1 Measure KPI volatility.
Margin is rooted in how much Tesla's actual delivery count has swung quarter-over-quarter.
We use however many quarters of history are available (here 8, giving 7 returns).
Tesla has ranged from −40% (Q1 2025) to +37% (Q3 2025) in a single quarter.
2 Inflate for limited data.
Small samples underestimate true volatility — the most extreme quarters may not have happened yet.
We scale the variance up to compensate; the correction automatically shrinks as more history accumulates.
3 Model fat tails + stress-test directly.
KPI outcomes have heavier tails than a bell curve: one guidance revision can move deliveries ±30%.
We use a distribution that assigns more probability to extreme outcomes, then also
directly plug in the worst and best historical quarters plus large hypothetical shocks and take
whichever approach gives the higher loss.
4 Expected Shortfall (ES) — average loss in the worst outcomes.
We run 12,000 simulations of next-quarter deliveries and compute the average dollar loss
across the worst 2.5% of outcomes. This answers "how bad is it on average when things go badly?"
5 Margin approaches max loss as settlement nears.
A convergence term starts at 0 when the contract is listed and grows toward the maximum possible loss
as settlement nears — because the outcome is no longer uncertain, it is imminent.
Combined with the statistical risk from step 4, total margin converges to the maximum possible
loss at maturity.
| Date | Days Left | Price | Lmax Live | Conv Term | Risk Floor | Req Balance | Total Posted | Cum P&L | Acct Balance | VM Call |
|---|