The risk model¶
risk.json in the project, or Arvo's defaults when there is none. One
model, in the backtest and in every session. A field not set is not set,
and the Risk tab says so rather than showing a default as if you chose it.
| Field | What |
|---|---|
stop_atr_multiple |
Stop distance as a multiple of ATR. Not set runs with no stop |
atr_period |
Bars the ATR is averaged over |
risk_per_trade |
Fraction of starting capital a stop-out may cost. Needs a stop |
max_position_fraction |
Fraction of starting capital one position may occupy. Above 1 is leverage |
max_drawdown |
Fall from the account's peak that stops the rule for the run. Permanent |
max_concurrent_positions |
The most positions held at once across a book |
max_daily_loss |
Realised loss in one day, as a fraction of starting capital, that refuses new entries |
correlation_cap |
Above this pairwise correlation, names are one bet; how many such bets may be held. Refuses when unknown |
sector_cap |
The most positions one sector may hold, with each name's sector. Refuses an unlabelled name |
day_trading |
unconstrained, or pattern_day_trader: three round trips per five days under $25k |
Costs — slippage, commission, a per-fill fee, an option spread — are on the experiment, recorded with every finding, and the paper session's measured slippage is reported against what the backtest assumed.