Operator's Manual

What it is

The Hedge Ideas cards scan every sector for a strong leader and a weak laggard scored far apart — and require both legs to be early in their moves. A leader that's already stretched, or a laggard that's already washed out, disqualifies the pair no matter how wide the gap. Δ is the score gap — bigger gap, stronger idea.

Each card shows the sector, both legs' Alpha scores, each leg's extension state (the freshness that earned it the slot), and the equal-risk sizing. Click a card and the whole thesis loads: leader on the chart, laggard in compare, spread lane underneath.

Fewer cards on some days is deliberate. Late in a stretched rally most leadership is already extended, so little qualifies — the count itself tells you how much fresh opportunity the tape is offering.

Why pairs

Long the leader, short the laggard — the market's moves hit both legs and cancel out. What's left is the gap between two businesses. Crash or melt-up, you're only exposed to the spread.

In testing, wider score gaps have preceded larger forward spreads — and unlike single-name signals, the edge persisted over multi-week horizons, because market drift is out of the equation.

Why freshness matters: a stretched leader has already spent most of its move — what remains is mostly snap-back risk. A washed-out laggard is primed to squeeze. In testing, pairs with both legs early in their moves carried by far the strongest forward spread, while pairs built from two exhausted moves showed none. The cards only offer the former.

Reading the spread lane

Bars = the spread's journey — the leader's relative line minus the laggard's — measured from the chart's six-month start, at every zoom: a day's value never changes when you switch 1M/3M/6M, only which days you see. The axis fits the visible range, and the zero line appears only when the window actually crosses it — no half-empty lane, no level pretending to matter while it's off-screen. Posture keeps the sign readable even then: a window entirely above its anchor stands its bars on the floor; one entirely underwater hangs them from the ceiling. Color carries the trend — grey when the spread is quiet, green through a Widening run, red through a Narrowing run, amber on transition moments: brief undecided handoffs between opposite runs (longer pauses stay grey — that's real quiet, not a handoff), and the third straight Narrowing day, where the tested exit fires (the label spells it out: Narrowing 3d — tested exit). Each colored stretch opens with its matching marker — ▲ Widening, ▼ Narrowing, ◆ transition — so the read survives without the color, and single-day flickers stay grey: a run needs two consecutive days to earn its color, with one-day gaps bridged, the same debounce the chart's ribbons use.

The number in parentheses is the current spread — it re-reads as you hover, so you can measure any stretch of the trade.

When comparing, both legs' earnings sit at the lane's floor — an amber E for the focus leg, a purple E for the compare leg — because the pair owns its whole calendar risk. While a Narrowing run is alive, the label counts it day by day. Track the pair and the lane gains two references. A dotted line at your entry level — bars above it mean the trade is paying — with the readout switching to percentage points since entry. And your holding itself as a band whose color is the verdict — green while the spread is paying, red while it's losing, amber while it's sideways within a point or two — gentle from your entry day, stepping a shade brighter through the tested payoff window, whose edges wear chevrons — > opens it at day 10, < closes it at day 21, each appearing once its day reaches the chart. While you're inside the window the lane header counts it — tested window — day 14 of 21. Outsized one-day breaks against the spread's trend don't mark the chart — they flag in your tracker, where there's room to explain them.

The spread's own trend — the lane also reads Widening (the spread is still opening), Flattening (momentum gone), or Narrowing (closing against the position), judged from its recent path against its own noise — no anchor, no window dependence. In our testing, pairs entered while the spread was still widening went on to clearly outperform those already flattening, and pairs entered against a narrowing spread lost on average. A tested expectation, not a schedule — any single pair can break the pattern. The same word appears on each Hedge Ideas card.

On the right of the lane: the hedge readout — the pair's net β (leftover market exposure at the shown sizing, colored green near zero), the dollar sizing, and the verdict. The same nβ chip rides the board cards, the ticket shows it as a Market lean scale — a dot between short-bias and long-bias with the verdict in words — and it freezes into the tracker at entry, so you can always see what market lean you actually put on. The tracker keeps watching after that: betas drift, and if a pair's net exposure at your frozen sizing has moved well away from what you entered with — and now carries a real lean — the row raises a hedge drift flag with both numbers. Drift toward neutral never warns; that's the market doing you a favor. Any pair served on the Hedge ideas strip reads ✓ hedgeable — on Hedge Ideas with the strip's own sizing, so the chart and the board always agree. Ad-hoc pairs you build yourself get graded on the spot.

How to trade it

Enter when the gap is wide and the spread lane is still trending up — you want the divergence and its momentum.

Exit or trim when the lane flattens or the laggard's score starts recovering — gap closure is the trade ending, not a dip to add into.

Give it two weeks. In our testing, the first days of a new pair did little — the spread's gains concentrated in weeks two and three, and by about a month the typical spread had stopped growing and began to give back. Treat that as a tested expectation, not a schedule: individual pairs pay early, late, or not at all. The spread lane, not the calendar, has the final word — a spread still widening is a trade still working, whatever day it is.

The exit that tested best: three straight Narrowing days, or about four weeks — whichever comes first. One reframe worth internalizing: this rule fires on winners too, and that's its best case — the tracker says so in words (exit with the win when you're ahead, plain tested exit when you're not). Red means "act," never automatically "bad." We made the exits compete on the same replayed trades. Cutting a pair after three consecutive Narrowing days kept the same overall return as holding a flat three weeks, in roughly a quarter less time, with a meaningfully smaller worst case. It works by asymmetry — many small early exits from pairs that die, while the ones still widening run their course. Three other results are worth as much as the rule itself: exiting the moment a pair drops off the Hedge Ideas board was far too early (conditions flicker daily — treat that as a review prompt, not an exit); price stop-losses tested worse than no rule at all, because a spread breathes with noise and a stop just locks the breathing in; and holding past four weeks gave most of the gains back. As always: a tested expectation across many trades, not a promise about any one.

