True strength — the relative strength chart
The line that ignores the market's mood.
What it is
Take the stock's move each day. Subtract what it should have moved, given its beta and what the market did. Add up the leftovers, day after day. That running total is the line — the stock's true strength: relative strength done properly, with beta and the market stripped out rather than a simple price ratio. (The textbook name is cumulative relative performance.)
You get it twice: once against the market (SPY), once against the stock's own sector.
The mental shift
Rising line ≠ rising price. Rising line = beating expectations.
The market carried it. The stock added nothing of its own.
This is the whole trick: strip out the market's motion, and what's left is the stock's own story.
The chart, element by element
Green line — vs market. Yellow line — vs sector. Dotted blue — the price itself, for context.
Alpha bars (right axis, 0–100) — the engine's graded verdict on all of it. 50 = no edge. The Alpha chip top-right colors itself: green above neutral, red below.
Status line (top-left) — the current phase in plain words: Accumulation or Distribution, how long the relative move has run and at what pace, plus the extension read (Fair, Extended Support, Very Extended…).
E under a bar — an earnings day. Expect the line to bend there; judge the slope after the gap, not through it.
The last bar tells the time. During the session it's live and moving. After the close, the solid bar is where the session settled; a faint bar with a blue stamp shows where extended-hours trading has taken it since. Pre-market, the blue stamp alone marks the early tape before today's bar exists.
Hover anywhere — every legend number re-reads for that date. Scrub back through a move to replay it.
The signal ribbons
▲ green run (under price) — confirmed relative strength: the score is strong and still building. Historically these stretches preceded further outperformance.
▼ red run (above price) — confirmed weakness: score low, relative performance still bleeding. Read it as avoid / underweight — in quality names weakness tends to mean-revert, so it's a warning, not a buy-and-hold short. When a flagged name is already deeply discounted, the ticket relabels it bounce risk — the snap-back is statistically the bigger danger there.
◆ yellow — the handoff: one regime has ended, the next isn't confirmed. Expect chop; size accordingly.
Runs, not one-offs. A signal must persist before it prints, and it keeps printing while conditions hold. No marks at all = no edge either way — that's information too.
Compare mode
Type any covered ticker into + compare — its own vs-market line drops in (solid purple), same beta-adjusted frame, plus its raw price as dotted purple (dotted always means price; the color says whose). To keep the frame readable, the vs-sector line steps aside while you compare — it returns the moment you clear the box. NVDA vs AMD, head-to-head, with the market's motion removed from both.
A histogram appears underneath: the pair spread — your stock minus the compare, in percentage points. That's the subject of the next entry.
The lane also grades the pairing. net β is the pair's leftover market exposure per $1 long at the shown sizing: each leg has a market beta (how much it moves when the market moves, weighted toward recent months), and the short leg's, scaled by the sizing, subtracts from the long's. Near zero (green) the pair is a pure relative bet — the tape cancels out. Away from zero (amber, then red) the "hedge" still quietly leans long or short the whole market, and a broad move drags your P&L regardless of who's winning the spread. The sizing ($1 : $0.82) equalizes daily movement, not market exposure — which is exactly why net β is worth a glance: an equal-risk pair can still carry a market lean. The number is directly usable: nβ × your long-side dollars ≈ the hidden market position you're carrying. At nβ −0.29 with $2,600 on the long side, you're effectively short about $750 of SPY on the side — a +1% market day costs ≈ $7.50, a −2% day pays ≈ $15, whoever's winning the spread. From there it's a choice, not a flaw. If the lean points the way you already think the tape is going, it's a small market opinion riding free on top of the spread — leave it on. If you want the spread pure, cancel it directly: add that many dollars of SPY in the offsetting direction (buy SPY when nβ is negative, short it when positive) and the lean nets to roughly zero while the spread sizing stays untouched. The third leg is optional either way — small leans are usually cheaper to accept than to trade away. Then the verdict: ✓ hedgeable — the names move together enough to trade as one spread (tight marks the closest). ✗ weak pair — not enough shared movement; you'd just hold two separate bets.
A red nβ is a quality caution, not a hedging to-do. In our replays, pairs entered with a large lean were the strongest cohort over the first two weeks — and the weakest by a month, with the heaviest tails of any group. A big lean usually means one leg's beta has collapsed or exploded: the pair is structurally lopsided, and its payoff is front-loaded. Cancelling the lean with a third leg didn't fix that (the lean is a tiny share of a pair's risk — removing it barely moved the book), so the tested read is simpler: on a red chip, prefer the front of the payoff window and let the Narrowing rule take you out early rather than late. Example: a red-chip pair that hasn't paid by day 10 has history leaning against it.
How to trade it
Buy strength that's building — the line rising, and still rising. Relative leadership persists: accumulation takes weeks, not hours.
Reduce when it rolls over. Don't wait for the price chart to agree — relative weakness is the early warning, not the aftermath.
Both lines rising — vs market and vs sector — is clean leadership. Rising vs sector only? That's a sector story wearing the stock's clothes.