Institutional-style technical report generated from the live DEV2 quant stack: Markov, Monte Carlo, predictive diagnostics, and decision governance.
The engine recommends a hedge ratio of 0.70 at spot 53.32. Current regime assessment is STRESS with dominant transition toward TRANSITION with flip probability 0.76. Monte Carlo uncertainty is 1.83x spot, which places the distribution in a EXTREME uncertainty bucket. Predictive diagnostics classify the setup as NO_EDGE / DO_NOT_TRADE. Directional bias is NEUTRAL. Composite decision confidence is 0.1839, which should be read as governance confidence rather than directional conviction.
The selected posture is PARTIAL_HEDGE with recommended hedge ratio 0.70. The top Monte Carlo candidate scored 477.50 at ratio 0.70 under profile BALANCED. Ratio stability is 0.80 with confidence band 0.50–0.70. Decision separation versus the next-best candidate is 110.55. Predictive diagnostics do not confirm a tradeable edge, so the hedge recommendation should be interpreted as risk-control governance rather than as an aggressive directional call. Transition risk is elevated, which argues against low-hedge complacency even if asymmetry is mildly constructive.
| Market | TTF |
| Current Price | 53.32 |
| Bias | NEUTRAL |
| Regime | STRESS |
| Flip Probability | 0.76 |
| Persistence | 0.24 |
| Half-life | 0.48 |
| Dominant Next Regime | TRANSITION |
| Volatility | 0.1110 |
| Price Percentile 20d | 0.88 |
| Transition Risk | HIGH |
| P10 | 7.16 |
| P50 | 29.54 |
| P90 | 104.47 |
| Distribution Width | 97.31 |
| Uncertainty / Spot | 1.83x |
| Upside Potential | 51.15 |
| Downside Risk | 46.16 |
| Asymmetry | 4.99 |
| Skew | BULLISH |
The Monte Carlo terminal distribution spans 7.16 to 104.47, with median 29.54 against current spot 53.32. Absolute distribution width is 97.31, which is materially wide for a 20-day horizon. Relative width equals 1.83x spot, signalling a regime where scenario dispersion dominates simple point forecasting. Simulation quality is labelled WEAK (score 0.30), based on 800 paths. Path agreement for the selected hedge region is 0.6667.
| Classification | NO_EDGE |
| Actionability | DO_NOT_TRADE |
| Direction Consensus | NEUTRAL |
| Conflict Level | HIGH |
| Conflict Score | 0.70 |
| Edge Strength | 0.0755 |
| Final Edge Score | 0.0000 |
| Confidence | 0.0000 |
Predictive diagnostics classify the setup as NO_EDGE with actionability DO_NOT_TRADE. Consensus direction is NEUTRAL. Signal conflict is HIGH (0.70). Final edge score is 0.0000. Predictive diagnostics indicate NO_EDGE with DO_NOT_TRADE posture. Consensus direction is NEUTRAL, signal alignment is 0.00, conflict is HIGH, and risk penalty is 0.33. Regime context is RANGE with dominant next regime RANGE.
| Storage Fill % | 29.01 |
| LNG Imports | 4,417.30 |
| Curve M+1 | 53.32 |
| Curve M+2 | 53.55 |
| Curve Q+1 | 56.90 |
| Curve Y+1 | 43.62 |
| Data Quality Score | 0.73 |
| Lag Days | 11 |
Storage fill is 29.01. LNG imports are 4,417.30. Fundamental data lag is 11 days. Snapshot data-quality score is 0.73. These fundamentals should be read as contextual constraints on the technical output, not as a substitute for the quant stack.
| Simulation Quality Label | WEAK |
| Simulation Quality Score | 0.30 |
| Tail Width | 1.83 |
| Path Count | 800 |
| Regime Coherence | 0.74 |
| Ratio Stability | 0.80 |
| Path Agreement | 0.6667 |
| Confidence Band | 0.50–0.70 |
This table ranks alternative hedge ratios using the Monte Carlo engine's internal ratio-evaluation framework.
