What it answers
You know your win rate and your reward-to-risk. What you rarely see is what they are worth once commissions, slippage and the
size you trade are counted, and how far a real year can land from the average one. Set your numbers and the board shows
three things at once: the arithmetic of a single trade, what a day, a month or a year makes on average, and what it makes
on a typical path and in a bad year, from 20,000 simulated paths of your own trading rules.
How to use it
Move the sliders in each group; every number and picture follows at once, and the address bar keeps your settings, so a
link you copy opens the board exactly as you left it. Full screen ↗ gives the board the whole window.
- Strategy: your win rate (of the trades that are decided), your reward-to-risk (average win over average loss), and
your scratch rate. A scratch is a trade you close at, or very near, your entry price: when a trade does not move
your way at once, you get out flat. It is neither a win nor a loss, but you still pay both commissions and the slippage,
so scratches only ever cost money. Scalpers scratch a lot. The scratch slider is the share of all trades closed that
way, and the win rate is then the share of the remaining trades, the decided ones, that win: at a 50% win rate and 20%
scratches, 40% of all trades win, 40% lose and 20% go out flat.
- Size, three ways: a fixed number of shares per trade; a risk % of equity, where the shares are that risk divided
by your stop (your size then grows and shrinks with the account, and the price does not enter); or a fixed number of dollars per trade,
the position’s value, where the shares are those dollars divided by the price. The size is set at each day’s open and
held all day. Two readouts show what the price does: the position value (shares × price) and the **buying power
used** (the position value as a multiple of your equity): with shares or a risk %, the price changes your size only when
the position needs more buying power than you have.
- Account: your starting capital and your buying power (margin, from ×1 to ×10). When the size you ask for needs more
buying power than you have, the board caps it and says so.
- Execution: the stock price, your stop in cents, the commission per share and its minimum per order, the slippage on
each fill, and the extra slippage you take when a stop is hit in a fast tape.
- Activity: trades per day (up to 300 for a hyper-scalper), trading days per year, and the horizon, from one day to a year.
- Safety: a daily loss limit (stop for the day), a quit level (stop for good), and gaps that jump through your stop.
The readouts come in two kinds. Per trade is plain arithmetic, exact. Over the horizon is simulated: the mean, the
median (the typical path), a bad case (the 10th percentile), a good case (the 90th), the chance of ending down, the chance
of hitting your quit level, and the median of each path’s worst drawdown.
The arithmetic
With $n$ shares, a stop of $s$ dollars per share, reward-to-risk $R$, an extra stop-out slippage $e$ per share, and
round-trip costs $k$ (two fills, each charged its commission or the minimum, whichever is larger, plus the slippage):
where $p_w$ and $p_l$ are the chances of a win and a loss once scratches and gaps are taken out. The win rate at which this
is zero is the breakeven win rate; without scratches or gaps it is
so costs matter only as a share of your stop: a tight stop makes every cent of cost count more.
Worked examples
A slow day trader. Open the board at these settings: a 50% win rate at 1.5 to 1, 500 shares of a
USD 10 stock with a 20-cent stop, four trades a day. Each trade expects USD 7.50 after costs of
USD 15.00, so the year expects USD 7,560, while the costs come to USD 15,120: the
broker and the tape take twice what you keep. Costs move the breakeven win rate from 40.0% to
47.1%. The simulated year ends at a median of plus USD 7,560, a bad case of
plus USD 2,460, and 3.2% of the paths end it down.
A hyper-scalper at 300 trades a day. Open the board at these settings: a 55% win rate at 1 to 1,
200 shares with a 5-cent stop, a commission of 0.3 cents a share with a 1-dollar minimum, no slippage on entry and one
cent on stop-outs. A 55% win rate sounds like an edge. It is not: each trade expects minus USD 1.90, the day
minus USD 570.00, and the month (21 days) minus USD 11,970; 100% of the simulated months end down.
The breakeven win rate is 63.6%, not 50%. The minimum commission is part of it; take it
away (open) and the day still expects minus USD 330.00 at a breakeven of 60.0%,
because one cent of stop slippage is a fifth of a 5-cent stop.
Margin ×10 with rare gaps. Open the board at these settings: a 50% win rate at 1.5 to 1, risking 2% of
equity per trade on a USD 50 stock with a 25-cent stop, buying power up to ×10, and two trades in a thousand gapping 5% of
the price through the stop. Without the gaps (open) the year’s median is plus USD 56,371,
the bad case plus USD 13,583, and 1.0% of the paths end down. With them the median falls to
plus USD 29,988, the bad case to plus USD 2,688, 6.2% end down, and the median
worst drawdown grows from 36% to 45%. Now move the risk slider. The chance of
ending the year down is 96% at 0.1% (the minimum commission eats trades that small),
3.2% at 0.5%, and 22% at 5%, where the gaps hit a size the margin allows.
The size panel shows the other side: at 5% the mean final equity is USD 791,823 and the median USD 80,110.
The average is carried by a few lucky paths; the typical one never sees it.
Assumptions and limits
- Trades are independent and the win rate is fixed: no streaks, no good or bad weeks, no change in the market.
- Every win is $R$ times the stop and every loss the stop plus the stop-out slippage; real trades vary around those averages.
- The size is set at each day’s open from that morning’s equity and held all day.
- Positions are intraday: no overnight financing, no margin calls beyond your own quit level, no taxes.
- A gap is a loss of the stop plus the gap size, as often as you set it; real gaps cluster.
- The board knows no broker’s prices and no regulator’s rules. It works with the numbers you give it.