Position sizing and strategy sandbox

Before you draw conclusions

2026 has been a good year for this kind of trading, and in a good year the aggressive settings look like free money: fewer, bigger positions, leverage, adding to stocks that keep falling. In weak years, such as 2011, 2018 and 2022 for dip buying, the same settings mostly made the losses deeper. So use the sandbox to see the trade-off between return and worst drop, not to find the setting with the biggest number. The sections below go through the controls in order. The trading rules step by step, when to buy and when to sell a gap up are on how to trade the strategy.

Positions

The number of positions decides how much one bad trade, or one bad week, can hurt. With equal sizes it also sets the size of each trade: 15 positions means 6.7% of the account per trade, 5 positions means 20%.

With few positions, a handful of stocks that keep falling decide the result, and luck plays a big part: which signals happened to arrive while a position was free. More positions make the ride smoother. Past 10 to 20 positions, more mostly means smaller trades and more cash waiting.

Try it: set Positions to 5 and hover over the return to see the range across runs.

Size per trade

The sizing modes decide what happens on days with fewer signals than free positions. Fixed size gives every trade the same share of the account and leaves the rest in cash. Fill free positions spreads the money meant for the free positions over the trades that came in, so quiet days get fewer, bigger trades. Deploy all capital spreads all spare buying power over the day's trades. Max per trade caps any single trade in the last two modes, so a day with one signal cannot put the whole account into one stock.

How big is too big? The Kelly criterion gives the share of capital to bet when you know your odds: the win rate minus the loss rate divided by the win/loss ratio (Kelly 1956). For dip trades it comes out far higher than most traders would bet, because it assumes independent bets one at a time, while dip trades fall together in a sell-off. Traders who use it usually bet half of the Kelly amount or less (Thorp 2011).

Try it: switch Sizing to Deploy all and compare the worst drop and capital in use.

Leverage and drawdowns

Leverage lets open positions add up to more than the account. The difference is borrowed and charged margin interest. It scales the gains and the losses alike, so the worst drop grows with it.

The worst drop matters more than the return, because losses and recoveries are not symmetric: a 20% drop needs a 25% gain to get back to even, a 50% drop needs 100%, and a 60% drop needs 150%.

Try it: load the All in, 2x template and look at the worst drop.

Repeat signals

A falling stock often signals two or three days in a row. Repeat signal decides whether you add another position each time, add half of one, or skip it. Adding is averaging down: it makes the most when the stock bounces and loses the most when it does not, because more money sits in stocks that are still falling. Each addition uses a position, so the total stays within the account.

Try it: set Repeat signal to Skip and watch Skipped signals and Capital in use.

Crowded days

On some days dozens of stocks signal at once. That is usually one broad sell-off, not dozens of separate chances. Crowd scaling makes each trade smaller on those days: at 1, ten signals get a tenth of the normal size each. It makes the worst days less painful, mostly by putting less money to work, so expect a lower return along with the smaller drops.

Try it: set Crowd scaling to 0.5 and compare Capital in use.

Questions

How much should I risk per trade?

A common rule of thumb is to lose no more than about 1% of the account if a trade goes wrong. For a strategy without a fixed stop, such as the RSI(2) dip rule, it is simpler to split the account into equal positions and choose a number of positions you can live with in a bad month.

How many positions should a swing trader hold?

Enough that no single trade decides the month. With a handful of positions, a few stocks that keep falling can decide the whole year; spreading over 10 to 20 positions makes the swings much smaller.

What is maximum drawdown?

The largest fall of an account from a peak to a later low, in percent. It matters because losses need larger gains to recover: a 50% drawdown needs a 100% gain to get back to even.

Should I average down on a losing position?

It depends on the strategy. For a dip-buying rule, a second signal on a stock you already hold is a new setup, and adding to it can raise returns. It also deepens the losses when the stock keeps falling, because more money ends up in it.

What is the Kelly criterion?

A formula for the share of capital to bet when you know your odds: win rate minus loss rate divided by the win/loss ratio. It assumes independent bets. Trades that fall together in a sell-off are not independent, so traders who use it usually bet a fraction of it.

References

1. Kelly, J. L. (1956). A New Interpretation of Information Rate. Bell System Technical Journal 35(4), 917–926. doi:10.1002/j.1538-7305.1956.tb03809.x
   Origin of the Kelly criterion for sizing repeated bets.
2. Thorp, E. O. (2011). The Kelly Criterion in Blackjack, Sports Betting, and the Stock Market. In The Kelly Capital Growth Investment Criterion, World Scientific, 789–832. doi:10.1142/9789814293501_0054
   Kelly in practice, and why fractional Kelly is used.