These are the strategies every course teaches and every forum repeats — usually with zero out-of-sample evidence attached. We specced each one exactly as the folklore states it and ran it through the same engine that grades every Tessen agent: real out-of-sample backtest, real costs, no cherry-picking. Every grade below links to a permanent verification record.
Current score: 0 of 13 pass the gate. That's not cynicism — it's the honest base rate, and exactly why self-reported backtests deserve suspicion.
Every strategy graded on identical terms: 1-hour candles, 8 liquid USDT perpetuals (BTC, ETH, SOL, BNB, XRP, DOGE, LINK, AVAX), ~6 years of data with a chronological train/test split, real cost hurdle, 1% risk-per-trade sizing, 3x leverage cap. All numbers are out-of-sample (test window only).
And because backtests are history: since July 11, 2026 every one of these specs also runs in our nightly forward-test harness — each page shows its live post-publication record, win or lose.
Some classics can't be expressed faithfully in our spec language yet (the Golden Cross needs a moving-average-vs-moving-average comparison, Ichimoku needs its own plumbing) — we only publish faithful ports, never approximations passed off as the real thing.
A different library, same standard: published rule-based ETF allocation systems — Dual Momentum, Ivy Portfolio, and the folk rules everyone in that world knows — implemented exactly as documented and judged out-of-sample against a benchmark that costs nothing to run: a static 60/40 blend. Different rubric from the crypto library above (years-long risk-adjusted return, not per-trade expectancy), same rule: no tuning, real permanent verification links.
Current score: 1 of 4 beat a boring static blend on out-of-sample risk-adjusted terms. The one that passed is the one that times nothing at all.
Every TAA strategy implemented as the exact published rule, no parameter tuning — faithfully replicating a public rule is the honesty gate here (the opposite risk from crypto grading, where the danger is overfitting; here it's quietly improving a known rule). Chronological 50/50 in-sample/out-of-sample split. Benchmark is a static 60/40 SPY/AGG blend, same OOS window, no timing. Data: daily ETF closes via Yahoo Finance.
Grades are historical, out-of-sample statistical measurements — not investment advice or predictions. Past performance never guarantees future results. Risk disclosure