Skip to main content

Simulated learner run

Test the adaptive policy before a learner sees it.

This page runs the probability/concentration gold bridge set against fixed synthetic learner profiles. It checks whether the policy routes learners through prerequisite claims before sending them to finite-class generalization.

Gold set

bridge v1

Profiles

4

Claims covered

8

Steps per run

10

Gold set

Probability / concentration bridge gold set v1

A curated 10-question diagnostic bridge from sub-Gaussian tails through finite-class uniform convergence.

probability-concentration-bridge-v1

Synthetic profilePrerequisite-aware policyJumpiest baseline

Probability basics gap

probability-basics-gap

Best raw accuracy: Prerequisite-aware at 100%.

Prerequisite-aware

0 prerequisite jumps

accuracy 100%

mastery 43%

Next checkpoint: probability-event-algebra-and-moments

First pick: question:diverse-state-theorem-001

Explore/exploit

7 prerequisite jumps

accuracy 80%

mastery 43%

Next checkpoint: probability-event-algebra-and-moments

First pick: question:logic-ml-foundations-030

Concentration assumptions gap

concentration-assumptions-gap

Best raw accuracy: Static path at 80%.

Prerequisite-aware

0 prerequisite jumps

accuracy 40%

mastery 42%

Next checkpoint: probability-event-algebra-and-moments

First pick: question:diverse-state-theorem-001

Explore/exploit

6 prerequisite jumps

accuracy 50%

mastery 42%

Next checkpoint: probability-event-algebra-and-moments

First pick: question:diverse-state-theorem-001

Uniform convergence bridge gap

uniform-convergence-bridge-gap

Best raw accuracy: Prerequisite-aware at 70%.

Prerequisite-aware

0 prerequisite jumps

accuracy 70%

mastery 37%

Next checkpoint: probability-event-algebra-and-moments

First pick: question:mcq-fundamental-theorem-012

Explore/exploit

7 prerequisite jumps

accuracy 60%

mastery 37%

Next checkpoint: probability-event-algebra-and-moments

First pick: question:logic-ml-foundations-030

Rusty probability theorist

rusty-probability-theorist

Best raw accuracy: Explore/exploit at 90%.

Prerequisite-aware

0 prerequisite jumps

accuracy 60%

mastery 55%

Next checkpoint: probability-event-algebra-and-moments

First pick: question:subgaussian-sum-tail-bound-012

Review due first

5 prerequisite jumps

accuracy 40%

mastery 54%

Next checkpoint: probability-event-algebra-and-moments

First pick: question:subgaussian-sum-tail-bound-012

Review focus

  • sub-Gaussian MGF assumptions
  • Hoeffding versus Chebyshev scope
  • finite union bound in learning theory
  • uniform convergence prerequisites

Project unlocks

  • confidence-radius-widget
  • finite-erm-sample-size-planner

Interpretation

This simulation validates routing behavior on fixed profiles. It does not estimate real learner ability. Real calibration starts after answer events are collected.