Last Updated: 2026-10-26
Scalping is a short-horizon form of intraday trading. Its small intended price moves make execution costs important, but neither fast decisions nor a high win rate establishes an advantage. This worksheet turns a setup idea into a paper study with an explicit cost model. It does not recommend a venue, leverage level or expected income.
All figures below are hypothetical, not trading results. Trading can lose capital, and simulation does not reproduce every live fill, queue position, interruption or funding charge.
Start with net expectancy, not the percentage of winners
Let R be the planned price loss on a losing trade, before costs. Assume every winner and loser has exactly the stated size:
Scroll sideways to see every column. Keyboard: focus the table region and use the arrow keys.
| Hypothetical setup | Win rate | Winner | Loser | Mean before costs |
|---|---|---|---|---|
| A | 70% | +1R | −1R | +0.40R |
| B | 40% | +3R | −1R | +0.60R |
The second has the lower win rate and higher modeled mean. This does not establish that it is better in practice: the frequencies and payoffs still require evidence, and the paths, costs and exposure differ. Subtract each setup's average cost in the same R units before comparing them. Report sample size and uncertainty; a run of wins is not a stable probability estimate.
Count both fills
In a synthetic example, buy 10 units at $100 and sell all 10 at $100.20. The gross price gain is $2. At an assumed 0.10% fee on each fill, the entry fee is $1 and the exit fee is $1.002. Net P&L is −$0.002, before any other charge. A 0.20% price move has already been consumed by the two fees.
If these are actual fill prices, their spread and slippage effects are already inside the price P&L. Do not subtract a second invented spread charge. In a simulation based on midpoint prices, model executable buy and sell prices instead, with a stated assumption for depth and queue behavior. Keep funding, borrowing and subscriptions as separate costs where applicable.
Percentage fees scale with notional; they do not inherently become a larger percentage just because the account is smaller. Minimum fees, order sizes and fixed subscriptions can create additional constraints. Leverage increases exposure and associated costs; it does not manufacture an edge.
Four ideas, four different tests
These are research hypotheses preserved from the original guide, not winning strategies:
Scroll sideways to see every column. Keyboard: focus the table region and use the arrow keys.
| Idea | Define before looking at outcomes | What could invalidate the interpretation |
|---|---|---|
| Return toward VWAP | Session anchor, observation interval, deviation measure and exit rule | A sustained trend; changing the anchor after seeing the move |
| Retest of a prior price area | Exact rule identifying the area and a maximum wait | Selecting only attractive examples after the fact |
| Reversal after a level break | Prior level, break threshold and confirmation window | Continuation rather than reversal; resting orders disappearing |
| Reaction after scheduled news | Source timestamp, release time and permitted observation window | Stale feed, wide spread, missing fills or unmodeled price gaps |
Do not label a candle an “institutional order block” or call a price move a deliberate trap without evidence of participants' orders and intent. A chart pattern can be defined and studied without that story.
Build a paper log that includes inconvenient cases
Freeze one rule set and evaluation period. For each qualifying signal, record its timestamp and timezone, information available at that moment, bid/ask or fill assumptions, quantity, fees, exit condition and result. Keep skipped, rejected, partial and missing-data cases. A limit price touched by a candle is not proof that your order filled.
Separate development data from a later evaluation period. If you revise a rule after examining losses, call the revised result a new test. Compare net results by the declared market regime without silently excluding poor sessions. Preserve the ledger so another reader can reproduce the arithmetic.
A stop-market trigger is not a guaranteed fill price; a stop-limit can fail to fill. The SEC's stop-order explanation covers this distinction for securities. Check the crypto venue's own contract and order documentation before assuming equivalent behavior. For exposure arithmetic, use the position-sizing worksheet.
Decide whether the study is usable
Report net P&L, trade count, average win and loss, exposure, maximum observed drawdown and excluded or missing records together. Compare the actual time spent with the schedule you can sustain. Thirty trades plus a ten-minute break after each of fifteen losses already requires 150 minutes of breaks alone; do not present such a plan as a two-hour session.
If price data, costs or fill assumptions are insufficient, the outcome is inconclusive, not profitable. A positive paper result is a reason to examine assumptions and out-of-sample behavior, not an instruction to fund an account. Check product eligibility directly; do not use location workarounds.
This page contains no affiliate signup recommendation. The site may earn commissions elsewhere; no personal trading performance or income claim is made here.