Sizing — one rule everywhere: 1:0.82 = $1 long for every $0.82 short, scaled so both legs swing the same dollars on a typical day. The quieter stock takes the bigger side; the P&L then tracks the spread, not whichever leg is more volatile. Equal dollars is the acceptable lazy version.

How pairs enter and leave the board

Ideas enter hard and stay patient. A pair joins Hedge Ideas only on strict conditions — a genuine leader against a genuine laggard, both legs fresh. But once seated, it isn't evicted by a single red day: it stays while the thesis holds, riding out small wobbles in either leg. Each card shows its tenure (day 4), and a pair currently inside that patience band wears a holding tag with a dashed border — still a live idea, just not a fresh entry.

This is tested, not taste: in our replays, pairs retained through a wobble went on to outperform fresh entries over the following weeks — a red Tuesday that nudges a leader's score is noise, not a broken thesis. The flip side is that leaving the board now means something: when a pair departs, its thesis genuinely died (laggard truly recovered, leader truly broke, or a leg went stale). Even then, our testing says treat the departure as a review prompt — the exit that tested best remains the three-day Narrowing run on the spread itself.

The bell names the reason each time a pair departs: laggard recovering — the short leg is no longer weak, so the thing you're short stopped falling behind; leader faded — the long leg lost its leadership; short leg washed out — the laggard has fallen so far it's in bounce territory, where the short side stops paying; leader stretched — the long leg is deep in chase-risk extension; gap narrowed — neither leg broke on its own, but the distance between them closed below the board's bar; and, rarely, laggard left the universe. On big regime days several pairs often leave at once — those collapse into a single bell event headlined by the most common reason, with each departing pair listed and clickable.

Timing matters too. When a sector has a violent day, the effect concentrates in the higher-beta names — after a violent surge they tend to give part of it back over roughly three weeks, and after a violent washout they tend to rebound over roughly two. So the flags name the name: a card might read ⚡ COIN headwind ~3wk — sector surged, meaning COIN specifically carries the tested drag — whether that works for or against the pair depends on which leg COIN is. The spread lane on the chart carries the same note beside its trend state, a pair tracked on such a day keeps the flag in your tracker, and the tracker tells you when a washout entry's window has passed. All of it is information, not a gate — the board's entry rules are unchanged, and the exit discipline stays the three-day Narrowing run.

Split sectors vs compressed sectors

A pair trade needs raw material: real winners and real losers. Sometimes a sector provides both — scores stretched wide, a crowd near the top and a crowd near the bottom, a sector genuinely at war with itself. Other times the whole sector huddles around neutral, and even the "leader" and "laggard" are barely distinguishable from the middle.

Why it matters: in our testing so far, the pairs that struggled most shared a profile — a gap that barely qualified, drawn from a compressed sector. The same modest gap inside a split sector held up better. Read it as a quality check on the whole neighborhood: a wide gap where winners and losers are everywhere is the market disagreeing loudly about business results; a narrow gap in a quiet sector may just be noise that happened to clear a bar.

This is early evidence — one stretch of market history — so the cards don't gate on it yet. Use it as context: when a sector looks compressed, ask more of the gap before taking the pair.

You don't have to eyeball this yourself: the Market tab's sector cards carry it as the split and quiet tags — split is the at-war sector that feeds the pair board, quiet is the compressed one that asks more of the gap.

The Relative ticket — a name's own basket

Every stock's panel can flip from Directional to Relative: instead of "buy it or don't," the ticket expresses the name against its own sector — long a strong name versus one or two genuinely weaker same-sector peers (or the sector fund when no peer qualifies), shorts mirrored the same way.

The hedge legs follow the same discipline as Hedge Ideas: a washed-out name never serves as a short leg — that's the bounce risk — and an already-stretched name never serves as a long hedge. Among what qualifies, the ticket prefers peers that actually move with the anchor, so the basket behaves like one spread instead of two unrelated bets, and it spreads hedge duty across the sector rather than pitting everyone against the same floor names.

Sizing is the one rule used everywhere: each leg scaled so both sides swing similar dollars on a typical day — the P&L then tracks the relative move, not whichever leg is noisier. The ticket shows this as a dollar template — the main leg anchored at $1,000, every other leg scaled off it; use any multiple, and convert to shares by dividing by price.

The hedge is insurance — in the strict sense. We replayed hedged expressions against the plain directional long across twenty years of regimes. In rising tapes the hedge costs: a strong name's sector and peers are strong too, and shorting them subtracts real drift (this year, hedged expressions clearly underperformed the plain long). In flat tapes it costs a little. But when the market breaks, the ranking flips hard — in the 2020 crash the hedged forms made money while unhedged leaders took their worst losses, and in the 2022 bear hedges roughly broke even while plain longs bled. The one constant everywhere: hedged worst-cases were far smaller. So reach for a Relative expression when you need protection — overnight gaps, a tape you distrust — and accept that in calm markets that protection has a premium. If you hedge, the two-leg peer basket was the steadiest form across every era; the sector-ETF hedge protects most in a falling market but costs most in a rising one.

Watch out