| Hedge Ratio | Profile | Expected PnL | P10 PnL | P50 PnL | P90 PnL | Breach Prob. | Score |
|---|---|---|---|---|---|---|---|
| 0.70 | BALANCED | 0.89 | -21.37 | 8.49 | 17.20 | 100.0% | 477.50 |
| 0.50 | BALANCED | 2.08 | -29.82 | 12.82 | 25.52 | 100.0% | 366.95 |
| 0.60 | BALANCED | 1.48 | -25.60 | 10.66 | 21.38 | 100.0% | 344.06 |
| 0.40 | LIGHT | 2.67 | -34.08 | 14.99 | 29.66 | 100.0% | 242.81 |
| 0.80 | DEFENSIVE | 0.29 | -17.14 | 6.27 | 13.10 | 100.0% | 223.99 |
| 0.25 | LIGHT | 3.57 | -40.48 | 18.28 | 35.85 | 100.0% | 9.97 |
| 0.90 | DEFENSIVE | -0.31 | -12.91 | 4.04 | 8.91 | 100.0% | -67.60 |
| 0.10 | OPTIONAL | 4.46 | -46.88 | 21.58 | 42.04 | 100.0% | -289.93 |
| 1.00 | MAX_DEFENSE | -0.90 | -8.63 | 1.77 | 4.82 | 100.0% | -565.81 |
| 0.00 | OPTIONAL | 5.06 | -51.15 | 23.78 | 46.16 | 77.6% | -647.34 |
The engine is not presenting a clean directional trade. It is presenting a risk-governance decision under high uncertainty and elevated transition instability. In that context, a partial hedge is a control action, not a bold alpha expression.
The hedge is supported by wide scenario dispersion, elevated flip/transition stress, and a Monte Carlo ranking that places the selected ratio above lower-protection alternatives on score and drawdown control.
Conviction is capped by weak predictive edge, high signal conflict, and a distribution that is too wide to justify aggressive directional confidence. The system is effectively saying: protect first, speculate later.
The causal chain begins with the STRESS regime reading. Markov assigns a flip probability of 0.76, which maps to a VERY HIGH transition-risk state. That means the engine does not trust the market to remain calmly inside one stable regime for long. Persistence is 0.24, which is a WEAK persistence reading. In plain language: the current regime does not have much staying power, and the most likely next state is TRANSITION. Once that unstable regime picture is passed into Monte Carlo, the simulation responds with a distribution width of 97.31 (1.83x spot). This is the engine's way of saying that path dispersion matters more than any single point forecast. Because dispersion is wide, the optimizer shifts from pure upside-seeking to protection-seeking. That is why the chosen hedge ratio is 0.70: it represents defensive partial protection, not a blind bearish bet. Asymmetry is 4.99. Positive asymmetry means upside scenarios are somewhat larger than downside scenarios, but the engine still prefers protection because transition instability and scenario width dominate the decision. Predictive diagnostics then act as a governor, not as the primary driver. Here they classify the setup as NO_EDGE / DO_NOT_TRADE, which prevents the system from framing this as a high-conviction alpha trade. So the full logic chain is: unstable regime → wide scenario distribution → protection-first optimization → PARTIAL_HEDGE.
Risk-wise, the report should be read as a control memo. Spot is 53.32, while the simulated 20-day range spans from 7.16 to 104.47. That spread is too large to justify overconfident directional positioning. Decision confidence is 0.1839. This is not a probability of being right in the ordinary sense; it is a governance-quality score telling you how robust the chosen posture looks after multiple engines are reconciled. Predictive conflict score is 0.70. Higher conflict means sub-signals disagree, so the report should be used with more humility and less directional aggression. Flip probability at 0.76 means regime change risk is materially present. That raises the penalty for low hedging because the market may leave today's state faster than a static view assumes. A good human interpretation is: preserve survivability first, then ask whether there is enough evidence left for a directional overlay. In this run, the answer to the second question is clearly limited.
The decision would become more constructive if three things improved together, not separately. First, Markov instability would need to cool: flip probability would need to fall meaningfully below the current 0.76, while persistence rises above the current 0.24. Second, Monte Carlo dispersion would need to tighten: uncertainty would need to compress below the current 1.83x spot, so the forecast range stops dominating the decision. Third, predictive diagnostics would need to move away from NO_EDGE and show lower conflict than the current 0.70. If those conditions improved together, the system could rationally migrate from PARTIAL_HEDGE toward a lighter hedge or even a hold posture. The reverse is also true: if flip risk rises further, predictive conflict remains high, or tail width expands, the system would have grounds to become even more defensive.
This flow compresses the full logic chain from market state to final action.
This section explains where the main outputs come from, what they are for, how a human reader should interpret them in practice, and what each one means for action.
