Pre-Earnings Confidence Radar — Read Between the Lines of Company Leadership

Four times a year, company leadership sits down for an hour-long call with analysts to discuss performance and future plans — the earnings call. Most investors focus only on the headline numbers, but the real gold lies in how the leadership speaks, not just what they say. The Transcripts / Pre-Earnings Confidence Radar (/transcripts) tab extracts signals from call transcripts using AI, saving you from manually reading through hundreds of pages.

---

What is an earnings call and why is tone a signal

An earnings call is a quarterly call where the CEO and CFO provide commentary and then answer live questions from analysts without a rehearsed script. The prepared presentation is polished PR. But when an analyst presses on a weakening margin, the way the response is delivered reveals true confidence:

• Evasive, vague answers (“we’ll get back to that,” “I don’t want to predict”) = caution, possibly masking.
• Confident, specific outlook with numbers = leadership knows where they’re headed.
• Nervous hedging around a topic = a weakness the market hasn’t priced in yet.
• Change in tone between quarters — from enthusiasm to “calibrating expectations” = warning.

★ Insight: The tone of leadership is a “soft” signal that’s hard to fake over the long term. Numbers can be polished for a one-time shine, but consistent confidence (or lack thereof) in live responses is costly to pretend.

---

What AI extracts from the transcript

• executive_confidence (0–100) — how confidently leadership speaks about the future.
• guidance_clarity (0–100) — how specific the outlook is. Note: clarity ≠ optimism (a company can clearly communicate a bad outlook).
• ai_sentiment (−1 to +1) — overall emotional tone.
• ceo_truthfulness — estimate of consistency (not a lie detector).
• risk_mentions / opportunity_mentions (counts) — shifts in the ratio between quarters are more interesting than absolute numbers.
• key_topics — main topics.

⚠️ WARNING about scales — don’t mix them. executive_confidence and guidance_clarity are 0–100. ai_sentiment is −1 to +1. Confidence 78 and sentiment 0.4 are in two different worlds. Mixing them (e.g., ×100 on sentiment) results in nonsense.

---

Four Baskets

1. PEAD candidates — strong beat + high confidence + clear outlook.
2. Quality compounders — long-term high stable confidence across quarters.
3. Caution / Trim — drop in confidence, evasive outlook, increase in risks → yellow flag, investigate (not an automatic sell).
4. Recent beats — who just delivered.

★ Insight: Baskets are not recommendations — they are attention filters. The PEAD basket doesn’t say “buy,” it says “look at the fundamentals and catalyst calendar.” It narrows 500 companies down to 15.

---

What is PEAD and why the radar is based on it

PEAD = Post-Earnings Announcement Drift — after a strong surprise, the price “drifts” in that direction for weeks. The radar expands on this: a PEAD candidate with high confidence + clear outlook has a higher chance that the drift continues (confident management signals that the beat wasn’t a one-time fluke).

---

Radar NEVER solo — connect to fundamentals

Radar is context, not a standalone signal. Tone tells you how management perceives the situation. It doesn’t tell you if the company is cheap or generating cash. A confident CEO of an overpriced company burning cash is still the CEO of an overpriced company burning cash.

It serves best as the final layer of context over fundamentals: the company passes the filter (5 pillars) → radar tells if the call supported or questioned the thesis.

★ Insight: Soft signals are most valuable as a tiebreaker (two similar companies → tone decides) and as an early warning (drop in confidence + rise in risks = warning before it shows in the numbers).

---

Connection: Transcripts → ATS → Catalyst Calendar

Data feeds the ATS pillar (part of SENTIMENT) on the stock detail. The Catalyst Calendar shows upcoming earnings → together they form a loop (calendar = when reporting, radar = how it sounded).

---

Practical Steps

1. Notice an interesting ATS score.
2. /transcripts — confidence, clarity, sentiment, basket.
3. Check fundamentals (5 pillars) — radar is context.
4. Catalyst Calendar — when the next report is.
5. Decision on confluence, not on one number.

---

Tips

1. Watch the change, not the absolute value (confidence 70 after previous 85 is worse than a stable 65).
2. The ratio of risk/opportunity mentions is a sensitive barometer.
3. Caution/Trim = “investigate,” not “sell.”
4. Combine clarity and sentiment (a clear outlook can be clearly bad).
5. For PEAD candidates, verify the quality of the beat (one-time tax relief ≠ revenue growth).

---

Common Mistakes

Overvaluing a single call · ignoring fundamentals · mixing scales (0–100 vs −1..+1) · taking AI analysis as absolute truth · entering blindly before a report.

---

Summary

The Pre-Earnings Confidence Radar makes visible what analysts have been manually reading for years: how leadership truly perceives their future — confidence, clarity of outlook, sentiment in live responses. Its strength isn’t in telling you what to buy — it’s in adding a layer of context to fundamentals, feeding the ATS pillar, complementing the Catalyst Calendar, and serving as a tiebreaker and early warning. Solo is a trap; as the final layer of confluence, it’s a tool for institutions.

---

Analytical and educational description of the QMA tool, not an investment recommendation. AI analysis of transcripts may contain errors — it serves as one of the inputs, not a replacement for your own decision-making. Verify against fundamentals; consider consulting a licensed advisor.