Guidance
How to use The 4 Horsemen
This guide covers both the Alpha Debate (equity thesis) and the Dividend Safety Suite. It assumes you've already run at least one analysis and have a report in your library.
Contents
Alpha Debate
- Writing a good thesis prompt
- Reading the report
- The Conviction Gauge
- Outcome
- Using the Thesis Tracker
- Watchlist and Buy Zones
- Earnings Alerts
- Portfolio Stress Tester
- When to rerun
Dividend Safety Suite
1. Writing a good thesis prompt
The system works best when you give it a clear starting point. The more specific the input, the more targeted the output — but you don't need to write a thesis yourself. A ticker, a company name, a sector, or a theme is enough.
What makes a good prompt
The guided mode builds the full debate task automatically from your ticker or theme — you don't need to write the prompt yourself. Use the custom mode only if you have a specific angle you want the agents to start from.
The investment horizon (3–10 years) affects how the agents frame macro context and risk. For most long-term equity theses, 7 years is a reasonable default.
2. Reading the report
The report is structured in order of what matters most to a long-term investor. Read it top to bottom the first time; on subsequent reviews, focus on the sections most relevant to current market conditions.
Investment Thesis
The 2–3 sentence core argument. If you can't summarise this after reading, the thesis isn't strong enough to act on.
Outcome
Specific asset recommendations — a Structural Play (lower risk, higher conviction) and a Moonshot (asymmetric opportunity). Each includes entry guidance and a link back to your original queried thesis.
Quantitative Evidence
Price history, valuation multiples, momentum signals. This is the data layer — it either corroborates or complicates the fundamental case.
The Bear Case
Written by the Contrarian agent. This is the most important section. Read it first on any rerun — it tells you what would need to be true for the thesis to fail.
Conviction Level
Not a sentiment score. It's the residual conviction after the Contrarian's attack has been fully heard.
Monitoring Checklist
The specific data points to track quarterly. Each item is a thesis pillar — if it breaks, reassess.
The Debate Replay (linked from the report) shows the full phase-by-phase exchange — every argument the Contrarian made and how the other agents responded. Read this when you want to understand the reasoning behind the verdict, not just the verdict itself.
3. The Conviction Gauge
The conviction level — HIGH, MEDIUM-HIGH, MEDIUM, or LOW — reflects what was left after the Contrarian's full attack. It is not a prediction of returns, and it is not a buy signal.
The Contrarian struggled to find structural weaknesses. The bull case survived rigorous challenge. Entry at the right price is well-supported.
Strong thesis with one or two identifiable risks. The Contrarian landed some hits but couldn't break the core argument. Monitor the flagged risks closely.
Genuine analytical tension. The bear case is credible. This doesn't mean avoid — it means be more selective on entry timing and position size.
The Contrarian found serious structural problems. This isn't necessarily a sell signal on an existing position, but it warrants a thorough review before adding.
4. Outcome
Every report produces two specific asset recommendations rather than a general verdict on the queried ticker. This is by design.
▲ Structural Play
A higher-conviction, lower-volatility proxy for the thesis. This is typically a market leader, infrastructure layer, or established compounder within the theme. The entry guidance gives a price range derived from the analysis — not a hard target, but a zone where the risk/reward is considered attractive.
⚡ Moonshot Play
A higher-risk, asymmetric opportunity within the same thesis. Lower position sizing is appropriate. This is the trade where you need the thesis to be directionally right AND execution to go well — both conditions must hold. Position accordingly.
Crucially, each play includes explicit reasoning explaining how it connects to your original queried ticker or theme. This link matters: if you queried AAPL and the structural play is TSMC, the report explains why TSMC is the better proxy for the specific thesis the debate produced — not just a generic recommendation.
5. Using the Thesis Tracker
Every completed report automatically generates a Thesis Tracker at /trackers/[id]. This is the tool you'll return to quarterly — not the full report.
The tracker preserves the original thesis alongside the Contrarian's bear case warning (verbatim), the monitoring checklist, and a log of any conviction changes over time. Think of it as your accountability document for the position.
A simple quarterly review workflow:
- Open the tracker for each active position.
