The scoreboard so far — Vietnam (backfilled)
| Strategy | Trades | Win rate | Return | Buy & hold | Difference | Worst fall | Hold’s worst | Per unit of pain | Hold’s | Days holding |
|---|---|---|---|---|---|---|---|---|---|---|
| RSI(2) mean reversion | 451 | 47% | -30% | +411% | -441% | -37% | -58% | -0.8 | 7.0 | 8% |
| Overnight edge | 9,530 | 9% | -100% | +60% | -160% | -100% | -52% | -1.0 | 1.1 | 91% |
| Time-series momentum | 852 | 50% | +209% | +821% | -611% | -54% | -61% | 3.9 | 13.5 | 53% |
| Time-series momentum, volatility-scaled | 852 | 52% | +558% | +821% | -262% | -68% | -61% | 8.2 | 13.5 | 53% |
| Dual momentum | 827 | 50% | +200% | +821% | -621% | -54% | -61% | 3.7 | 13.5 | 52% |
| Basket trend (Vietnam only) | 48 | 56% | +1018% | +904% | +114% | -28% | -63% | 36.6 | 14.4 | 63% |
Read the difference column before anything else. Every one of these strategies is being asked a single question — did it beat simply owning the same things over the same years? — and a return that looks handsome on its own can still be a way of losing to doing nothing. Worst fall is the other half of it: most of these exist to sit out the bad part, and giving up return to do that can be a fair trade.
Per unit of pain is return divided by the worst fall — how much you got paid for the worst moment you had to sit through. It is the column where most of these stop looking like failures: losing to buy-and-hold on return while halving the drawdown is not the same thing as losing. It is also the column that flatters a strategy which has barely traded, so read it next to the trade count, not instead of it.
$100 — RSI(2) mean reversion
$100 — Overnight edge
$100 — Time-series momentum
$100 — Time-series momentum, volatility-scaled
$100 — Dual momentum
$100 — Trend-gated 3x (US only)
$100 — Basket trend (Vietnam only)
$100 — BTC trend (crypto only)
About these
Live calls only. Everything in the scoreboard above is backfilled — the rules replayed over history they had already seen — and turning that into dollars would be money invented out of hindsight. These start moving when a live trade closes, and each one resets independently in your own browser.
What each one is
RSI(2) mean reversion
In an uptrend, a sharp two-day drop is usually bought back within days.
Overnight edge
Almost all of the index's long-run gain has arrived between the close and the next open, not during the trading day.
Time-series momentum
What went up over the last year tends to keep going, and stepping aside when it has not avoids the worst drawdowns.
Time-series momentum, volatility-scaled
The same signal, but never betting the same stake on a calm month and a violent one.
Dual momentum
Hold it only when it is beating cash, not merely when it is going up.
Trend-gated 3x (US only)
Leverage compounds wonderfully in trends and dies in chop; a trend filter keeps the leverage and steps aside from the chop.
Basket trend (Vietnam only)
Ask the whole basket one question a month instead of asking every stock its own — in a market where the stocks herd together, the basket's answer is the cleaner one.
BTC trend (crypto only)
Nothing tested beats holding Bitcoin — except holding Bitcoin only when it is above its own 100-day average.
What changes on HOSE
Four things here are not true of the US page, and all four change what the numbers mean.
A hard 7% daily band. HOSE stops a stock moving more than ±7% in a session. That caps how far any winner can run on the day, which matters most for the rules that make their money from a few large moves — the momentum three — and least for RSI(2), which is aiming at small ones anyway.
No short selling, which happens to cost nothing here: every one of these five rules is long-only, so unlike the breakout agent there are no signals being counted that could never have been placed.
These are individual companies, not index funds. The US basket is twelve broad ETFs, so a result there is a claim about the market. Seven of the eight here are single stocks, so a win can just be one company having a good year, and that is a different and weaker kind of evidence. E1VFVN30 is the one index fund in the basket — it is the closest thing to a like-for-like comparison with the US page.
Dual momentum is measured against the wrong cash. Its safe asset is BIL, a US treasury fund priced in dollars, while these stocks are priced in dong. Comparing the two mixes an interest rate with an exchange rate. The rule still runs, because changing the safe asset per market would be inventing a variant Antonacci never wrote — but its number here is the least trustworthy on the page and it is listed rather than quietly dropped.
The research programme is concluded
60 ideas tested once each against a fixed bar; 1 survived. Concluded 11 August 2026. Why the zero is credible, the closest miss, the nastiest data trap found, and what the one survivor actually proves — read the findings.