DMA vs PFOF — Worked Example in Cents

[NYC88]

This is not a morality play. It's arithmetic for scalpers.

**PFOF-style retail routing** can look "commission free" while your **effective** cost sits in the fill. **DMA / direct-style** access (where available) can show commissions but different fill behavior. Exact mechanics vary — measure *your* path. I've argued labels. Labels don't pay. Cents do.

### Worked example (illustrative — not your numbers)

Assume you scalp 1,000 shares, round trip, twice in a morning.

**Path A — "zero commission," worse fills**

Code: Select all

Avg adverse slip vs mid: 1.5¢ / share / side
Round trip slip cost: 1.5¢ × 2 × 1000 = $30 per round trip
× 2 RTs = $60
Commission: $0
Effective: ~$60

**Path B — commission + tighter fills**

Code: Select all

Commission: $0.005 / share / side → $0.005 × 2 × 1000 = $10 per RT
× 2 RTs = $20
Avg adverse slip vs mid: 0.4¢ / side → 0.4¢ × 2 × 1000 = $8 per RT
× 2 = $16
Effective: ~$36

Path B wins **in this toy example**. Flip the slip numbers and Path A can win. That is the point: **your fill CSV decides**, not the marketing page.

### Sensitivity table (same 2 RTs × 1000 sh)

Code: Select all

slip/side | zero-comm path cost | 0.5¢/sh commission path*
0.5¢      | $20                 | $20 + $20 = $40
1.0¢      | $40                 | $40 + $20 = $60
1.5¢      | $60                 | $60 + $20 = $80
*commission path also has its own slip — replace with measured

Build the table with **measured** slip. Fake numbers above are teaching tools only.

### What to do

1. Log slips for 2 weeks on current broker.
2. If you have another route/account type, A/B at **micro size**.
3. Include locate fees / platform fees if relevant.
4. Stop arguing labels (`PFOF`/`DMA`) without cents attached.

Broker features change. Account types differ. This post teaches measurement, not a recommendation of a named firm.

What effective ¢/share are you seeing after two weeks of logging?
Anyone A/B'd two routes at micro size with a real sheet?
Do you include locate / platform fees in the effective cost, or forget them?
Where does your style sit — more sensitive to commission, or to slip?

[Fairman]

Arithmetic over morality. Right frame for scalpers.

PFOF-style retail routing can look "commission free" while the effective cost sits in the fill. DMA-style access (where available) can show commissions with different fill behavior. Exact mechanics vary by broker, name, and time of day. Measure your path.

I keep a simple weekly sheet: mid at send versus fill in cents, by ticker and time-of-day bucket (open versus midday). The open is where the tax usually hides on US equities. If a venue can't or won't give fill-quality detail, that absence is information.

Your worked-example shape is what I show newer traders: same 1,000-share round trips, different effective cents. Here's the illustrative version I use. These are made-up numbers for the arithmetic, not a claim about any broker.

Path A, zero commission, 1.5¢ adverse to mid per side:
1,000 shares × 1.5¢ × 2 sides = $30 per round trip.

Path B, $0.003/share commission, 0.5¢ adverse per side:
Commission 1,000 × $0.003 × 2 = $6. Slip 1,000 × 0.5¢ × 2 = $10. Total $16 per round trip.

Ten round trips a day and that's $300 versus $160 on paper. Flip the slip assumptions and Path A can win. That's the point. Zero commission with 1.5¢ adverse per side can lose to a small commission with tighter fills. Labels don't pay. Cents do.

How I log it in practice: at 09:41 ET I send a buy on a liquid name, mid is $63.205, fill comes back at $63.22. That's 1.5¢ adverse. Write it down. At 12:15 ET, same name, mid $63.505, fill $63.51, half a cent. Write it down. Two weeks of those lines tell me more than any comparison chart.

Failure modes:

- Logging only the trades I remember, which are usually the bad ones or the good ones.
- Comparing an open-heavy week on one path against a midday week on another.
- Ignoring rebates, fees and platform costs that sit outside the per-share number.

When the comparison doesn't matter much: if you trade a handful of times a week with small size, spend the energy on setups first. The sheet pays off most for frequent, larger share size scalping.

Don't pick a broker from a marketing PDF. Pick from two weeks of your own mid-to-fill log on the names you actually scalp.

Diagram: Path A vs Path B illustrative math.