Will WTI oil on Nov 3, 2026 close above $95.99?
6 seats worked this question alone; the panel agreed within a few points, so no debate was run. Code ran every test and computed the number. How the desk works
Judged by Kalshi KXWTI-26NOV0314-T95.99 settlement at 2026-11-03: If the daily settlement price for WTI crude oil(December 2026 contract) on November 03, 2026 is above 95.99 USD/Bbl, then the market resolves to Yes.
The dispute that matteredNot a disagreement in point estimates (all cluster 0.52-0.56) but in interpretation of the -7.64% shock from 2026-11-02: most analysts treat it as a draw from the existing unconditional return distribution (implying ~57-60% P(YES) before modest downward adjustment), while the Causal Skeptic argues the shock likely comes from an unidentified regime (news/liquidation/data artifact) for which the historical distribution's skew/kurtosis parameters may not apply, and shows the one correlation structure available in the data is unstable across regimes. I resolve this by treating the unconditional-distribution estimate (~57-60%) as the correct base rate but applying the Causal Skeptic's regime-instability finding as the dominant source of downward adjustment, landing below the panel's naive distributional math at 0.53 rather than averaging toward 0.54-0.56.
Synthesis
- WTI Dec26 settled at $96.46 on 2026-11-02, which is $0.47 (0.49%) above the $95.99 strike.
- The 2026-11-02 daily return was -7.64%, the most extreme single observation in the ~61-179 day sample window used across analysts (sample minimum).
- Historical daily log-return series exhibits mean near 0 to +0.36% (varies by analyst's window), sd ~2.7%-4.1% (varies by analyst's window and return definition), negative skew (~-0.73 cited by one analyst), and near-zero return autocorrelation (0.058).
- Normal/lognormal approximations to the daily return distribution imply roughly 57%-65% probability of staying above the strike before discretionary adjustment (varies by analyst depending on which window/sd estimate used).
- Bootstrap resampling (10,000x cited by one analyst) of 1-day returns places the -0.49% breach threshold near the 40th-41st percentile of simulated returns, implying ~59-60% unconditional P(YES).
- Regime-split correlation test (Causal Skeptic) shows the only cross-series relationship available in the data flips sign across regimes and is not robust to single-point removal, undermining confidence in any conditional (vs. unconditional) estimate.
- The question is effectively a near-at-the-money 1-day volatility threshold bet, not a directional fundamentals call, since the cushion (0.49%) is small relative to typical daily moves.
- No statistically significant drift exists in the return series (bootstrap CI on mean return includes zero), so no directional edge can be assumed from historical mean return alone.
- P(WTI settles above $95.99 on 2026-11-03) = 0.53, with 90% CI [0.46, 0.60].
- The causal origin of the -7.64% move on 11/02 (news, liquidation cascade, contract-roll artifact, or data error) is unknown and not addressed by any analyst's evidence — this materially affects whether continuation or mean-reversion is more likely tomorrow.
- Whether the historical unconditional return distribution (built from a calmer regime) is the right generator to sample from, given the panel's own regime-instability finding.
- No inventory, curve (contango/backwardation), positioning/COT, or news-flow data was available to any analyst despite several having commodities/positioning in their title — this is a real data gap, not just conservative framing.