- Read the Bear Case Warning first. Has anything occurred that makes this more credible?
- Work through the Monitoring Checklist. Which pillars are intact, which are weakening?
- Check the Catalyst Calendar. Are the expected catalysts on track?
- Decide: rerun, hold, or exit. If the core thesis is intact, hold. If a thesis-killer event has occurred or is more likely, rerun or exit.
You can add update log entries directly in the tracker to record conviction changes and the reasons behind them. This creates a durable record of your thinking over time — useful both for accountability and for future reviews.
6. Watchlist and Buy Zones
Every stock that appears as a Structural or Moonshot Play is automatically added to your Watchlist at /watchlist, with a live price feed and a buy zone derived from the entry guidance in the report.
The buy zone is a price range — not a target price. It reflects the level at which the analysis considered the risk/reward to be attractive relative to the thesis. When a stock is above the zone, the analysis considers it to be pricing in the upside already. When it's below the zone, it may have sold off past your entry range.
Buy Zone Alerts (configurable in Account → Alert Preferences) send an email when a watchlisted stock enters its range. You can remove any ticker from the watchlist at any time — this doesn't affect the underlying report or tracker.
7. Earnings Alerts
Configurable in Account → Alert Preferences. Two options:
Earnings Alert (all tiers)
An email 48 hours before any watchlisted stock reports earnings. The alert prompts you to review your thesis before the numbers land — giving you time to decide whether your conviction still holds or whether you want to reduce risk ahead of the announcement.
Auto-Run Earnings Analysis (Pro and Alchemist)
Instead of a reminder, the system automatically runs a fresh four-agent debate on each watchlisted stock 48 hours before it reports. One credit is consumed per run. You receive the completed report by email — a current analysis rather than a nudge to do one yourself.
If you have both options enabled, Auto-Run takes precedence — you receive the report, not the reminder.
Earnings alerts are sent once per earnings event. If you've already received an alert or auto-run for a given date, it won't repeat.
8. Portfolio Stress Tester
Available to Pro and Alchemist subscribers at /stress-test. This tests your whole portfolio against four historical crash scenarios, rather than evaluating a single thesis.
Enter your current equity holdings — each as a ticker and a percentage weight, totalling 100%. The system calculates how your portfolio would have performed in each scenario, then generates an AI narrative identifying where concentration risk and correlation failure would have caused the most damage.
2008 Financial Crisis
Stress-tests financial sector exposure, leverage, and any positions correlated with credit markets.
2020 COVID Crash
Rapid liquidity-driven drawdown. Tests whether your holdings are correlated with market beta or have defensive characteristics.
2022 Rate Shock
Duration risk test. Long-duration growth equities repriced severely. Tests whether your portfolio is rate-sensitive.
Dot-com Bust (2000–2002)
Valuation-driven collapse. Relevant for any portfolio with high-multiple tech exposure.
9. When to rerun
A debate report reflects market conditions, available data, and the analyst debate at a specific point in time. Markets change — rerunning is how you keep your conviction current.
Good reasons to rerun:
- Approximately 100 days after the original report — the 100-Day Rerun Reminder alert covers this automatically.
- After the company reports earnings — especially if results significantly missed or beat expectations.
- After a major macro regime change: a rate cycle inflection, a significant geopolitical event, or a policy shift that directly affects the thesis.
- If the Contrarian's bear case from the original debate starts looking more credible based on current events.
- Before adding meaningfully to a position — especially if more than 60 days have passed since the last report.
Rerunning consumes one credit and produces a new report and tracker. The original report remains in your library — useful for comparing conviction levels and monitoring checklist status over time.
10. Running a Dividend Safety Analysis
The Dividend Safety Analysis is a separate four-phase debate run by four income-specialist agents — distinct from the Alpha Debate team. Launch it from /dividend/run or from the Dividend section in the nav.
Credit usage — important
Each Dividend Safety Analysis consumes one credit from your monthly allowance — the same pool used for Alpha Debates. A Free account includes 1 credit per month total, shared across both report types. If you run a dividend report, that is your free credit for the month. Check your remaining credits in the nav before running.