| Output | Current Value | Comes From | What It Is | Why It Exists | How To Interpret It | So What? |
|---|---|---|---|---|---|---|
| Spot | 53.32 | Snapshot / market state | Current anchor price used by the report. | Reference point for all scenario comparisons. | Use it to judge whether P50 sits above or below the market and how large width is relative to the present price. | Anchor everything to this level: compare P50 and scenario width to spot before drawing any directional conclusion. |
| Regime | STRESS | Markov regime layer | Current market state label inferred from recent behavior. | Sets the behavioral context for simulation and governance. | Read it as the market's current 'mode' rather than a price target. | The market is behaving like STRESS right now, so behavior matters more than any single price print. |
| Flip Probability | 0.76 | Markov transition matrix | Probability that the current regime does not hold and may transition. | Measures regime instability. | Higher = less trust in a static market story; it usually supports more defensive decisions. | High flip means low trust in a static market story; it usually argues against complacent low-hedge positioning. |
| Persistence | 0.24 | Markov self-transition probability | Probability that the current regime continues. | Counterweight to flip probability. | Higher = regime is stickier and easier to trade directionally; lower = fragile regime. | Low persistence means the current state is fragile; high persistence would make directional conviction easier to justify. |
| Half-life | 0.48 | Markov persistence math | Approximate time for current regime influence to decay materially. | Adds time interpretation to persistence. | Short half-life means the regime can change quickly; long half-life means more staying power. | This adds time meaning: short half-life says today's regime can decay fast, so decisions should stay adaptive. |
| Dominant Next Regime | TRANSITION | Markov transition ranking | Most likely destination if the market transitions away from the current regime. | Shows where instability is pointing. | Use it as scenario direction, not certainty. | Treat this as the most likely destination of instability, not as a certain forecast. |
| Recommended Hedge | 0.70 | Monte Carlo ratio optimization + governance | Fraction of exposure the engine suggests protecting. | Primary control output for risk action. | Higher = more protection; lower = more openness to market moves. | This is the practical control action: higher means protect more, lower means leave more exposure open. |
| Decision | PARTIAL_HEDGE | Decision governance layer | Human-readable action label attached to the hedge ratio. | Summarizes posture for execution/governance. | Interpret together with hedge ratio; the label alone is never enough. | Read the action label together with the hedge ratio; the label alone is too coarse for real risk decisions. |
| Confidence | 0.1839 | Cross-engine governance aggregation | How coherent and defensible the chosen posture is after reconciling engines. | Measures robustness of the recommendation. | Not a direct win probability. Low confidence means treat the action as cautious governance. | This is governance robustness, not a plain probability of being right. Low values mean protect with humility. |
| P10 / P50 / P90 | 7.16 / 29.54 / 104.47 | Monte Carlo terminal distribution | Key percentiles of simulated terminal outcomes. | Describe downside, median, and upside scenario landmarks. | P10 is adverse, P50 is central, P90 is optimistic; compare all three to spot. | Read these as downside / central / upside landmarks. When they are very spread out, scenario thinking beats point forecasting. |
| Distribution Width | 97.31 | Monte Carlo percentile spread | Distance between the upper and lower important scenario bounds. | Measures absolute scenario dispersion. | Wide width means point forecasts are unreliable and risk control matters more. | Wide width means dispersion dominates the setup, so risk control matters more than narrative conviction. |
| Uncertainty / Spot | 1.83x | Derived from width vs spot | Distribution width scaled by current price. | Normalizes uncertainty so different price regimes are comparable. | Above 1.0x is already substantial; very high values argue for humility. | When width is large relative to spot, forecasting precision is weak and defensive posture becomes easier to justify. |
| Asymmetry | 4.99 | Derived from upside vs downside | Difference between upside potential and downside risk. | Shows whether the distribution leans favorable or unfavorable. | Positive does not automatically mean bullish action if instability is high. | Positive asymmetry can be constructive, but it should not overrule instability and wide tails on its own. |
| Predictive Classification | NO_EDGE | Predictive diagnostics engine | High-level edge verdict. | Acts as a governor on directional conviction. | NO_EDGE means the system does not see enough aligned evidence for a strong directional trade. | If this says NO_EDGE, the system is explicitly refusing to bless a strong directional trade. |
| Actionability | DO_NOT_TRADE | Predictive diagnostics engine | Recommended trading posture from predictive layer. | Filters whether a signal is operationally usable. | DO_NOT_TRADE means even interesting signals are not clean enough for aggressive action. | This tells you whether an interesting signal is actually clean enough to use in practice. |
| Conflict Score | 0.70 | Predictive diagnostics signal reconciliation | Level of disagreement among predictive sub-signals. | Measures internal disagreement. | Higher conflict reduces trust in directional overlays. | High conflict means sub-signals disagree, which reduces trust in aggressive directional overlays. |
| Data Quality Score | 0.73 | Snapshot quality controls | Quality/readiness score of the input snapshot. | Helps judge how much trust to place in the run itself. | Lower quality means outputs may still be directionally useful, but should be handled more cautiously. | This is the trust score of the run itself; lower quality means the outputs deserve more caution. |
| Regime | STRESS |
| Flip Probability | 0.7619 |
| Persistence | 0.2381 |
| Half-life | 0.48 |
| Dominant Next Regime | TRANSITION |
| P10 | 7.16 |
| P50 | 29.54 |
| P90 | 104.47 |
| Distribution Width | 97.31 |
| Hedge Ratio | 0.70 |
| Decision | PARTIAL_HEDGE |
| Confidence | 0.1839 |