| Commodities Analyst | 0.20 | Highest domain relevance (1.0 commodities) and honest about missing fundamental data (inventory/curve); evidence is the primary price/vol series, treated as foundational — but its conclusion is one of six using the same series, so weight capped despite relevance. |
| Weather Analyst | 0.08 | Calibration stats are in line with peers, but domain relevance to oil pricing is secondary (self-rated 0.7, role mismatch); evidence set (d1, bootstrap) duplicates commodities/markets analysts almost exactly — downweighted for redundancy, not competence. |
| Markets Analyst | 0.17 | Strong relevance to price-threshold/return-distribution reasoning, best-in-panel accuracy/brier tradeoff among non-skeptic analysts; evidence again largely the same return series — moderate independent value from framing but not a new data source. |
| Quant Researcher | 0.15 | Adds explicit bootstrap CI on drift (includes zero) — a slightly independent statistical check, decent calibration; domain relevance diluted (flat 0.7-0.8 across all domains suggests generic model application rather than oil-specific insight). |
| Positioning Analyst | 0.10 | Lowest confidence (0.35) of the panel, weakest calibration (brier 0.281, worst in group); evidence is the same return series with no positioning/flow data actually used despite the role name — discounted for lack of role-specific independent evidence. |
| Causal Skeptic | 0.30 | Best calibration (brier 0.268, highest in panel) and the only analyst contributing genuinely independent evidence: regime-split correlation sign flip (s5) and non-robustness to single-point removal (s6), directly questioning whether the shock is a stable, model-able event. This is the one line of evidence not duplicated elsewhere and most relevant to resolving the top disagreement, so it gets the highest weight despite moderate confidence. |
- If WTI Dec26 settles below $95.99 on 2026-11-03, this prediction (P(YES)=0.53) is falsified as a directional call, though a well-calibrated 53% forecast is not falsified by a single binary outcome — track realized frequency across repeated near-ATM 1-day threshold calls to assess calibration.
- If next-day volume/implied-vol data (unavailable as of 2026-11-03 12:30 UTC) shows the 11/02 move was a data artifact or thin-liquidity air-pocket rather than a fundamentals-driven shock, the unconditional-distribution-based estimate should be revised toward the higher end of the range (~57-60%) rather than the skeptic-discounted 0.53.
The seats
Commodities Analyst52%
57% Unconditional probability that a one-day WTI log return lands above the ~-0.49% threshold needed to stay above $95.99 from a $96.46 base, using the empirical daily-return distribution (mean 0.0022, sd 0.0413, n=173) under a normal approximation.
The Nov-2 drop of 7.6% in one session likely reflects a specific bearish catalyst (e.g., a large crude/product inventory build, a demand-growth downgrade, or OPEC+ supply signal) that this dataset doesn't name. If that catalyst is still unfolding or data releases continue in the same direction on Nov 3 (e.g., a weekly EIA report or confirmed oversupply trend), the move could extend well past the small 0.49% cushion, making the 'low autocorrelation, independent draw' assumption wrong and the true YES probability meaningfully lower than 0.52 — possibly closer to 0.35-0.40 if the selloff has further to run intraday.
- Last price 96.46 on 2026-11-02 sits only ~0.49% above the 95.99 strike, making this essentially a coin-flip-range question dominated by one day of volatility, not a directional macro call.
- Empirical daily return distribution (sd=4.13%, skew=-0.73) implies the 0.49% cushion is small relative to typical daily moves, and tail risk is skewed to the downside.
If Kalshi KXWTI-26NOV0314-T95.99 settles with the Dec26 WTI contract at or below $95.99 on 2026-11-03, this forecast (P≈0.52 YES) is falsified; conversely a settlement materially above $97 would suggest the skew-adjustment was excessive.
- No inventory, EIA stock, or futures-curve (contango/backwardation) data was supplied, so this call relies only on price and return series, not a physical balance model.
- The last observed WTI Dec26 contract price (96.46 on 2026-11-02) is a reasonable proxy for the opening/reference level on 2026-11-03.
- Daily log-return distribution (d2/s2) is approximately stable enough to extrapolate one more day ahead despite the recent volatility spike.
- Settlement on Nov 3 behaves like a random draw from the recent empirical return distribution, not a scheduled-news day with asymmetric risk (no EIA/OPEC event flagged in data).
debate skipped: panel unanimous within 0.04
Weather Analyst56%
Loading…
Public Markets Analyst54%
Loading…
Quantitative Researcher52%
Loading…
Market Positioning Analyst56%
Loading…
Causal Skeptic55%
Loading…
Claims and evidence
The debate
| Commod | Weathe | Public | Quanti | Market | Causal | |
|---|---|---|---|---|---|---|
| Commodities | ||||||
| Weather | ||||||
| Public Marke | ||||||
| Quantitative | ||||||
| Market Posit | ||||||
| Causal Skept |
With only price/return series available (no inventory or curve data), WTI Dec26 at $96.46 sits just 0.49% above the $95.99 strike. Given a daily return sd of ~4.1%, this is close to a coin flip; a plain normal model gives ~57% YES, but the empirical negative skew (-0.73) and the preceding day's large 7.6% drop (even though returns show low serial correlation) justify trimming the estimate to ~52% rather than asserting a physical-market-driven directional view I cannot support with the data on hand.
