Coinseekly - AI driven crypto insights, details, patternscoinseekly
Back-tested on real history

Our signals, graded against real price history

Anyone can claim their signals work. We ran ours over 5.0 years of daily candles across 30 coins and graded every one — then split the results by bull and bear markets so you can see the edge isn't just a rising tide.

Signals graded
3,027
Bullish signals · 30d
+5.2%

1,486 signals

Buy & hold · 30d
+2.8%

any random day

Bullish edge · 30d
+2.4%

vs buy & hold

Across the full window, bullish signals returned +5.2% over 30 days vs +2.8% for buying a random day — a +2.4% edge.

Best-performing signal · 30 days
Golden cross Bullish
Win rate 50%Avg return +9.1%Fired 78×
MACD bullish cross Bullish

MACD line crosses above its signal line — momentum turning up.

Fired 1,058
Win 7d/30d 45%/46%
Avg 7d/30d -0.2%/+5.2%
MACD bearish cross Bearish

MACD line crosses below its signal line — momentum turning down.

Fired 1,041
Win 7d/30d 54%/58%
Avg 7d/30d +0.1%/-3.3%
RSI overbought Bearish

RSI(14) pushes into overbought territory (≥ 70) — elevated pullback risk.

Fired 424
Win 7d/30d 47%/52%
Avg 7d/30d -4.2%/-11.1%
RSI oversold Bullish

RSI(14) drops into oversold territory (≤ 30) — historically a bounce zone.

Fired 350
Win 7d/30d 59%/51%
Avg 7d/30d +1.9%/+4.6%
Golden cross Bullish

50-day SMA crosses above the 200-day SMA — a classic long-term uptrend signal.

Fired 78
Win 7d/30d 47%/50%
Avg 7d/30d +5.4%/+9.1%
Death cross Bearish

50-day SMA crosses below the 200-day SMA — a classic long-term downtrend signal.

Fired 76
Win 7d/30d 48%/50%
Avg 7d/30d +0.2%/-4.3%
Buy & hold — any random day
baseline
Avg 30d +2.8%

How we measure this

  • Universe & window. 30 top coins, daily candles from Jul 2021 to Jul 2026.
  • Trigger. Each signal is counted only on the bar it transitions (e.g. RSI crossing below 30), never every day it stays there — so the sample reflects fresh signals, not double-counts.
  • Grading. We take the closing price the day a signal fires and compare it to the close 7 and 30 days later. A bullish signal “wins” if price rose; a bearish signal wins if it fell. Average return is signed in the signal's favour.
  • Measured against a baseline. Over this window the market rose, so simply buying on any random day returned +2.8% in 30 days. A signal only has real edge if it beats that — which is why we show the baseline next to every number, and split everything by bull/bear regime.
  • Bull vs bear. A day is “bull” when Bitcoin closed above its 200-day moving average, “bear” when below — the standard market-regime proxy. Each signal is bucketed by the regime on the day it fired.
  • Same maths as live. The exact RSI, MACD and moving-average code that grades history is what powers the live screener and alerts — nothing is cherry-picked. No fees, slippage or position sizing are modelled.

Want these signals the moment they fire?

CoinSeekly scans the market around the clock and alerts you by email and Telegram the instant a signal triggers — on any coin you track.

Start free — 7 days of Premium Open the live screener

Want the full analysis behind these numbers — key findings, regime tables, and the chart-pattern statistics? Read the signal win-rates study →

Want to see these signals traded live, every position public? Follow the agent portfolios →

Past performance is not a prediction. These figures are a historical back-test of mechanical signal rules over past price data — they do not include fees or slippage, ignore position sizing, and are not a strategy you can blindly trade. Crypto is volatile; signals fail. This is educational, not financial advice. Always do your own research. See our methodology for how each indicator is computed.