Skip to main content
Track Fixed contract sizing
2 / 7
Search the archiveLibrary
Library

1994issue C041-10

Starting capital from worst-case portfolio walk-forwards

The historical workflow set a 2 percent per-trade cap and a 35 percent overall-capital cap first, then replaced a single-date capital figure with a walk-forward distribution of required starting capital.

  • A 2 percent per-trade cap and a 35 percent overall-capital cap were set as simultaneous exposure bounds before any start-date capital study.
  • A single January 1 1981 start produced a 5459 capital-required figure that the same study later treated as an optimistic artifact of a favorable start date.
  • Consecutive daily starts through 1990 produced an average capital requirement of 10576 and a standard deviation of 2049, which pushed a conservative start band to at least 17000 to 19000.
  • The same 2 percent per-position rule on a 25000 account forced many signals to be skipped because initial risk exceeded 500.
Entries in this reading3 entries
Drawdown limit5 readingsRisk of ruin6 readingsCurrent trackFixed contract sizing7 readings

Exposure bounds came first

A 2 percent per-trade cap and a 35 percent overall-capital cap were specified as simultaneous exposure bounds before any start-date capital study.

Editorial note: TradersWeek maps that setup onto Fixed contract sizing, Risk of ruin, and Drawdown limit. Those filters keep a loss or exposure decision bounded before a trade is placed and throughout the position. The archive describes the caps and the later capital study without assigning those method names.

A loss streak was not treated as proof

Multiplying one observed streak of 13 losses by an average loss of 1455 plus a 3000 margin buffer produced a 21915 starting-capital estimate that the text rejects as under-supported.

One start date was later treated as favorable

A single January 1 1981 start produced a 5459 capital-required figure that the same study later treated as an optimistic artifact of a favorable start date.

Repeating the portfolio simulation from consecutive daily starts through 1990 produced an average capital requirement of 10576 with a standard deviation of 2049.

Starting capital from one date versus walk-forward starts

A single January 1981 start made the book look cheap at $5,459. Across walk-forward start dates the same rules needed $10,576 on average and $14,105 in the worst case, so the one-date figure is the easy outcome, not the funding rule. Those three amounts are the figures the article states from the original simulation and from the worst-case study table.
A single January 1981 start made the book look cheap at $5,459. Across walk-forward start dates the same rules needed $10,576 on average and $14,105 in the worst case, so the one-date figure is the easy outcome, not the funding rule. Those three amounts are the figures the article states from the original simulation and from the worst-case study table.Six-futures portfolio (British pound, cotton, Deutschemark, yen, orange juice, Swiss franc) · Walk-forward start dates, 1981–1990 · 1981-01-01T00:00:00.000Z to 1990-12-31T00:00:00.000Z

Walk-forward starts after the end of 1990 were dropped because they covered less than a year. Each run used a 2 percent per-trade cap, a 35 percent overall-capital cap, six-tick slippage and a $50 commission.

A small shortfall chance was still rejected

Under a normality assumption, 14105 of starting capital left about a 4 percent chance of still being underfunded, which the text called unacceptable.

Planning for a three- or four-standard-deviation shortfall, rather than the 1.75-sigma 14105 level, raised the conservative start band to at least 17000 to 19000.

Editorial note: TradersWeek reads the refusal to accept that remaining underfunded chance as a Risk of ruin filter, and the higher start band as a Drawdown limit on opening capital. The archive reports the figures without those labels.

The same cap skipped signals

The same 2 percent per-position rule on a 25000 account forced many signals to be skipped because initial risk exceeded 500.

Editorial note: TradersWeek reads that skip rule as Fixed contract sizing applied before entry.

Educational research material, not investment advice. Historical source context does not establish present-day performance.
2 of 7 in the Fixed contract sizing track
19941-6 pp.Next on Fixed contract sizingBound small-account risk before adding leverageCap the cash lost on one fill first, so a short run of bounded losses cannot shrink equity until the next trade cannot be placed.
All readings on this track · 7 readings
  1. 1987Volatility-layered mechanical system with fixed contracts
  2. 1994Starting capital from worst-case portfolio walk-forwards
  3. 1994Bound small-account risk before adding leverage
  4. 1996Variable position size after entry
  5. 1996Equity path filters for contract size and drawdown
  6. 1997Stop distance, equity caps, and trading halts
  7. 1999Size-matched buy-and-hold evaluation for stock systems
By fieldSizing and leverage6 readingsBy year19949 readings
All 7 readings tagged Fixed contract sizing
Also on Fixed contract sizing5 readings
2018Evaluating a normalized risk index for drawdown and exposure limitsDrawdown limit1996Equity path filters for contract size and drawdownDrawdown limit1994Bound small-account risk before adding leverageRisk of ruin1992Equity-curve average as a live-capital gateDrawdown limit1994When standing puts fail the drawdown testDrawdown limit