Published 2026-09-06 19:02:20 CEST
Trading Lab — Experiment Brief
First screen · Sep 6, 2026 · Paper money only · $1,000 start
Best idea so far
BTC “fast vs slow average”
SMA 5 / 20 — short average vs long average (simple trend rule)
Paper growth
$1,000 → $6,196
+$5,196 · fees already subtracted
Worst dip along the way
about −18%
from a peak (drawdown = peak-to-trough drop)
Beat buy & hold BTC
$6,196 vs $4,619
same period, same starting cash
What we tested
plain English
- We compared simple rules for when to be invested in Bitcoin (BTC-USD) and Algorand (ALGO-USD) vs sitting in cash.
- Period: Jan 1, 2023 → Sep 5, 2026, using one price per day (not minute-by-minute).
- Costs: we subtracted trading fees so results are a bit pessimistic.
Story cards
what happened
1. Buy & hold baseline
BASELINE
Just buy once and hold — no timing rules.
BTC
$1000→$4619 (+362%)
BTC rough ride
−53%
ALGO
$1000→$567 (−43%)
ALGO rough ride
−83%
2. Winner — short avg above long avg
WINNER
Rule: be invested when the 5-day average is above the 20-day average (trend up); otherwise cash. On BTC this was the standout.
BTC 5&20
$6196
Worst dip
−18%
Trades
41
Same rule · ALGO
$4290
Same rule on ALGO: $4290 — a huge upgrade vs holding ALGO ($567).
3. Loser family — “Trend + RSI”
LOSER
Rules that combine trend with RSI (a “hot / cold” momentum gauge). Mostly never traded — stayed at $1000. Not useful yet.
Typical end
$1000
Verdict
Drop / redesign
My take
researcher
Lab pick
SMA-style trend following on BTC — strong on paper, smoother than buy&hold.
Still testing
Later: confirm it wasn’t one lucky bull market (split the years).
Dropped
For now: trend_rsi as configured.
Next
Suggested (for Othman to approve): stress-test SMA 5/20 on 2023 / 2024 / 2025 separately before adding new strategy families.
Path thermometer
where we are