What it is
Scalping the DOM (depth of market, the price ladder) means reading the order-book and time-and-sales to trade very short moves, often two to six ticks, dozens of times per session. Instead of a chart pattern, the trigger is order-flow behaviour: a large resting bid that absorbs aggressive selling, a stack of offers that gets pulled, a burst of market buys that fails to move price. The stop is a tick or two, the target is a few ticks more, and the whole trade is over in seconds to minutes.
This is the most execution-dependent playbook in the wiki. It is included because people ask about it constantly, not because it is a good starting point.
The logic
In a liquid futures market, price moves when aggressive orders exhaust the passive liquidity at a level. A scalper tries to detect exhaustion or absorption before the chart shows it. When 3,000 contracts trade into a bid and the bid does not drop, someone is absorbing; if the sellers then stop, the path of least resistance is up by a few ticks as the sellers cover. The scalper joins the absorber.
On the other side are the aggressive traders who just hit that bid, and other scalpers who misread the same tape. The edge, when it exists, is tiny per trade and depends entirely on paying less in costs than the average opponent.
Setup rules
- Market: the most liquid futures only, such as es, treasury futures, or cl during the US session. Anything thinner is dominated by spoofing and noise. Stocks are worse because liquidity is fragmented across venues.
- Tools: a DOM with volume-at-price, a footprint or bid-ask delta display, and a data feed with unfiltered tick data. A delayed or aggregated feed makes this unworkable.
- Timeframe: none in the conventional sense; use a 500-tick or 1-minute chart only for context.
- Conditions: trade only during high-volume windows (first 90 minutes and last 60 minutes of regular hours); trade only at a reference level (vwap, prior day high or low, a point-of-control) where large passive orders are likely to be real.
- Absorption trigger: cumulative delta of at least 3x the average bar delta into a level with price movement of one tick or less.
Entry, stop, target
Join the absorbing side with a limit order one tick inside the absorbing level (to avoid queue position problems), or hit the market if the tape shows the aggressors giving up. Stop is two ticks beyond the level. Target is four to six ticks.
| Item | Level | Notes |
|---|---|---|
| Absorbing bid | 5,000.00 | 4,000 contracts absorbed without printing lower |
| Entry | 5,000.25 | Limit, filled as sellers pause |
| Stop | 4,999.50 | 3 ticks, risk $37.50 per contract on ES |
| Target | 5,001.50 | 5 ticks, reward $62.50, 1.7R |
Scalping requires an immediate exit if the level breaks; there is no "give it room". Many scalpers scratch (exit at breakeven) far more often than they take a loss, which is what keeps the win-rate high enough to survive.
Position sizing and risk
A three-tick stop on one ES contract is a small dollar risk, which tempts people to trade far too many contracts. Set risk per trade by the account, not by the tick size, using /tools/position-size, and impose a hard daily loss limit at 2 percent of equity, after which the platform is closed, per /learn/risk-management. Scalpers blow up through frequency, not through any single trade.
What breaks it
- Costs dominate. With a target of five ticks and a stop of three, a round-trip commission of $4 plus one tick of slippage is more than a third of the average win. Your net expectancy can be negative while your gross is positive.
- Spoofing and pulled orders. Large displayed size is often not real and vanishes before it trades. Reading absorption from executed volume, not displayed size, is the only defence, and it is imperfect. See spoofing.
- Latency. Retail infrastructure is slower than the firms on the other side. A tape that "looks" one way may already be the other way by the time you click.
- Edge decay. Order-flow signals that were reliable a decade ago have been automated; the discretionary scalper competes directly with machines.
- Fatigue and tilt. Making 50 decisions a day at full attention degrades quickly; most losses cluster late in the session. See tilt.
- Regime. Absorption at levels works in balanced, two-sided markets; during high-volatility news it is noise.
How to test it
Backtesting discretionary tape reading is nearly impossible, so the protocol is forward: 20 sessions on a simulator with realistic fill logic (queue position matters; assume you are last in the queue at your limit price), logging every trade with the level, the trigger, the outcome and the time of day. You need at least 300 trades to distinguish a 1.5R-at-55-percent edge from zero. Then 20 sessions with one micro contract live, comparing actual fills against the simulator. If live results lag the sim by more than 20 percent, the difference is your costs and latency, and no amount of screen time fixes that.
Variations
- Level scalping without absorption, trading only at prior day high, low and settlement.
- Delta divergence scalps: price makes a new low while cumulative delta does not.
- Momentum scalps that join a burst of aggression rather than fading it; see momentum-ignition-volume.
Further reading
order-book, level-2, time-and-sales, iceberg-order, spoofing, tick, tick-value, scalping, slippage, day-trading-margin.
Related playbooks: momentum-ignition-volume, range-day-playbook, trend-day-playbook, vwap-reclaim-reject