α
AlphaSteve
← Brain index

AM / PM Bidaily Note — codified template

This file codifies the AM/PM bidaily template that was operating implicitly until the 2026-05-28 SNOW audit (see audit-log entry #001). The implicit template let earnings claims propagate without primary-source verification, let price-action commentary expand at the expense of fundamentals, and tolerated internal-vocabulary loops that obscured analytical errors. The codified template below adds three discipline rules that address each failure mode directly.

The AM/PM template freeze (_FROZEN-2026-05-28) lifts only after the user has reviewed and signed off on this codification.

Cadence

  • AM note: written 6:30–7:30 AM ET. Reads overnight tape (Asia, Europe), pre-market US, and any breaking news. Reconciles against _house-view positions. Sets discipline for the day's pre-deployment posture.
  • PM note: written 5:30–6:30 PM ET. Reads the US cash session, post-close earnings, after-hours reactions, and any policy/macro updates. Reconciles against _house-view positions and the AM-note framing. Sets up overnight discipline.

Structure

---
tags:
  - research
  - daily-research
date: YYYY-MM-DD
session: AM | PM
---

# Research — YYYY-MM-DD AM|PM

## Top of mind
[2-4 paragraphs. The session's load-bearing observations. Lead with the *fundamental* or *business* shift the session named, not the tape reaction. Tape commentary supports, not leads.]

## Tape close (PM only) | Overnight tape (AM only)
[Index levels, key sector rotation, VIX, yields, oil, FX. Source-tier-tagged. No more than ~30% of total word count.]

## Business & corporates
[Name-level: earnings prints, capital allocation, strategic moves, regulatory actions. Each name as a bullet with:
- Lead with the business/operational read, not the price reaction
- Primary-source citation for any earnings claim (8-K, press release, or transcript) — see Discipline #1 below
- Price reaction at end of bullet, not start
- Carries / extends / new evidence tag]

## Geopolitics & macro
[Policy, central banks, data releases, geopolitical developments. Source-tier-tagged.]

## Technology & sectors
[Sector-level themes, supply-chain developments, technology shifts. Source-tier-tagged.]

## Day ahead (AM) | Themes emerging (PM)
[Forward-looking. What's the discriminating event in the next session? What pattern is forming across sessions?]

## Implications for AlphaSteve
[Pre-deployment posture. Watchlist proximity. Thesis-pipeline implications. Sector views. Discount-rate posture.]

## House view reconciliation
[Mandatory. For each relevant house-view position: extends / conflicts / no change. Cite the specific edit landing in [_house-view](/dailies#house-view).]

## House view changes this run
[Mandatory end section. List every change to [_house-view](/dailies#house-view) this run produced. "No changes" is valid; say so explicitly.]

## Cross-references
[Linked notes, prior sessions, thesis files, frameworks.]

Three discipline rules

Discipline #1 — Primary-source pre-flight for every earnings claim

Rule. Every earnings claim cited in an AM/PM note must reference a Tier 1 source (8-K, company press release, or earnings transcript). A Tier 3 aggregator (CNBC, Yahoo Finance, Stocktwits, etc.) can be cited as well, but not instead of. The pre-flight is a checklist box that must be ticked before the claim is written, not assumed.

Why. The 2026-05-27 PM SNOW print recorded "muted AH reaction" from aggregator color without anyone opening the 8-K or press release, where the EPS beat magnitude, the guide-rate acceleration, and the $6B AWS deal all sat on page one. The pre-flight discipline catches this kind of sourcing failure structurally.

How to apply. Before any bullet that includes an earnings number, the writer ticks a mental checklist:

  • Have I opened the 8-K?
  • Have I read the headline EPS, revenue, and guide?
  • Have I checked for same-day strategic announcements (partnerships, M&A, buybacks)?
  • Have I verified the after-hours reaction against at least one aggregator quote?
  • If any box is unticked, the bullet does not get written until it is.

Failure mode this prevents. Repeating an aggregator summary that misses the load-bearing element of the print.

Discipline #2 — Price-action cap

Rule. No more than ~30% of total AM/PM word count on tape commentary (index moves, multiples, AH reactions, gap-ups, VIX moves, oil moves, yield moves). The remainder on fundamentals, business operations, revenue mix, guide composition, capital allocation, end-market signals, customer-demand evidence, competitive dynamics.

Why. The user feedback on 2026-05-28: "AM and PM too much about price action, not enough about revenue, business, etc." The implicit template was letting tape commentary expand because tape data is easier to write than business analysis. The cap forces the analyst's attention back to the fundamentals.

How to apply. At the end of each AM/PM, the writer estimates the word-count split between (a) tape/price/multiple discussion and (b) business/fundamentals discussion. If (a) is materially more than 30%, the note gets re-balanced before filing.

Failure mode this prevents. A note that reads as "what the tape did today" instead of "what changed in the underlying businesses today."

Discipline #3 — Plain-English constraint

Rule. No internal-naming chain longer than two references in a single sentence. Every sentence should survive removal of the internal-vocabulary scaffold. If a sentence reads naturally only because the reader knows what "the symmetric-operation refinement of the cycle-late-selectivity theme" means, the sentence has failed the test.

Why. The user feedback on 2026-05-28 quoted this sentence as the failure example: "The compressed-setup-doesn't-pop framing the PM-27 named is now resolved with a modest negative rather than flat reaction; the symmetric-operation refinement of the cycle-late-selectivity theme is confirmed with one additional notch of conviction." That sentence does not parse for any external reader. It also masks the underlying analytical mistake (the SNOW mis-read) by wrapping it in jargon that sounds analytical.

How to apply. Each sentence the writer drafts gets tested against this question: would a portfolio manager outside the kit, reading this for the first time, follow the argument? If the answer requires them to look up two or more internal terms, the sentence gets rewritten.

A useful test: replace each internal term with the plain-English thing it points at, and see if the sentence still works. "The compressed-setup-doesn't-pop framing the PM-27 named is now resolved with a modest negative…" becomes "Yesterday I claimed compressed names don't pop on confirmation prints; today's CRM −2% close-of-AH is a modest negative version of that." That second sentence parses. The first does not.

Failure mode this prevents. Private-vocabulary loops that mask analytical errors and create the illusion of rigor.

Linked