Minimum-Quantity / All-or-None Constraint Risk: A Slippage Modeling Playbook

Date: 2026-03-09
Category: research (execution microstructure)

Why this matters

MinQty and AON constraints are often used to avoid toxic tiny fills.
But in practice they can turn displayed liquidity into conditional liquidity:

• top-of-book looks available,
• yet your order is ineligible for most contra flow,
• completion risk rises,
• and the final execution often becomes a late aggressive catch-up.

So the cost is not just spread/impact. The hidden cost is eligibility friction.

Failure pattern (common in production)

1. Desk sets large MinQty (or uses AON) to avoid micro-fills.
2. Market fragment size shifts smaller (news burst, close approach, venue mix change).
3. Eligibility rate collapses; order stays live but inactive.
4. Residual grows as deadline approaches.
5. Router flips to urgent marketable flow.
6. Realized slippage is worse than if we had accepted controlled partial fills earlier.

This is a classic “quality-first becomes completion panic” path.

Cost decomposition

For parent order notional (Q):

[ C_{total}=C_{px}+C_{impact}+C_{delay}+C_{opp}+C_{reset} ]

where

• (C_{px}): explicit spread/fees
• (C_{impact}): market impact from executed child flow
• (C_{delay}): adverse drift while waiting for eligible counterparties
• (C_{opp}): opportunity loss from unfilled residual at cutoff
• (C_{reset}): additional cost from switching late to urgent tactics

With hard size constraints, (C_{delay}) and (C_{opp}) dominate tails.

Core signals (new)

1. Eligibility Hit Rate (EHR)

   • Fraction of observed contra events that satisfy current MinQty/AON condition.

2. Constraint-Induced Idle Time (CIIT)

   • Share of parent lifetime where order is live but has no eligible match opportunities.

3. Constraint Regret (CRG)

   • Ex-post delta between realized cost and counterfactual cost under relaxed threshold policy.

4. Late Urgency Conversion Rate (LUCR)

   • Probability of switching from constrained passive policy to urgent aggressive policy near deadline.

5. Eligible Depth Ratio (EDR)

   • Effective executable depth under current constraint divided by displayed depth.

Displayed depth can look healthy while EDR collapses.

Modeling architecture

1) Branch model for constrained orders

Model parent outcome as branch process:

• (B_1): eligible fill path (good)
• (B_2): delayed partial completion (mixed)
• (B_3): no eligible flow then urgent conversion (bad)
• (B_4): miss cutoff / residual carry (worst under strict mandates)

Estimate branch probabilities conditional on state (x_t):

[ P(B_k\mid x_t,\theta_{min}) ]

where (\theta_{min}) is active size constraint.

2) Eligible-flow hazard

Estimate hazard of receiving eligible contra flow:

[ \lambda_{elig}(t)=h(\text{queue state, lot-size mix, venue, TOD, vol regime},\theta_{min}) ]

Lower (\lambda_{elig}) implies higher delay and deadline risk.

3) Counterfactual policy engine

For each decision step, evaluate candidate thresholds:

[ \theta \in {AON,\ MinQty=2000,1000,500,none} ]

with expected objective:

[ J(\theta)=\mathbb{E}[C_{total}\mid x_t,\theta]+\alpha\cdot \text{ResidualRisk}_{q95} ]

Pick (\theta^*=\arg\min J(\theta)), subject to mandate constraints.

State controller (practical)

• STATE 1: FLOW-HEALTHY

  • EHR high, CIIT low
  • Keep constraint, passive-first

• STATE 2: FRAGMENTING

  • EHR falling, lot-size distribution shifting smaller
  • Step-down MinQty gradually; widen venue set

• STATE 3: BLOCKED

  • CIIT high, LUCR risk rising
  • Relax constraint aggressively; allow controlled partials

• STATE 4: DEADLINE-RISK

  • high residual + low time-left
  • convert to completion-first policy with explicit slippage budget

Use hysteresis (entry/exit buffers) so policy does not flap on transient noise.

Guardrails

1. No hard AON near cutoff

   • Auto-disable AON inside final execution window unless legally mandated.

2. MinQty decay schedule

   • Define time-based or risk-based relaxation curve in advance.

3. Residual cap by time-left

   • If residual exceeds threshold at T-minus buckets, force policy relaxation.

4. Venue-specific eligibility map

   • Track each venue’s observed eligible-flow profile; avoid one-size-fits-all thresholds.

5. Tail-budget gate

   • If projected q95 C_total breaches budget, prioritize completion over fill-size purity.

Backtest / replay requirements

• Event-level lot-size distribution and queue updates
• Exact child-order constraints (AON, MinQty) in order log
• Deadline/cutoff semantics (close, auction windows, internal risk cutoff)
• Policy-switch telemetry (when and why constraints were relaxed)
• Counterfactual replay under alternative threshold schedules

Without counterfactual replay, constraint regret stays invisible.

Minimal rollout checklist

• Add EHR/CIIT/EDR/LUCR metrics to real-time dashboard
• Implement branch-probability nowcast with q95 residual risk
• Add automatic MinQty decay and AON cutoff guardrail
• Build venue-level eligible-flow profile table
• Gate rollout on q95 total-cost + completion-rate improvement
• Run 2-week shadow comparison (current vs adaptive-threshold policy)

Bottom line

MinQty/AON is not a harmless order-typing choice.
It is a latent slippage lever that trades early fill quality against late completion risk.

If you don’t model eligibility friction explicitly, the desk will keep “saving spread” early and paying it back with interest near deadline.