Enter a ticker and click Analyse. The debate runs for 8–10 minutes across four phases: Income Thesis → Data Validation → The Great Debate → Dividend Safety Report. You can watch it stream live or leave and return — the report will be in your library when complete.
Good starting points for a first dividend run: JNJ (Dividend King, 60+ consecutive years), KO (Aristocrat, strong brand moat), ABBV (high yield with near-term patent risk — good for testing the bear case), O (REIT monthly payer).
11. Reading a Dividend Safety Report
The dividend report is structured differently from the Alpha Report — it is focused entirely on income sustainability rather than price upside. The key sections:
Dividend Safety Score (1–10)
The headline verdict. Built from FCF payout coverage, interest coverage, and debt load — not from yield or consecutive growth years alone. A stock with a 10-year Aristocrat streak and a 95% FCF payout ratio will score lower than one with 5 years of growth and 45% coverage.
Payout Sustainability
LOW / MEDIUM / HIGH — reflects whether current FCF, earnings trajectory, and capital allocation commitments can sustain the dividend in a downside scenario. This is not a verdict on growth; a MEDIUM rating can still be a quality income holding.
Cut Risk Level
HIGH / MEDIUM / LOW — the Dividend Bear's verdict on how likely a cut is within the next 3 years, with the specific scenarios that would trigger it. Read this section every time before adding to a position.
Dividend Growth Analysis
1/3/5/10-year CAGR table, consecutive growth streak, and King/Aristocrat/Achiever/Challenger/Starter classification. The trajectory matters as much as the streak — a King with decelerating growth needs scrutiny.
Yield-on-Cost Projections
The table shows what your current-price yield becomes over 1, 3, 5, 10, and 20 years at base DGR and conservative DGR. This is the income you are actually buying, not the yield on today's price in 20 years.
The Bear Case Warning
The single most dangerous cut scenario, named specifically. This is not a generic risk — it is the scenario the Dividend Bear identified as most credible given the current balance sheet, macro environment, and management history.
The Debate Replay for a dividend report shows all six income agents — Income Strategist, Balance Sheet Analyst, Dividend Quant, the Dividend Bear, the Balance Sheet defence, and the Quant referee verdict. It is the full reasoning trail behind the Safety Score.
12. Dividend Watchlist and Calculators
Every completed Dividend Safety Report is automatically added to your Dividend Watchlist at /dividend/watchlist. This is separate from the equity Watchlist — it shows income-specific columns: Safety Score, cut risk, current yield, 5-year DGR, and days to the next ex-dividend date.
Only the most recent report per ticker is shown — if you rerun JNJ, the watchlist updates to the latest analysis automatically.
The Dividend Calculators at /dividend/tools include two tools:
DRIP Calculator
Models dividend reinvestment over up to 40 years. Adjust shares, price, dividend per share, DGR, and toggle reinvestment on/off. Shows portfolio income and share count at key years.
Monthly Income Projector
Enter multiple holdings (ticker, shares, annual dividend per share, DGR). The tool projects total portfolio income at base and conservative DGR scenarios across 1, 3, 5, 10, and 20-year horizons.
13. Dividend Alerts
Three alert types are available for dividend holdings, configurable in Account → Alert Preferences under "Dividend Alerts". All three are opt-in by default.
Ex-Dividend Reminder (7 days before)
Sent 7 days before each upcoming ex-dividend date for tickers in your Dividend Watchlist. You must hold shares before the ex-dividend date to receive the payment — this alert gives you time to act if you intended to add before the cutoff.
Cut Risk Warning
Sent once when a holding from one of your reports has a HIGH cut risk rating. This is based on the AI analysis — it does not pull live earnings data. Treat it as a prompt to rerun the analysis or review the holding, not as a real-time signal.
Dividend Increase Alert
Sent when live market data detects a dividend payment higher than the amount stored in your most recent report. A dividend increase is a management signal worth noticing — the alert prompts you to rerun the analysis with the new payout figure.
Not financial advice. Everything on this platform — reports, debate outputs, trackers, alerts, and stress test results — is for informational and research purposes only. It does not constitute investment advice or a recommendation to buy or sell any security. All investment decisions remain yours. Full disclaimer →