killed here, before deployment.

04

Alpha registry

Survivors are versioned, promoted and tracked.

05

Distribution

Decisions reach people and machines.

SignalsTeamsAPIsAgents
01HypothesisA market idea, stated so that it can fail.
02Point-in-time researchTested only on data that existed at the time.
03FalsificationWeak ideas are killed before deployment.
04Alpha registrySurvivors are versioned and promoted.
05Signals · Teams · APIs · Agents
Schematic. Line counts illustrate attrition, not data.Killed by evidence
01In development

Research

Formalize market ideas. Test them against point-in-time evidence. Kill weak ideas before deployment.

HypothesesPoint-in-time dataFalsification
02Available

Signals

Actionable market decisions with lifecycle tracking and a public record that keeps every outcome.

LifecyclePublic recordTelegram
Explore the record
03Open-source alpha

Agents + developers

Read-only programmatic access to the same record, for your code and your agents.

APIPythonCLIMCP
See the interface
04Agents + developers

Same record. API, CLI, MCP.

The object on this page is the object your code reads — same fields, same hash, read-only. No endpoint can place an order.

Human interfaceRecord view
crn_sig_0098Example record
SymbolBTCUSDT
SideLong
StatusTracking ended
Peak R+3.42R
Published14 SEP 2026 · 08:15 UTC
Peak R = maximum favorable price excursion.
Machine interface
› curren_get_signal("crn_sig_0098"){ "symbol": "BTCUSDT","side": "long","peak_r": 3.42,"published_at": "2026-09-14T08:15:00Z" }// excerpt · illustrative values · not a live call
Field names follow the public API contract (v0.4, alpha). Response shown as an excerpt; values are illustrative, not a live call.
v0.4 · alpha

Read-only API

Signals, lifecycle, results and verification over HTTPS. No endpoint can place an order.

GET /v1/signals?status=active
v0.4 · alpha

Python client

The async client the CLI and MCP server are built on. Restricted fields are never reconstructed.

curren.client
v0.4 · alpha

CLI

Signals, lifecycle, results, track record and verification from a terminal.

curren signals --symbol BTCUSDT --json
v0.4 · alpha

MCP tools

Six read-only tools for agents, from active signals to verification.

curren_get_signal · curren_verify_signal
Developer docs Source on GitHub
05Methodology

Out-of-samplebyconstruction.

Published before the result. Every signal stays on the record.

Entry, stop and publication time are hashed on write. Results are appended, never edited.

EvidenceLook-ahead biasSurvivorship bias

A backtest

Built inFitted after the data.LikelyLosing parameter sets disappear.

A screenshot

Built inChosen after the outcome.TotalOnly winners get posted.

A Curren signal

NoneTimestamped before the move.NoneStop-outs stay in the denominator.
The decisionHashed when first recorded
Published14 SEP · 08:15 UTCSymbol · side · publication timecannot be rewritten
What happened nextAppended, never rewritten
Peak ratchetswhile trackedTracking endsPeak R frozen

Each publication is hashed when Curren first records it, and the final Peak R is frozen when tracking ends; curren verify <id> checks both. These hashes detect changes inside Curren’s own store — they are not an independent notary or a profitability guarantee. Illustrative chain.

Read the full record.

Open the public record Signals on TelegramBuild on the API
View signalsTelegram