Signals
Standalone setups we watch — each card is a specific market condition that has historically preceded above-baseline forward returns. These sit outside the main BATS composite; the BATS gauge is a broad, slow-moving mood ring, while these are single-purpose event triggers. New signals will drop in over time.
Bond Fear Spike (MOVE Index)
Treasury volatility · spike vs 50-day average · short-term bounce, daily checkThe MOVE index is the bond market's version of the VIX: it measures how big a swing traders expect in Treasury yields over the next month. When bond fear spikes, stocks are usually already falling — and the top of the spike has tended to land within days of the stock market's low. A spike here means MOVE closes at least 25% above its own 50-day average. Since 2003 that has happened — times. Measured from the day each spike peaked — where the chart below draws its markers — SPY was higher a month later — of the time, averaging — against — for a random month. The catch: you only know the peak a few days late. Buying the day after the rule fires, which you can actually do, still beat a random month: — and higher — of the time. Either way the edge is short-lived: by twelve months a spike looks like any other day.
| Holding period | From the peak | Higher | After it fires | Higher | Any day | Higher |
|---|---|---|---|---|---|---|
| — | ||||||
What to expect if you follow this signal
The peak is only known afterwards. The chart marks each spike where MOVE topped out, and the stock market's low has usually been within days of that point. But MOVE has to stop rising before you can call it a peak. The rule fires on the way up, a median — trading days before the top, and buying the day after it fires captured most of the move (— a month later vs — from the exact peak). Waiting for MOVE to turn down first did not help: by the time a 5–10% drop from the spike high confirmed a peak, the first bounce had usually already happened.
It is a bounce signal, not a bottom call. Even from the peak, the lowest close of the next three months came a median — trading days later, with SPY dropping another — on average on the way. Twice it kept falling much further — — — because those spikes came early in a longer decline, not at the end of a scare.
The edge fades with time. One-week and one-month returns after a spike beat a random week or month by a wide margin. Three months out the gap has narrowed (— vs —), six months out it is smaller still (— vs —), and at twelve months a spike looks like any other day (— vs —). Treat it as a reason to buy a dip, not to change a long-term plan.
The falling ceiling is a pattern, not a rule. Every spike since March 2023 has topped out lower than the one before, along a straight line through the 2023 and 2025 highs; the March 2026 spike stopped right at it. If a future spike closes clearly above that line, the calming trend is broken and the 2022 playbook — repeated spikes into a falling market — becomes the better guide.
When the signal fires — and how to actually trade it
MOVE is published once a day, after the bond market closes (around 4:30 PM ET). Yahoo posts the reading in the evening, so this card picks it up in the nightly refresh (~7:30 PM ET), plus a quick live check when the page loads. Because the number lands after the stock market has closed, the tradable version of the backtest buys SPY at the next day's close — you cannot act on a spike the day it prints, and you cannot know it has peaked until it turns down.
- Evening: if the pill says SPIKE ACTIVE and the trigger date is today, plan a SPY buy for tomorrow's close.
- Sizing: the historical edge is a few percent over a month. Size the trade so another — (the average further drop after a firing) or a crisis-sized — is survivable.
- Exit: the stats stop at one month and there is no exit rule in the data, so pick your holding period before you buy.
Study — – — (— yrs) on daily closes of the ICE BofA MOVE index and SPY, price only. Spike = MOVE close at least 25% above its 50-day simple average, one event per spike (it re-arms once MOVE falls back under +10%). "From the peak" returns start at SPY's close on the day MOVE topped out within that spike, which is only identifiable in hindsight; "after it fires" returns start at the close after the trigger day and are the tradable version. The same pattern holds at 20%, 30% and 35% thresholds; 25% is shown because it is a round number, not because it tested best. The ceiling line is drawn through the highest closes of the 2023 and 2025 spikes in this data and is a chart observation, not part of the rule. Results ignore trading costs, taxes and dividends. Not investment advice.
Breadth Ratio (50MA / 200MA)
Loading… S&P 500 · daily% of large-caps above their 50-day MA divided by % above their 200-day. A deep dip (≤0.35) that turns up marks a short-term breadth washout followed by an early bounce — historically a good spot to buy. Vertical bars mark past fires.
Backtest 2005–2026 on ~100 S&P constituents. Fires when 20-day min of the ratio hit ≤0.35 and today's ratio ticks up. Baseline 250d return: +10.2% / 80% hit.
Breadth Washout (pct50 ≤ 15%)
Loading… S&P 500 · dailyWhen the share of large-caps trading above their 50-day MA collapses to ≤15% and then ticks back up, that's a washout bottom being formed. Catches the shock-crash setups the Breadth Ratio misses (Apr 2025 Liberation Day, Mar 2020 COVID, Dec 2018 Powell) plus a wider set of shallower dip-buys. Threshold widened from 10 to 15 to catch nearly 2× the entries with better short-term timing — fires slightly earlier so the market has usually bottomed by the time you'd act, instead of overshooting down for a week.
Backtest 2005–2026 on ~100 S&P constituents. Fires when the 20-day minimum of pct50 hit ≤15 and today's pct50 ticks up. Baseline 250d return: +10.2% / 81% hit. Threshold widened from the earlier ≤10 setting after testing showed 15 catches nearly 2× the entries (49 vs 29 over the same window) with better short-term timing — the 5-day post-fire return went from −0.6% (catching a knife) to +0.2% (bottom mostly in). The 60d+ window is still where the edge shows up.