The headline is the medium-term score, the only one validated: low readings have preceded deeper drawdowns, not lower returns. Short and long use the same words but are descriptions, not tested forecasts.
For money already converted to USD — never convert per trade (1.5% each way).
CAD-listed: no currency conversion needed.
● validated index · ○ research-supported sector · For money already converted to USD — never convert per trade (1.5% each way).
● validated index · ○ research-supported sector · CAD-listed: no currency conversion needed.
When a fund that has been rising for months drops sharply for a day or two, buy it at the next morning's open. Sell once it bounces back above its average price of the past week, or after two weeks at most.
On 18 country funds it had never been tried on, it made about 0.2% per trade on average. If the rule had no real edge, a result that strong would turn up by luck only about once in 2,000 tries. On Canadian-listed funds, expect about 0.1% per trade.
Small, frequent gains, not big wins: about 1 trade in 3 loses money, and each trade lasts two weeks at most. The gain is per trade, not per year. Money waiting between trades earns only the cash rate, so this is not a way to beat simply holding XEQT.
Exact rule: Buy at the next open when an ETF closes with RSI(2) below 10 while above its 200-day average. Sell at the next open after it closes above its 5-day average, or after 10 sessions.
Exact evidence: +0.20% per trade on 18 untouched country ETFs (p = 0.0005); about +0.1% expected on CAD-listed ETFs. About a third of trades lose.
Lines show each sector's last six weeks · grouped by quadrant
How to read the three: cash for money needed within a year; for 1–5 years, hold 64% of your usual stock share (it moves daily with conditions); for 5+ years, XEQT on schedule, and lump sums go in now. Waiting for dips lost 2–3% in testing.
What the dial buys: tested on XEQT in CAD from 2023, its worst drop was 7% against 15% fully invested, and it returned about 15% a year against 22%. Simply holding a fixed 79% (its average) fell 12% at worst and returned 17%. It is insurance for money you need soon, not extra return.
| Option | Return | Worst 1y | Worst 3y |
|---|---|---|---|
| Cash | 2.25% today | 0.0% | 0.0% |
| Short-term bonds | 3.23% (2-yr Canada yield) | −6.5% | −4.4% |
| XEQT | ~11% historically, before fees | −35.8% | −33.2% |
History, not a forecast · 4529 windows, 4 independent
| Fund | Market | Weight |
|---|---|---|
| XTOT | US | 31.1% |
| XIC | Canada | 24.8% |
| XEF | Intl | 24.0% |
| ITOT | US | 15.0% |
| XEC | Emerging | 4.9% |
| CAD | Cash | 0.1% |
| USD | Cash | 0.0% |
| 1W | 1M | 3M | 1Y | YTD | High | |
|---|---|---|---|---|---|---|
| SPY | +1.9% | +1.7% | +3.7% | +17.0% | +14.9% | −0.2% |
| QQQ | +2.4% | +5.6% | +4.9% | +25.3% | +23.8% | −0.3% |
| IWM | −0.1% | −5.5% | −6.3% | +13.7% | +13.6% | −8.7% |
| RSP | +1.2% | −2.5% | −1.0% | +11.9% | +11.3% | −5.1% |
| XIC | −0.5% | −2.8% | +0.8% | +17.3% | +12.3% | −4.9% |
| EFA | −0.8% | −4.3% | −0.8% | +11.9% | +9.1% | −5.2% |
| EEM | +0.9% | −2.1% | +0.9% | +26.1% | +23.8% | −5.4% |
| XEQT | +0.7% | +1.2% | +2.5% | +17.9% | +16.4% | −0.8% |
Above 50-day: SPY, QQQ, EEM, XEQT · above 200-day: all 8
A score, not a forecast. Sector ranking showed no edge in testing; the rotation words belong to the rotation graph.
Stocks fell worldwide as yields rose and Fed minutes pointed to another hike; the S&P 500 slipped 0.2% from its record, but the TSX fell 1.7% and XEQT 0.6%.