WTI (Dec26) settled at $96.46 on Nov 2, just $0.47 (0.49%) above the $95.99 strike, after a sharp -7.64% single-day drop. With one trading day to settlement, the question reduces to: does WTI fall by more than 0.49% on Nov 3? Using the empirical daily-return distribution (n=171), this threshold sits near the 41st percentile of the return distribution, implying roughly 59% probability of staying above, which I discount modestly for negative skew and fat tails (elevated continuation risk after a large down-move) to ~56%.
WTI Dec26 closed Nov 2 at 96.46, only 0.49 percent above the 95.99 settlement threshold, right after a historic one day drop of -7.64 percent, the largest move in the 61 day return sample by far. The question reduces to whether the contract avoids a further ~0.5 percent decline into Nov 3 settlement. Using the observed daily log return distribution (mean 0.36 percent, sd 2.68 percent), the breakeven return of -0.49 percent sits only about 0.2 to 0.3 standard deviations below center, implying roughly 57 to 62 percent probability of staying above threshold under a normal approximation. Negative skew and the just-occurred outlier argue for somewhat elevated downside tail risk, while the bootstrap CI for mean return spans zero, so positive drift cannot be relied upon.
Base rate under an unconditional-normal approximation to the historical daily-return distribution (n=179, mean≈0.0024, sd≈0.0407) puts P(YES) at ≈0.57, since the strike (95.99) sits only ~0.49% below the last observed settlement (96.46). Adjusting downward for (a) statistically non-significant drift (bootstrap CI includes 0), (b) negative skew and excess kurtosis in the return series implying fatter left-tail risk, and (c) the fact that the most recent realized return was a -7.6% shock (a likely volatility-clustering event not reflected in the full-sample sd), I revise the probability down modestly to 0.52 — essentially a coin flip with a very slight lean toward YES because the strike is below, not above, last price.
The strike ($95.99) sits only ~0.5% below the last observed settlement ($96.46 on 11/02), right after a sharp one-day drop of -7.64%. This places the question almost exactly at-the-money for the Nov 3 settlement, with the outcome dominated by one more day of return noise rather than durable fundamentals. A lognormal approximation using the historical daily-return distribution (mean 0.36%, sd 2.68%) implies roughly 60-65% probability the Nov 3 settlement stays above -0.49% return, i.e. above 95.99. I discount slightly toward 0.56 because the prior day's move was a ~2.9 sd shock, and such shocks often see volatility clustering/continuation that an unconditional model understates; return autocorrelation is negligible (0.058) so there is no statistical basis to lean on reversion either. Net: a modest lean to YES, not a high-conviction call; most of the apparent edge is just the distributional math already implied by proximity to strike, with no clean mispricing to exploit.
The question reduces to a near-coin-flip on a one-day return threshold (-0.49% from the last observed 96.46 to breach 95.99), set immediately after the single largest one-day decline in the entire 61-obs return sample (-7.64% on 11/02, which is literally the sample minimum per s2). None of the supplied statistical evidence establishes a causal, stable mechanism that would let us confidently predict whether that shock reverses, continues, or is noise — regime split (s5) shows the key correlation flips sign across regimes, and influence testing (s6) shows the correlation is not robust to single-point removal. Treating the empirical return distribution (s2/s3) as the best available unconditional estimator, the fraction of 1-day returns historically falling below the -0.49% breach threshold is roughly 40%, implying ~60% unconditional probability of staying above 95.99 — but I discount this because the triggering event (a 7.6% crash) is itself outside the normal regime that generated that empirical distribution, and the model offers no causal explanation for the crash (news? liquidation? data artifact?) that would let us condition correctly. I settle below the naive 60% to account for this unresolved uncertainty and the asymmetric tail risk of continued momentum in a already-stressed/falling market.
