In income-scaled DCA, the window picks the winner

Same rule, two leaders

Apply one income-scaled contribution rule — 30% of a $5,000/month income growing 5% a year — to 26 assets. Rank by XIRR with each asset starting at its own first investable month, and TSLA leads at 43.36%. Rank on W2018, the fair window where every asset receives the identical contribution schedule, and NVDA leads at 70.98%. Nothing about the assets changed; the window did.

Why DCA makes it worse

For price series, unequal windows already bias rankings toward lucky start dates. Money-weighted returns amplify this: an income-scaled saver invests their largest dollars late, so whichever asset happened to be strong during the final years of its own window dominates its XIRR. Own-inception rankings therefore mix three things — the asset, the era it listed in, and where its biggest contributions landed.

The discipline

Rank DCA outcomes only on common windows (the study designates W2018 because all 26 assets are present), keep the own-inception view as descriptive context, and read every extreme XIRR next to its account-path drawdown. A cash-flow-weighted ranking without a common window is a ranking of contribution calendars.