Open calls, the track record and the full review are in Predictive.
The evidence behind the page: composite detail, sector strength, internals, frozen forward tests and method notes. Research, not guidance.
Risk indicators flashing. Expect turbulence — historically ~2x deeper drawdowns than calm regimes.NEUTRAL LINE = 50
Biggest sector moves since 2026-10-06: XLE -5 · XLV +5 · XLI -5
VIX term structure holds contango (1.18) — options markets price calm ahead; junk bonds are outperforming quality (HYG>LQD); HY spreads widened 0.36pp in a month.
Score attribution (medium-term): breadth is the biggest drag (10/100, -7.9 weighted pts vs neutral); volatility is the biggest support (72/100, +4.5 pts).
Soft underbelly: Solar / Clean Energy (TAN) scores just 4 inside top-ranked Energy (59) — the sector's strength isn't uniform.
Whole-market participation and news tone
Drawdown dial: 66% exposure would historically have meant shallower drawdowns, at a cost in long-run return. Not a reason to hold back long-term money.
Previous session's value, applied during the 2026-10-07 session. Today's value is the drawdown dial in Where New Money Goes.
Buy-the-dip monitor · swing & long-term horizons
Contribution context — descriptive. No validated timing edge.
Waiting one quarter in cash earns about 0.56% risk-free (Bank of Canada rate (Canadian cash) at 2.25% annualised). The comparison, not a recommendation.
Three registered rules, so a verdict requires the 98.3% interval (alpha 0.0167) to exclude zero, not 95%. Fixed before any result exists.
in-sample descriptive summary — NOT a validated edge
| Drawdown | Mean fwd 63d | Positive | n | independent |
|---|---|---|---|---|
| 0-2% | 3.75% | 83% | 905 | 21 |
| 2-5% | 4.43% | 86% | 292 | 19 |
| 5-10% | 2.46% | 63% | 185 | 13 |
| 10%+ | 6.59% | 80% | 225 | 7 |
| any day | 4.12% | 81% | 1607 | — |
Every 63-session window overlaps its neighbours, so independent is the count to judge a row by — not n. Buckets were fixed before any of them were looked at.
Descriptive context and an in-sample summary. The composite does not forecast returns (Test A, refuted), so nothing here tells you when to buy. The forward test began 2026-09 and needs roughly 20 independent quarters.
Two sessions before Canadian Thanksgiving (TSX closed Monday, October 12). The question is whether rising yields keep pressing the average stock while the largest names hold up.
An elevated risk reading has come before deeper drawdowns, not lower returns: bumpier, not worse. With the Fed minutes leaning toward another hike, CPI on October 14 matters more than usual; Q3 earnings arrive against a high bar.
The long-term composite is Neutral / Selective (58.9) and drifting down. Long yields near a generation high, a Fed expecting one more hike, a slowing US job market, a firmer US dollar and narrow leadership are the backdrop.
Each star is a scored call. It stays commentary until 50 are scored and it beats the no-news baseline.
RSP and small caps beating the S&P 500 for a few weeks, with breadth back above 50, would ease my concern about the narrow market. High-yield spreads back above 3.24%, the VIX above 20, or a hot CPI would say the stress is building.
Stocks fell worldwide as yields rose and Fed minutes pointed to another hike; the S&P 500 slipped 0.2% from its record, but the TSX fell 1.7% and XEQT 0.6%.
An elevated risk reading has come before deeper drawdowns, not lower returns: bumpier, not worse. With the Fed minutes leaning toward another hike, CPI on October 14 matters more than usual; Q3 earnings arrive against a high bar.
Written after each close by an AI reading the dashboard, headlines and its own track record. Every call is scored against real prices by code it cannot change. Commentary, not financial advice.
Forward-looking bias per timeframe, with the signals driving it
Experimental — failed historical validation (2016-2026 backtest): outlook labels showed no reliable forward edge, and headwind-labeled days were often followed by rebounds. Shown for transparency and ongoing live scorecarding, not as a trading signal. See EVALUATION.md.