sidebet.a sports-bets-as-options primer vol. 01
nba model · σ=13.5 module 01 / greeks module 02 / convexity profile

Every moneyline is a binary option.
This is the trade ticket.

DELTA dP/dD — sensitivity to score GAMMA d²P/dD² — convexity THETA dP/dt — clock decay VEGA dP/dσ — vol exposure

A moneyline bet pays $1 or $0 at terminal time. That's a binary option on the score differential, with the pregame line setting the drift and game-level scoring variance setting the volatility. Under a Brownian motion model — Stern (1994), well-replicated for NBA — the Greeks have closed-form expressions. Inputs on the left. Greeks and payoff curves on the right.

Trade ticket01
Bet side underdog
Favorite odds (American) -190
Underdog odds (American) +160
Stake (USD) $100
Game state02
Time elapsed Q1 · 0.0 / 48 min
Score differential (fav − dog) tied
Game volatility σ (pts) 13.5
Devigged · proportional
fav fair: 63.0%
dog fair: 37.0%
vig: 3.98%
implied drift μ: 4.48 pts
Position readout · underdog 03
PWIN PROB
37.0%
entry: 37.0%
ΔDELTA
-0.0280
per 1 pt of D
ΓGAMMA
+0.00069
d²WP / dD²
ΘTHETA
+0.0627
per unit t
𝒱VEGA
+0.0093
per +1.0 σ
FAIR VALUE NOW
$-3.83
stake-adjusted EV at current state
PAYOUT IF WIN
$160.00
profit on $100 stake
PREGAME EV
$-3.83
at devigged fair
Payoff surface · bet value vs. score differential slices: pregame · half · 4Q early · final min · now
Theta sheet · WP through game time, by score differential curves: D = −8, −3, 0, +3, +8
Reading the Greeks04
DELTA-0.0280
GAMMA0.00069
THETA0.0627
VEGA0.0093
Position math05
sideUNDERDOG
odds+160
stake$100.00
payout if win$160.00
implied prob (raw)38.46%
implied prob (devig)36.99%
model live WP36.99%
WP delta vs. entry+0.00 pp
fair value now$-3.83
pregame EV (post-vig)$-3.83
drift μ (model)4.484 pts
sigma σ13.5 pts
Module 02 · The Convexity Profile

Cheap dogs are options.
How much option, exactly?

Pregame, an underdog at p₀ looks like a cheap lottery ticket. The model says: across 40,000 simulated game paths per price, the expected peak in-game probability divided by entry follows a tight log-linear law — E[max pdog] / p₀ ≈ 1.05 − 0.91·ln(p₀) with R² ≈ 0.9998. The 2× hit rate is essentially constant. Cheap dogs do not double up more often; they double up by larger amounts.

Your position
underdog at p₀ = 37.0%
Expected peak
72.1%
Peak multiple
1.95×
P(reach 2×)
48%
γ / δ ratio
0.0246
Zone
DIRECTIONAL
Peak multiple · E[max pdog] / p₀ across pregame prices 40k MC paths per point · NBA σ=13.5
Fit: peakMult ≈ 1.05 − 0.91·ln(p₀), R² = 0.9998.
2× hit rate · the flat curve P(max pdog ≥ 2·p₀)
Hovers 46–50% across every entry price. The probability of doubling is constant. What varies is the absolute size of the move.
γ / δ ratio · convexity-to-direction closed-form at entry · log scale
Falls roughly 75× from p₀=3% to p₀=49%. This is the real story — convexity per unit directional exposure collapses smoothly with price.
Live simulation · your bet specifically06
Run 5,000 fresh Monte Carlo paths for p₀ = 37.0%
Click to run a fresh 5,000-path simulation. Takes about 1–2 seconds. Histogram and distribution stats appear here.

The takeaway,
stated plainly

The folk wisdom that "below 25¢ is the line" for live-betting underdog plays is partially right and partially wrong. Right: cheap underdogs are structurally more convex. Wrong: there's no threshold. The relationship is smooth. The biggest marginal gains in convexity come from going from 10% to 5%, not from crossing some arbitrary 25% line.

The non-obvious finding: the probability of any 2× repricing is approximately constant across pregame underdog prices. What changes is not how often you get a favorable move; it's how big that move is relative to what you paid. This is the precise content of "cheap optionality" — same hit rate, larger relative payoff.

Notes &
caveats

i. σ = 13.5 points-per-game is the consensus NBA score-differential volatility from Stern (1994) and replicated in Polson & Stern (2015). Other sports require a different model — NFL has discrete possessions that violate the Brownian assumption — so this v1 is NBA-only. Don't generalize.

ii. The model assumes constant within-game volatility. Real basketball has stochastic vol from foul rate, pace changes, and end-game intentional fouling. Greeks should be read as first-order approximations, not as exact prices a market maker would charge.

What this is
not

iii. This is not a betting system, an edge-finder, or financial advice. It is a visualization tool — a way of seeing the structural similarity between sports markets and derivatives markets.

iv. The empirical claim of Module 02: expected peak in-game repricing of an underdog follows a tight log-linear law — E[peak]/p₀ ≈ 1.05 − 0.91·ln(p₀), R² = 0.9998 from 40k Monte Carlo paths per price. Cheap underdogs are structurally convex. Folk wisdom had the shape right and the threshold wrong.