LP Position Sizing: Unit Size and Scaling Rules

How to size DLMM positions so a losing streak cannot kill the book: unit sizing, scaling gates, and portfolio-level thinking.

Short answer

Size every position as one fixed unit of SOL or USD. Raise the unit only after a net-profit milestone that you set in advance.

In this guide 5 sections

Size by units, not by vibes

A documented copy-LP wallet ran ~10.75 SOL per position across 9,806 closed positions, turning consistent unit sizing into a 78% win rate and ~40 SOL per month. The unit size is what makes the statistics meaningful: with wildly varying sizes, one oversized loss erases twenty wins.

Practical rule: define one position unit (a fixed SOL or USD amount) and never exceed one unit per new setup. Let the count of positions, rather than the size, express conviction.

Scale on net, gate on proof

The 36-positions-per-week trader held sizing below $100 until net PnL grew by a further $800, and only then increased size. That is the discipline: scale after a measured net milestone, not after a good day.

Set your own gate before you start: for example, scale up one notch per $X of cumulative net profit, and drop back a notch after a drawdown of one unit. Write it down before the emotions show up.

Test size and exit liquidity

A $100 practice budget let one trader test ideas and mistakes before raising size. The reported return was one personal month, not a target for new LPs. Decide your own net-PnL gate in advance (see also the fee kickback for cost recovery) and keep the test size affordable to lose.

A trader who kept a fixed LP size for a long period warned that a larger memecoin exit can suffer heavy slippage. Before raising size, estimate what your full token inventory could sell for after a downside range break. Pool TVL and market cap do not guarantee an exit at the quoted price.

Portfolio over position

Degen basket strategy explicitly runs on total portfolio PnL: no per-token targets, no closing winners early, rugs left alone at -99%. The unit economics only work if individual outcomes are allowed to happen.

The failure mode is watching single positions. A position sitting at the bottom of its range does not need action if the basket is net positive and the exit rules have not triggered.

Expectation setting

Even good LP books have losing weeks: a 78% win rate still produced five losing trades out of 36 in one week, including a -$47.81. What keeps the book alive is that average wins ($10-55) are protected by hard exits and the unit size caps each loss.

Never size so large that one out-of-range exit or one rug forces a strategy change. If a single position outcome would change your plan, the unit is too big.

Sources & inspiration