Nine sectors · up to five in each · every week
Stock tips written by a research engine, read in four minutes.
Every week our model scores a broad universe of U.S. stocks and mails you the standouts with entry and exit levels — plus a separate weekly scorecard of how the last calls did.
Risk-adjusted, in plain English: the amber line swung 25% a year against the S&P 500's 12% (annualized volatility). Sharpe is return per unit of that swing (computed gross of a cash rate); Sortino counts only downward swings; max drawdown is the worst peak-to-trough dip since the start. All marked daily.
How to read this: we ran the model back from Jul 2025 to Oct 2, 2026 and pretended to invest along with it — every 4 weeks, put an equal amount into each stock it ranked top of its sector that week, hold for 4 weeks, repeat. The amber line is what that portfolio would have grown to; the dotted lines are simply owning an index fund instead — the S&P 500, the Nasdaq 100 (QQQ) or the Dow (DIA). The start date is fixed and the chart extends every Saturday; it is never restarted from a better week.
The live line is not a simulation: from Jul 31, 2026 it applies the same every-4-weeks rule to the picks we actually published (longs and shorts, as sent), marked at daily closes — −1.4% versus +3.3% for the S&P 500 over the same weeks. It is a short record so far; judge it in quarters, not weeks.
The honest caveats: the amber line is a simulation, not money we managed — no trading costs or taxes, and it scores today's stock list, which quietly favors survivors (companies that failed along the way never make such a list). Past results, real or simulated, don't predict future ones.
Top 5 High Conviction · track record
The five highest-scoring stocks each week, replayed and live
How the list is made. Every weekend the model scores a universe of about 500 U.S. stocks from 0 to 100 on fundamentals, trend, technicals and volume. The Top 5 is simply the five highest scores across all sectors. Short calls, contrarian calls and names carrying a fresh news red flag are left out, and an AI review can veto a name but never add one or change the order. It goes out Monday morning, before the open.
How the numbers are counted. Put an equal amount in each of the five names, hold for a week, and at the next list sell whatever dropped off and re-split the whole portfolio equally across the new five. The live line does exactly that with the lists we actually sent: each name bought at the Monday open after the email and sold at the next Monday open (a Monday holiday uses Tuesday's open). The amber backtest line applies the same rule to past weeks, Friday close to Friday close.
The held-1-year line is the long-term version: each week's five are bought and kept for 52 weeks, then sold, so every sale is a long-term one. The portfolio is split equally across all the weekly lists still held (one list in the first week, up to 52 after a year), backtested Friday close to Friday close. Holding for a year lets one big winner dominate: a stock that stays on the list for months ends up in many of the held lists at once, so this line swings with a handful of names even more than the weekly one does.
| $10,000 in… | Worth now | Return | Profit / loss |
|---|---|---|---|
| Top 5 live, from Aug 4, 2026 | $11,643 | +16.4% | +$1,643 |
| S&P 500, same weeks | $10,141 | +1.4% | +$141 |
| Top 5 backtest, from Jul 2025 | $14,667 | +46.7% | +$4,667 |
| Top 5 held 1 year, from Jul 2025 | $28,571 | +185.7% | +$18,571 |
| S&P 500, same period | $12,503 | +25.0% | +$2,503 |
A worked example. Start with $10,000: $2,000 goes into each of the first week's five names at the Aug 4, 2026 open. Each Monday the portfolio is sold down and split equally again across that week's five, so a name that stays on the list is topped up or trimmed back to a fifth. The table is what that $10,000 would be worth today, before trading costs, spreads and taxes. Rotating up to five names a week costs real money in spreads, so a real account would have made less.
Why the weekly email shows a different number. The email's phantom portfolio started from the prior Friday's close in its first week (a price no reader could have traded) and never resizes a name that stays on the list. This page uses only prices you could actually have bought and sold at, so it reads lower.
The honest caveats: five stocks swing far harder than an index. Look at the max drawdown and volatility above before anything else. The backtest rebuilds past scores with fewer inputs than the live model, can't replay the news filter or the AI veto, and uses today's stock list, which quietly favors survivors. The live record is only weeks long, so judge it in quarters, not weeks. Research only, not investment advice, and past results don't predict future ones.
Two narrower reports · access arranged by email
Two things the free email doesn't carry
The weekly sector tips stay free, and they stay the same. The engine also produces a weekly Top 5 High Conviction list and a daily Index Levels read on the index complex. Both are handled by email, one reader at a time.
See a real sample of both → Genuine past output, redacted only where the withheld part is the whole point of the report.
How access is turned on. Write to signal@shapethezone.ai saying which one you're interested in. Access is switched on by hand against your email address, so write from the address you want the reports sent to.
Research only — these are the same model's opinions, kept separate because they're narrower work, not because they're guaranteed. Nothing here is investment advice, and past results don't predict future ones.
Two emails a week — the calls, then the receipts
By the numbers: 10Y 4.61% (+0.06) · WTI $84.25 (+4.22) · VIX 18.21 (−0.56) · DXY 120.71 (+0.40)
- US launches 'heavy' strikes on Iran after attempted attack on American troops
- Samsung says chip crunch will last until 2028 as quarterly profit soars
~1σ weekly band (≈68% probability). "in-band" = share of the last ~6 months our band held, backtested.
+ 8 more sectors · Unsubscribe
| Pick | Bought at | Now | Your P&L | Sector | vs sector | Since we called it |
|---|---|---|---|---|---|---|
| AAPL | $334.54 | $338.69 | +1.2% | −3.8% | +5.0% | |
| CSCO | $115.16 | $115.22 | +0.1% | −3.8% | +3.9% | |
| SNOW | $268.02 | $269.90 | +0.7% | −3.8% | +4.5% | |
| NET | $262.14 | $263.71 | +0.6% | −3.8% | +4.4% | |
| DDOG short | $246.83 | $251.01 | −1.7% | −3.8% | −5.5% |
| Pick | Bought at | Now | Your P&L | Sector | vs sector | Since we called it |
|---|---|---|---|---|---|---|
| FANG | $148.20 | $164.50 | +11.0% | +1.4% | +9.6% | |
| SLB | $44.18 | $46.02 | +4.2% | +1.4% | +2.8% | |
| OXY | $51.06 | $51.68 | +1.2% | +1.4% | −0.2% | |
| KLXE short | $8.42 | $8.66 | −2.9% | +1.4% | −1.5% |
Priced at the first open after the email — the price you could actually have paid. Every pick counted, none dropped.
The model, in plain english
What's behind a pick — and what an STZ score means
Every week the engine scores a broad universe of U.S. stocks across nine sectors and mails the standouts in each. No black box theatrics and no chatbot guesses: each pick carries a one-line, auditable "why" built from the signals that actually drove its score.
Fundamentals
How the business is actually doing — earnings delivery, growth and what the street expects next. The quarterly-report half of the product.
Trend
Whether the market is already voting with the thesis — the shape of the price path over the past year, not a single day's move.
Timing
Technical texture that separates "good company" from "good entry this week", drawn from daily price action.
Participation
Whether money is actually moving into the name right now, judged against the stock's own recent baseline.
How the research is produced → — the full pipeline, the operating principles, and exactly what stays proprietary.
The STZ score, pick methodology and research engine are proprietary. © 2026 ShapeTheZone LLC. All rights reserved.
The weekly email is free, and stays free. The model behind it took months to build, runs every week, and gets checked against what actually happened — a contribution covers the server time and model calls behind an issue, and the hours that go into keeping it honest. That's what keeps the research independent, with no paywall, no ads and nothing sold about you. Give whatever it's worth to you. Entirely your call.