What optimization does this use?

It uses greedy marginal-value allocation with diminishing returns. Each point goes to the stat whose next point currently adds the most weighted effectiveness. Because each stat's marginal value falls as points are added, the result is a balanced, near-optimal distribution rather than dumping everything into one stat.

Why model diminishing returns?

Most RPG stats give less benefit per point as they pile up, the way CRIT rate or ATK percent does in practice. The tool divides each stat's gain by one plus its current investment, so the first points in a stat are worth far more than the last, which steers a balanced build.

Does Genshin literally have Strength and Intelligence?

No. Genshin Impact uses ATK, HP, DEF, CRIT, and bonus percentages rather than classic Strength, Agility, and Intelligence. This is a general RPG allocation model you adapt by labeling the stats to your game and entering realistic weights and per-point gains.

How do the weight and per-point gain interact?

Marginal value is weight times per-point gain divided by one plus current points in that stat. Weight reflects how much you care about the stat; per-point gain reflects how much raw stat one point grants. Both push points toward that stat.

Can I force a minimum in a stat?

Set a high weight and gain for a stat you must invest in, and the greedy step will favor it early. For hard minimums, allocate those points manually first and run the optimizer on the remainder.

Genshin Impact Stat Optimizer

Email me this result

Get this tool's output sent to your inbox, plus one useful tool a week. No spam, unsubscribe any time.

Spreading stat points by feel usually leaves effectiveness on the table. This tool runs a real greedy marginal-value optimization with diminishing returns: it hands each point to the stat that currently buys the most weighted effectiveness, which naturally balances your build instead of over-stacking a single stat.

How it works

Each stat has a weight and a per-point gain, and its marginal value falls as you invest:

marginalValue(stat) = weight × perPointGain / (1 + pointsAllocated)

The optimizer loops once per available point, gives that point to the stat with the highest current marginal value, then recomputes. The total score is the sum of each stat’s weight × perPointGain × points.

Example and tips

With 20 points across an offense stat (weight 3, gain 2) and a defense stat (weight 1, gain 2), the optimizer front-loads offense but starts feeding defense once offense’s marginal value drops below it, landing on a split like 15 / 5. Tune the weights to your playstyle: a glass-cannon build raises the offense weight, a survival build raises defense.