Why one-time milestones change the incentive to repeat an action

By DX Research Group · · Incentives and participation

A marginal-reward example explains the difference between recognition and repeatable loss rewards.

A milestone awarded once changes what the next identical action can earn. That matters when the action carries financial risk. We would examine the marginal reward for repeating a qualifying event, rather than judging an incentive solely by the size of its first award.

The dated Chapter 0 terms describe wallet-level achievement bonuses that stack once during the September 22 to October 6, 2026 at 00:00 UTC window. Loss recognition is fixed by milestone, with the stated amounts independent of dollars lost. These mechanics recognize an event without giving every repeat of that event another identical grant.

Calculate the second event

Use an illustrative mechanism with a one-time 500-point award for the first qualifying loss event. The first event yields 500 points. A second event meeting the same threshold yields zero additional points from that milestone. The third also yields zero. The marginal grant falls to zero after the award is earned, even though trading costs and potential losses can continue.

Compare a hypothetical repeatable rule paying 500 points for every qualifying event. Ten events would yield 5,000 points, creating a reason to repeat the measured behavior if a participant values the points enough. That hypothetical rule is a design counterexample, rather than DXAP's published mechanism. It shows why the repeat condition matters as much as the threshold.

One-time treatment still leaves a boundary worth inspecting. A participant near a higher tier may perceive a reason to cross it. We would evaluate whether the interface clearly states that already-earned recognition is retained and whether it presents the cost and risk context without suggesting another trade. The public terms say no trade is needed to preserve an earned volume tier and identify points as program points.

Test the award state, then test behavior

A proposed technical fixture begins with a confirmed qualifying event, awards the milestone once, and then replays the same event after a data refresh. The point delta on refresh should be zero. A second distinct qualifying event should also leave that milestone's award unchanged. Different milestone tiers remain separate records under the declared stacking rules.

The fixture needs a wallet-and-period award identity, an event identity and a review state. A late-reconciled event can change eligibility while duplicate delivery cannot create another grant. These are proposed accounting requirements derived from the one-time mechanism; they do not establish how the current backend implements them.

A behavioral study asks a separate question. Compare aggregate activity around tier boundaries with suitable periods and participants, recording account age and prior activity. A descriptive jump near a threshold would justify investigation, while causal conclusions require a design that handles selection into participation. We would also inspect feedback reports about pressure to trade, rather than allowing volume alone to define success.

The fees and risks reference establishes that live trading carries costs and loss exposure. Recognition belongs beside that reality. Its useful role is to acknowledge participation without implying that points compensate for losses or that crossing another milestone improves the agent's economic performance.

Sources

Related field notes