Lemma 2 — lower bounds for both constrained optimal values
DisprovedFreemiumActions.Threshold.lemma_2Let and . The publisher chooses a deterministic stationary policy . At each level the player probabilities satisfy , with all probabilities and in . Removing the offer gives and . Conditional on action the player advances one level with constant probability or stays, and stays at level if already there; quitting is absorbing. Assume are nondecreasing, nonincreasing, both quitting probabilities strictly positive, , and . The immediate revenues are and , with zero revenue after quitting. Let and . These are Assumptions 1–4 with the standing probability and revenue conventions.
For every engagement level , the optimal values constrained to action or at every visit to satisfy both inequalities
Both denominators are strictly positive quitting probabilities. Choices away from remain optimized over all stationary policies.
Differences from the source:
- encoding: engagement levels 0 through N, quit state −1 and controls 0 or 1 → Fin (N+1), Option with none for quitting, and Bool controls; reason: order-preserving state labels and true = offer; N=0 is retained
- reading: increase, decrease and increasing in Assumption 3 → weak monotonicity, with quitting positivity and tauM > tauI still strict; reason: the appendix uses nondecreasing and nonincreasing comparisons
- encoding: earns revenue, with muI < muM in Assumption 4 → 0 ≤ muI < muM; reason: makes the model’s implicit nonnegative-revenue convention explicit without normalizing either revenue
- encoding: real expected infinite-horizon total reward and visit counts → nonnegative infinite sums in ENNReal; policy values are finite sums of visits times rewards; reason: positive quitting on the finite state set makes these values finite; no truncation, discount or conversion of an infinite value to zero is used
- encoding: maximum over all stationary policies → indexed supremum over the finite nonempty set of all Bool-valued stationary policies; reason: a finite nonempty set of values has an attained maximum
- reading: prose defines constrained stationary values while the second line of (27) uses the unconstrained continuation value in a self-loop → constrain the stationary action on every visit to e; reason: matches the prose on p. 17 and the resolved self-loop formulas in the proof on p. 32; it is not a one-step deviation
import Definitions.Def_FreemiumActions_Threshold_Values
namespace FreemiumActions.Threshold
theorem lemma_2 {N : ℕ} (d : Data N) (h : Assumptions14 d)
(e : Level N) (he : 1 ≤ e.val) :
ENNReal.ofReal (d.qM e * d.muM / (1 - d.qM e)) ≤ constrainedValue d e false ∧
ENNReal.ofReal ((d.pM e * d.muM + d.pI e * d.muI) /
(1 - d.pM e - d.pI e)) ≤ constrainedValue d e true := by sorry
end FreemiumActions.Threshold
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.