Summary¶
TOMIC teaches a static risk posture: fixed max loss per trade, fixed portfolio heat limits, delta managed back to targets as the market drifts13. The blind spot is that a static delta target is itself a regime bet: a market-neutral short-vega book is short delta to crashes through its short puts and its convexity, and it earns its keep only in calm, contango, positive-VRP conditions4. This concept's thesis: the portfolio delta target and preferred structure set should be a function of the regime label defined in Regime Definition, with the detection methods of Regime Detection Methods supplying the label with known lag.
Why Static Delta Targets Fail¶
- Same target, different risk. A book at −20 deltas of index exposure has a completely different loss distribution in a calm contango regime (mean-reverting grind) than in a stressed backwardation regime (fat left tail, gap risk). The books acknowledge "market risk" hedged with OTM puts and VIX calls1 but stop short of making the delta target itself conditional.
- Short premium is regime-conditional by construction. Selling IV above realized vol is the edge only when VRP is positive and the term structure rewards carry64. A static book keeps selling premium into backwardation.
- Transitions are the worst time to have a target at all. During detection lag (1–5 days per the methods table), the regime label is unreliable; the rational posture is to run the book flat and add no new risk until the label re-confirms.
Regime → Delta Target → Structures Mapping¶
Baseline labels from the default set (VIX band × term slope × VRP sign). Delta targets are portfolio-level deltas on a hypothetical 100Δ underlying reference, per unit of portfolio heat, not per-trade sizes. "Forbidden" means not new-position eligible; existing positions are managed per their own risk rules.
| Regime (VIX / slope / VRP) | Portfolio delta target | Preferred structures | Forbidden structures |
|---|---|---|---|
| Calm (<20) / steep contango / VRP+ | Neutral-to-short: −10 to −30 per unit | Put credit spreads, iron condors, covered calls sized to carry (naked premium excluded by the risk never-rules) | Long vega naked; heavy long-gamma bleed trades (theta cost dominates) |
| Normal (20–30) / contango or flat / VRP+ | Near-neutral: −5 to +5 | Iron condors, calendars, diagonals, ratio spreads skewed long | Large short strangles; naked puts beyond standard sizing |
| Normal / flat / VRP≈0 | Neutral: ±5 | Defined-risk spreads both directions; wait-and-see sizing | Any carry-motivated short vega (no premium to harvest) |
| Stressed (>30) / backwardation / VRP− | Small long delta or flat: 0 to +10 | Defined-risk long-vega/long-gamma tilts: put backspreads, call/put calendars sized small, VIX-call units, long puts as insurance1 | All naked short premium; short strangles/strangles in any form; short VIX ETP-style carry exposure |
| Stressed / any slope / VRP+ (IV rich but spike risk) | Flat to short, half-size | Deeper-OTM credit spreads, defined-risk only, reduced heat | Naked positions; aggressive vega shorts on size |
| Transition (label changed ≤5 days ago, any endpoint) | Flat: ±5, no new risk | Exits, rolls, hedges only1 | Everything new — detection lag means the label is provisional |
The mapping encodes the books' qualitative guidance (sell premium in calm conditions, buy units/insurance when fear is priced)24 but makes the delta target — not just the structure list — the regime-conditional variable.
Case Studies and Evidence¶
- 2018 Volmageddon (Feb 5, 2018): XIV — a short-VIX-futures carry vehicle in exactly the "calm contango" regime — lost ~96% in one session when the term structure snapped to backwardation. A regime-aware system would have been flat or defined-risk after the first backwardation confirmation; a static-delta system had no signal to change anything. See Case 2018 Volmageddon.
- March 2020 COVID crash: multi-week backwardation with VRP deeply negative; short-premium books that re-achieved their "static neutral" delta by adding short puts were adding the wrong exposure. See Case 2020 COVID Crash and Case 2020 Vol Spike.
- Baselines for the VRP input: the regime gate is only as honest as the VRP computation; use the forecast stack in Vol Forecasting Baselines, never raw IV minus trailing RV of the regime you're in.
Implementation Notes¶
- Recompute the label daily at the fixed snapshot time from Regime Definition; move the delta target with the confirmed label only, and treat transitions as a separate state.
- Rebalance toward the target gradually (e.g., 1/3 of the gap per day) so the regime rule doesn't become a new source of churn and slippage; costs are part of the EV contract.
- Log every target change with the label, the label's posterior/probability, and the snapshot data — auditable in the TOMIC infrastructure spirit1.
- Acceptance test: regime-gated backtest must beat (a) static-delta baseline and (b) the VRP/GARCH baseline, net of costs, across threshold perturbations — otherwise the rule is noise.
Links¶
- Regime Definition — state variables and default label set
- Regime Detection Methods — latency and honesty about detection delay
- Case 2018 Volmageddon
- Case 2020 COVID Crash
- Vol Forecasting Baselines
- TOMIC Risk Management — the static framework this thesis extends
References¶
- Chen, D. & Sebastian, J. (2012). The Option Trader's Hedge Fund. Wiley — TOMIC's risk framework that this concept makes regime-conditional.
- Sinclair, Euan. Positional Option Trading. Wiley, 2020 — regime-aware positioning and delta management in practice.
- CBOE (2018). "CBOE Statement Regarding the Termination of VelocityShares Daily Inverse VIX Short-Term ETN (XIV)." Feb 2018, https://www.cboe.com/about/press/released-products/volmageddon-etp-note/
- Bollerslev, Tauchen & Zhou (2009). "Expected Stock Returns and Variance Risk Premia." Review of Financial Studies 22(11) — VRP sign as a conditioning variable.
Amendment Log¶
| Version | Amendment | Status |
|---|---|---|
| v0 | initial mapping | in force |
| v1 (proposed) | VRP size threshold — premium-selling eligibility requires VRP_size ≥ θ (calibrated); activation gated on fitted GARCH/HAR forecast + calibration + human approval | pending |
-
Market-risk hedges (OTM puts, VIX calls) and heat limits in TOMIC risk management. ↩↩↩↩↩
-
TOMIC's condition-based structure selection in Strategies. ↩
-
Portfolio-level greeks management in TOMIC Greeks. ↩
-
Conditions for selling volatility and the carry risk of short VIX positions in Volatility Selling. ↩↩↩
-
Delta as a chosen exposure with gamma/theta costs in Delta-Neutral Trading. ↩
-
Structure behavior under vol moves (backspreads, ratio spreads) in Spreads. ↩