Limitations of IPO Allotment Probability Models
Introduction
IPO allotment probability models are widely used to estimate the likelihood of receiving an IPO allotment based on live subscription data. While useful for setting expectations, these models have inherent limitations due to data availability, regulatory processes, and simplifying assumptions.
This document explains the key limitations investors should understand before relying on IPO allotment probability calculations.
Limitation 1: Exact Number of Applications Is Unknown
Stock exchanges publish subscription in terms of shares or lots bid, not the exact number of unique applications.
Probability models must approximate application counts, usually by assuming one-lot bids in the retail category, which may not always be accurate.
Limitation 2: Assumption of One Lot per Retail Application
Most retail-focused models assume each valid retail investor applies for exactly one lot.
In reality, some investors apply for multiple lots, which inflates demand figures without proportionally increasing the number of applications.
Limitation 3: Invalid and Rejected Applications
Probability models usually ignore invalid, duplicate, or rejected applications caused by PAN mismatches, UPI failures, or technical errors.
The actual rejection rate is only known to the registrar after bid reconciliation, which can change final probabilities.
Limitation 4: Bid Modifications and Cancellations
Live subscription data includes bids that may later be modified or withdrawn before issue close.
Probability models using intraday data cannot fully adjust for these changes, especially on the final bidding day.
Limitation 5: Cut-Off Price Assumptions
Models typically assume that all retail investors bid at the cut-off price.
Applicants bidding below the final cut-off price are excluded from allotment, which can reduce the effective competition pool.
Limitation 6: Registrar Allotment Algorithms Are Not Public
While SEBI mandates fair and transparent allotment, the exact randomization algorithms used by registrars are not publicly disclosed.
Probability models assume uniform randomness, but the actual implementation details are unknown.
Limitation 7: Category-Level Aggregation
Subscription figures are published at a category level (RII, NII, QIB), not at an individual application level.
This aggregation hides variations such as bid sizes, pricing strategies, and investor behavior within each category.
Limitation 8: SME IPO-Specific Constraints
In SME IPOs, small issue sizes and large lot values result in very few total applications.
Even minor changes in bids or rejections can significantly alter allotment outcomes, making probability estimates less stable.
Limitation 9: Timing Differences Between Data and Allotment
Probability models are based on snapshot subscription data, while actual allotment happens days later after reconciliation and regulatory checks.
Late-stage changes are not captured in most public-facing models.
How Investors Should Use Probability Models
- Use probabilities as a directional indicator, not a guarantee
- Compare probabilities across IPOs rather than focusing on absolute numbers
- Combine probability estimates with fundamentals, valuation, and risk assessment
Final Note
IPO allotment probability models are analytical tools, not predictors. Final allotment depends entirely on registrar execution under SEBI regulations, and actual results may differ materially from any estimated probability.