Spreads: Open vs Midday — Measure Once

[NYC88]

If your backtest assumes midday spreads at 09:31, your expectancy is fan fiction.

I killed a "3¢ edge" setup by measuring OPEN15 spreads once. Median was 6¢. Setup wasn't broken. The bucket was. Felt obvious after. Wasn't obvious before the sheet.

### One-time measurement drill (per watchlist name)

For each ticker, sample:

Code: Select all

time_et, bid, ask, spread_cents, last, notes
09:31
09:45
10:30
12:00
15:30

Repeat across **5 sessions**. Store median spread by bucket.

### Worked mini-example

Code: Select all

Ticker XYZ median spread_cents
OPEN15: 6
AM:     3
LUNCH:  4
PM:     2

Setup edge estimate after costs: ~4
Conclusion: OPEN15 marketable entries are structurally disadvantaged → limits only or skip bucket

### Use in playbook

| Bucket | Max marketable chase | Preferred order type |
|--------|----------------------|----------------------|
| OPEN15 | smaller / none | limits more often |
| AM | normal | per setup |
| LUNCH | often skip | — |
| PM | normal | per setup |

### Strategy implication

Mean-reversion that needs 3¢ edge dies if median open spread is 6¢. Skip the bucket, change order type, or admit the setup is midday-only.

### Maintenance

Re-measure when ADV collapses, you change broker route, or earnings week liquidity shifts. Old spread tables are how quiet accounts bleed.

Post anonymized median tables. Helps the board more than another broker rant.

Have you measured OPEN15 vs midday spreads on your top 3 names?
What bucket do you refuse to take marketable orders in?
Anyone's "edge" disappear after an honest spread table?
How often do you refresh the table — monthly, or only after pain?

[Fairman]

On **Spreads: Open vs Midday — Measure Once**, the desk standard stays practical. Spreads at the open and spreads at midday are different instruments wearing the same ticker. I measure both. I do not size from memory of "this name is usually tight."

Replying in-thread with operable constraints — not forum combat. If these rails disagree with your framing, treat it as a checklist offer.

Execution quality is fill-to-plan adverse cents on your names at your TOD buckets — not a brand slogan. Size follows live spread width and session phase.

## Rails that stay on the sheet
- Measure once per regime; do not invent.
- Spread regime is part of the setup.
- Invalidation written in cents or an auditable structure break before entry.

## Measure-once protocol (actually: measure in buckets)
1. Pick your A-list names for a measurement week.
2. Log typical spread (and adverse cents on fills) in open bucket (09:30–10:00 ET) and midday bucket.
3. Hold route/broker constant while measuring.
4. Separate ETFs from single names — same share count is not the same tax.
5. Set live max spread per name/bucket. If live spread exceeds max, size cuts or observe — automatically.
6. Re-check after platform/broker changes; do not inherit last quarter's numbers forever.

## Worked example — Using midday spread memory at the open
Assumes 2-cent midday spread at 09:31; pays 7–9. Measure open vs midday once; set separate live max. Open and midday are different instruments.

I keep examples in cents and clock time so Sunday review can tag the behavior — not rewrite it as a mood.

## Failure modes
- Ignoring the rails above on **Spreads: Open vs Midday — Measure Once** and substituting confidence.
- Adding size after adverse selection because the story still feels right.
- Paying open microstructure tax that midday memory never modeled.

## When to skip
Skipping is a first-class decision — even when the chart still looks persuasive:
- Checklist incomplete at 09:20 ET — observe or one pre-written A+ only.
- Live spread outside max for that symbol and session bucket.
- Index violently disagrees with a single-name thesis and no reduced-size RS plan is on the sheet.
- Platform, hotkeys, locate, or flatten path unverified after reconnect/update.
- Daily loss, PDT slot, or prop buffer tripwire already touched.
- News/FOMC/CPI window where the size-cut card says stand down.

## Closing
For **Spreads: Open vs Midday — Measure Once**, keep operable boredom: written rails, measured cents, index context when relevant, and a skip that needs no apology. Process quality beats ticket volume.