debate skipped: panel unanimous within 0.04
debate skipped: panel unanimous within 0.04
debate skipped: panel unanimous within 0.04
debate skipped: panel unanimous within 0.04
debate skipped: panel unanimous within 0.04
debate skipped: panel unanimous within 0.04
All six analysts are, in substance, reporting the same underlying evidence through the same lens: last price ($96.46) is 0.49% above strike, and the historical daily-return distribution (sd roughly 2.7-4.1% depending on window) implies a near-coin-flip with a modest lean toward YES (57-65% before downward adjustment for the just-occurred -7.64% outlier and negative skew). This is one independent line of evidence, not six — the panel converged not because of genuine corroboration across different data sources, but because everyone ran essentially the same calculation on the same price series. The only materially independent contribution is the Causal Skeptic's regime-split/robustness test (s5, s6), which shows that whatever cross-series relationship exists in the data is unstable and not trustworthy for conditioning — this is the strongest reason to discount the naive 57-60% distributional estimate down toward 0.53 rather than up toward 0.56. No analyst had access to inventory, futures curve, positioning, or news data, which is a real gap for a commodities question and caps my confidence (data_quality 0.45). Statistical quality is moderate (0.55): the bootstrap and quantile methods are sound, but applied to a short, overlapping, non-independent set of windows. Novelty is low (0.15) because the entire panel's edge reduces to 'proximity to strike plus vol scaling,' a textbook distributional calculation with no new informational content. Crowding is high (0.85): the deterministic pool (0.54) and my consensus (0.53) sit on top of six analysts who are really expressing one view with cosmetic differences — there is no genuine edge here beyond what a single competent quant could compute from the price series alone. TRADE DECISION: Not tradeable as a high-conviction position. The estimate (0.53) is barely above a coin flip and the edge over 50/50 is within the noise of the calibration uncertainty of every analyst in the panel (brier scores all ~0.27-0.28, i.e., historically close to uninformative on this type of near-ATM 1-day threshold question). If a market is pricing this meaningfully away from 0.50-0.55 (e.g., below 0.45 or above 0.65), that would be the trigger — not the probability estimate itself, which lacks sufficient edge or independent evidence density to size a position confidently. Recommended action: no directional position; if forced to express a view, a small, capped-risk exposure sized to reflect the thin ~3pp edge over 50/50, explicitly not scaled to the full 0.53 estimate given the duplicated-evidence problem and unresolved causal uncertainty around the preceding shock.
Assumptions
- holdingNo inventory, EIA stock, or futures-curve (contango/backwardation) data was supplied, so this call relies only on price and return series, not a physical balance model.
- holdingThe last observed WTI Dec26 contract price (96.46 on 2026-11-02) is a reasonable proxy for the opening/reference level on 2026-11-03.
- holdingDaily log-return distribution (d2/s2) is approximately stable enough to extrapolate one more day ahead despite the recent volatility spike.
- holdingSettlement on Nov 3 behaves like a random draw from the recent empirical return distribution, not a scheduled-news day with asymmetric risk (no EIA/OPEC event flagged in data).
- holdingKalshi settlement tracks the CME WTI Dec26 futures daily settle, not an intraday print.
- holdingThe Nov 2 close of $96.46 is the most recent observed settlement prior to the Nov 3 session.
- holdingNo material scheduled news (EIA inventory, OPEC+ meeting outcome) is already known to be pending for Nov 3 that would shift the distribution materially (untested — I have no calendar data).
- holdingReturn distribution (d2/s2/s3) is representative of near-term conditional volatility; no regime break is explicitly modeled.
- holding171 observations of daily returns are i.i.d.-ish for bootstrap purposes despite autocorr diagnostics being weak (lag-1 autocorr ~0.007, not significant).
- holdingDec26 WTI contract tracks the price level series in d1 without major contract roll distortion between Nov 2 and Nov 3.
- holdingThe 61-day return sample is representative of near-term volatility regime despite containing the outlier crash.
- holdingNo material new fundamental catalyst materializes between Nov 2 close and Nov 3 settlement beyond what is already reflected in the -7.64 percent move.
- holdingSettlement price approximates final traded price near session close.
- holdingSettlement price for the Dec 2026 WTI contract on Nov 3 is close to last observed price 96.46 (Nov 2) absorbed into d1 series; threshold 95.99 is 0.49% below last observed level.
