Armaan Sharma · Sports Analytics ↗ Open App

Portfolio Project · Sports Analytics

UCLA Men’s Basketball
NIL & Roster Model

A GM-style operating model that treats UCLA’s basketball program like a salary-capped franchise — valuing players, scoring portal decisions, and building a defensible 2026-27 roster on a $6.2M budget.

$6.2M
UCLA MBB Budget
10.7%
Valuation MAE
5,000
Monte Carlo Sims
15
Roster Slots Modeled

The Problem

UCLA can’t out-spend.
It has to out-construct.

The House v. NCAA settlement (July 2025) let schools pay athletes directly for the first time — up to $21.3M per year. UCLA committed to the full cap. But with a $6.2M MBB budget, they sit at the Power Four median, well below Kentucky’s $22M war chest.

This project builds the analytical tools a front office actually needs to compete at a structural disadvantage: a player valuation model calibrated against real contracts, a Monte Carlo portal decision engine, and a cap-constrained roster optimizer.

2025-26 MBB Roster Spend

Kentucky$22.0M
BYU$13.0M
Duke$12.0M
P4 Average$8.5M
UCLA (this model)$6.2M

UCLA’s $6.2M puts them at the Power Four median. The edge has to come from construction quality, not spending power.

What’s Inside

Five tools. One framework.

Every section solves a specific decision a real front office faces. Start with the budget, price a player, evaluate a portal target, build a roster, then pick the right contract structure.

01 · BUDGET

Budget Allocation

Walks from the $21.3M House settlement cap through UCLA’s sport-share formula to a $6.2M MBB budget. Every number is sourced — AD Jarmond’s commitment, the Wilken back-damages formula, Big Ten media revenue.

Donut chart: $21.3M split across all sports
UCLA vs. Kentucky, BYU, Duke, P4 average
Rev-share layer vs. third-party NIL clearly split

2026-27 MBB Budget

Rev-share (15% of $21.3M)$3,198,000
Third-party NIL / collective$3,000,000
Total MBB budget$6,198,000
02 · PLAYER VALUE

Player Valuation Model

Select a player’s production tier, position, eligibility, and portal demand. The model returns a dollar estimate, confidence band, and reservation price — all from four dropdowns, no spreadsheet required.

Calibrated on 11 confirmed 2025-26 deals
10.7% MAE on confirmed contracts (down from 34%)
NBA opt-out + retention premiums built in

Example — Senior PG, 5+ P4 offers

$2,970,000

Range: $2,524,500 — $3,415,500

Production tier$1,800,000
PG scarcity multiplier×1.20
Portal demand (5+ offers)×1.45
LA market bonus+$97,000
Reservation price$1,400,000
03 · PORTAL DECISIONS

Portal Decision Engine

Enter a target’s numbers and get a MATCH / COUNTER / PASS recommendation backed by 2,000 Monte Carlo simulations. The output is a probability distribution — not just a recommendation, but a confidence level.

Decision score: (wins × fit) ÷ cost − opp-cost
Real April 2026 portal players pre-loaded
Bidunga: 100% PASS · Wright: 72% COUNTER

Rob Wright III — BYU PG (real 2026 scenario)

⟳ COUNTER Market $2.0M · Our val $1.6M
MATCH
3%
COUNTER
72%
PASS
25%

Pursue — but negotiate below the $2.0M asking price

04 · PROPOSED ROSTER

Cap Sheet & Roster Builder

Fifteen slots, dollars assigned to each, rationale for every decision. Built on the Balanced Veteran archetype — three proven starters, no mega-contract, coherent system fit. Budget bar updates live as you build.

ILP optimizer finds the globally optimal assignment
Switch archetypes: Stars & Scrubs, Balanced, Flywheel
Projected record: 21-26 wins, Sweet 16 ceiling

Starting five — proposed 2026-27

1 Rob Wright III PORTAL PG $1.8M
2 Tyler Bilodeau PF $1.1M
3 Xavier Booker C $800K
4 Eric Dailey Jr. SF $750K
5 TBD portal SG PORTAL SG $550K
Top-5 payroll$5,000,000
05 · CONTRACT STRATEGY

Contract Strategy

Written strategic analysis covering the three NIL contract structures, when to use each, and a tier × duration recommendation matrix. How to gain a structural edge when you can’t outspend the competition.

Flat contracts: high-certainty seniors, win-now
Multi-year: lock in sophomores before market rises
Incentive-laden: align interests with development players

Tier × Duration matrix

1-yr flat
Multi-yr
Elite starter
✓ Best
If retained
Rotation JR/SO
Avoid
✓ Best
Dev FR
Risky
✓ Incentive

On methodology

“The methodology is more valuable than any single output — every dollar figure is tagged as confirmed, modeled, or assumed. Nothing is made up.

10.7%
Mean Absolute Error
7 confirmed deals
down from 34% in v1

Built with

The stack

🐍
Python
Core modeling engine and data pipeline
📊
Streamlit
Interactive web app, multi-page layout
📈
Plotly
All charts — donut, bar, histogram
🔢
Pandas / NumPy
Data wrangling and back-testing
🧮
PuLP (ILP)
Optimal roster slot assignment
🎲
Monte Carlo
5,000-sim portal decision confidence

Ready to explore it?

The full app is live. Price a player, run a portal scenario, build your own roster.

▶ Open the App
© 2026 Armaan Sharma Valuation MAE 10.7% · All figures are modeled estimates, not confirmed contracts