Does raising leverage as price falls improve leveraged DCA?
Not beyond arithmetic. Raising leverage at each step of a dip-buying ladder lowers the average entry price by construction, but on twelve crypto perpetual futures from 2020 to 2026 its advantage over a decreasing schedule (+0.44 points of budget per episode) was smaller than the same ladders show on directionless random paths (+0.76), and it did no better than equal leverage. Liquidation rates were set by the ladder’s geometry.
The claim
Pandey asks how to spread a fixed total amount of leverage across the steps of a dollar-cost-averaging ladder into a drawdown. By the rearrangement inequality, an increasing schedule — little leverage on the early, expensive buys and more on the late, cheap ones — gives the lowest average entry price on a falling path. Back-tests on gold, bitcoin, ether, SPY and QQQ show the increasing schedule with a better Sharpe ratio and fewer liquidations in all five markets.
What we tested
Twelve perpetual futures, hourly, January 2020 to October 2026, with fees and actual funding. We used the paper’s ladder — five buys at 3% intervals from −3% to −15%, equal margin per step, isolated margin — with leverage rising from 2× to 10×, falling from 10× to 2×, or 6× throughout. The comparison that matters runs the same ladders on randomised paths that keep the market’s volatility but have no direction, so that any difference left there comes from geometry alone.
Result
Per episode, as a share of the budget, the increasing schedule returned −1.01%, equal leverage −0.93% and the decreasing schedule −1.45%. The increasing schedule’s advantage over the decreasing one, 0.44 points, is smaller than the 0.76 the same ladders produce on directionless paths, and against equal leverage it is −0.07. Liquidations hit 16.3%, 18.3% and 32.4% of episodes, almost exactly the 15.5%, 17.8% and 31.9% seen on directionless paths. Against the decreasing schedule the increasing one is a bet on depth: it wins big when all five steps fill and loses when the dip is shallow.
Five-step ladder to −15%, twelve perpetual futures, 2020 – 2026
| Increasing 2× → 10× | Equal 6× | Decreasing 10× → 2× | |
|---|---|---|---|
| Result per episode, % of budget | −1.01 | −0.93 | −1.45 |
| Episodes with a liquidation | 16.3% | 18.3% | 32.4% |
| Same, on directionless paths | 15.5% | 17.8% | 31.9% |
What it means
The lower average entry price is guaranteed on any path that falls step by step, which is why it shows up in every market — and it says nothing about whether the dip was worth buying. On crypto perpetual futures over this period the ladders lost money under every schedule, and liquidation risk was set by how deep the ladder reaches with how much leverage, not by the order of the steps.
What would change the verdict
Deeper ladders, to −25% or −40%, gave the increasing schedule about one point per episode over equal leverage, not enough to be reliable; a test fixed in advance on episodes after this date would settle it. Exchanges with different maintenance-margin rules change the liquidation geometry and need their own run.
This note re-tests a claim made by the cited authors on our own crypto market data; the summary of their method is ours. Tested · published · last reviewed . Research, not investment advice.