- holdingDaily log-return series d2/d3 (n=179/344) is representative of the relevant 1-day-ahead return distribution for this contract.
- holdingNo material scheduled macro/OPEC event is known to me beyond the data provided (I have no post-as_of information).
- holdingSettlement-to-settlement dynamics are assumed well-approximated by the historical unconditional return distribution; no explicit GARCH/vol-clustering model was fit, so I only approximate the effect of the clustering.
- holdingNov 3 settlement is well-approximated by last close (96.46) plus one more day's return drawn from a distribution similar to the trailing 61-day empirical series.
- holdingNo major scheduled catalyst between 11/02 close and 11/03 settlement beyond what already drove the -7.64% move.
- holdingUsing unconditional (not conditional/GARCH) return sd likely understates near-term vol given the shock just observed.
- holdingd1 is the WTI daily settlement/close series and its last value (96.46 on 11/02) is the most recent pre-settlement price anchor for 11/03.
- holdingd2 is the corresponding daily log/simple return series aligned with d1.
- holdingNo new information between 11/02 close and 11/03 settlement is available to this analysis (as_of is 11/03 12:30 UTC, before settlement).
- holdingThe bootstrap distribution in s3 is a reasonable proxy for the conditional distribution of the next-day return despite being estimated unconditionally over 61 days spanning mixed regimes.
- holdingd3/d4/d5 (volume, implied vol or OI, and a second return-like series) are not directly used in my estimate because no test ties them causally to next-day settlement; s5 and s6 specifically undermine using them as predictors.
Simulated trade
The portfolio manager proposed no paper trade. The ~3pp consensus edge over 50/50 derives almost entirely from one duplicated calculation (price-proximity plus vol-scaling) run six times on the same series, not independent corroboration. The only independent input argues for less confidence, not more. This edge sits within the panel's own historical calibration noise and below realistic transaction costs and crowding drag for a WTI expression. No fundamental data (inventories, curve, positioning) was used. Correct action: wait for an external price dislocation (market <0.45 or >0.65) rather than trade the 0.53 estimate itself.
Forecasts
| Seat | Stage | P | Base | Resolves | Outcome | Brier |
|---|---|---|---|---|---|---|
| Committee | final | 54% | — | 2026-11-03 | Yes | 0.210 |
| Commodities Analyst | prior | 80% | 57% | 2026-11-03 | Yes | 0.040 |
| Commodities Analyst | independent | 52% | 57% | 2026-11-03 | Yes | 0.230 |
| Commodities Analyst | revised | 52% | 57% | 2026-11-03 | Yes | 0.230 |
| Weather Analyst | prior | 80% | 50% | 2026-11-03 | Yes | 0.040 |
| Weather Analyst | independent | 56% | 50% | 2026-11-03 | Yes | 0.194 |
| Weather Analyst | revised | 56% | 50% | 2026-11-03 | Yes | 0.194 |
| Public Markets Analyst | prior | 90% | 58% | 2026-11-03 | Yes | 0.010 |
| Public Markets Analyst | independent | 54% | 58% | 2026-11-03 | Yes | 0.212 |
| Public Markets Analyst | revised | 54% | 58% | 2026-11-03 | Yes | 0.212 |
| Quantitative Researcher | prior | 86% | 57% | 2026-11-03 | Yes | 0.020 |
| Quantitative Researcher | independent | 52% | 57% | 2026-11-03 | Yes | 0.230 |
| Quantitative Researcher | revised | 52% | 57% | 2026-11-03 | Yes | 0.230 |
| Market Positioning Analyst | prior | 88% | 62% | 2026-11-03 | Yes | 0.014 |
| Market Positioning Analyst | independent | 56% | 62% | 2026-11-03 | Yes | 0.194 |
| Market Positioning Analyst | revised | 56% | 62% | 2026-11-03 | Yes | 0.194 |
| Causal Skeptic | prior | 60% | 60% | 2026-11-03 | Yes | 0.160 |
| Causal Skeptic | independent | 55% | 60% | 2026-11-03 | Yes | 0.202 |
| Causal Skeptic | revised | 55% | 60% | 2026-11-03 | Yes | 0.202 |