Will WTI oil on Nov 10, 2026 close above $89.99?
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-26NOV1014-T89.99 settlement at 2026-11-10: If the daily settlement price for WTI crude oil(December 2026 contract) on November 10, 2026 is above 89.99 USD/Bbl, then the market resolves to Yes.
The dispute that matteredAll six analysts land in a tight 0.56-0.60 band, so the apparent 'disagreement' is not directional but methodological: (1) normal-approximation diffusion (A, E) implies ~0.59 using 2-day vol scaling, vs (2) bootstrapped empirical return distributions (B, D) implying 0.61-0.63 with explicit left-skew/fat-tail discounting, vs (3) pure skeptical random-walk framing (F) that refuses to apply any drift or skew adjustment and lands at 0.56. The practical disagreement is whether the recent 2-day recovery (88.12 -> 90.52 -> 90.99) constitutes a real stabilization signal (B, D lean on this) or is noise indistinguishable from the sharp prior 104->88 drop (F, E explicitly reject this, citing the failed OOS retest of the mean-reversion insight: hit rate 45.5%, p=0.806). I resolve this by treating the bootstrap median drift (+0.8-0.9%) as unreliable given it is sourced from the same recent-rebound data the skeptics discount, and instead anchor primarily on the model-free vol-scaling base rate (~0.57-0.59), nudging slightly above the pure-noise 0.56 floor only because price is already above strike with no demonstrated drift either way -- not because the rebound is informative.
Synthesis
- WTI (D1 proxy) closed at 90.99 on 2026-11-06, which is 1.10% above the 89.99 strike.
- Prior 6 closes show a sharp drop from 104.1 to 88.12 (Nov4 shock low) followed by two consecutive up days to 90.52 and 90.99.
- Trailing 20-day daily log-return standard deviation is approximately 2.9%-3.5% depending on the analyst's window/estimator.
- The library's mean-reversion insight fails its own walk-forward out-of-sample retest today: 45.5% hit rate, p=0.806 (i1_wf); regression beta=-0.189, CI includes 0, p=0.273, with Durbin-Watson=0.20 flagging severe residual autocorrelation and Cook's D=0.131 flagging an outlier (i1_reg).
- Normal-approximation diffusion over the remaining ~2 trading sessions to Nov10, using realized vol ~3.5%/day, implies P(stay above strike) approximately 0.59 (Analyst A/E logic).
- 5000x bootstrap resample of 4-day forward returns (B, D) gives median +0.83-0.88%, 25th percentile around -3.0% to -3.6%, implying roughly 61-63% probability of staying above the strike by interpolation.
- Calibration-weighted log-odds pooling across all six post-debate probabilities (0.56-0.60) yields the deterministic committee value of 0.57.
- The 1.1% cushion is thin relative to typical 1-2 day volatility, making this a near-coin-flip binary rather than a confident directional call.
- No analyst identifies a statistically reliable directional driver (momentum or mean-reversion) for the next 2-3 trading sessions; the question resolves close to a noise-dominated random walk.
- The recent 2-day uptick (88.12 -> 90.99) is ambiguous: it could reflect genuine stabilization (B, D's lean) or be statistically indistinguishable from noise given the failed mean-reversion retest (F, E's lean).
- WTI (Dec26 proxy) settles above $89.99 on 2026-11-10 with consensus probability 0.58, credible interval [0.49, 0.67].
- Expected 4-day forward log-return of the underlying is approximately +0.5%, consistent with the near-zero-drift bootstrap medians reported by B and D, tempered toward zero by the skeptics' no-drift framing.
- Whether the Nov4-6 rebound reflects a real stabilizing force or is statistically indistinguishable from the volatility noise that produced the preceding 104->88 drop.
- The correct vol-scaling horizon (2 vs 3 vs 4 trading days) is inconsistently applied across analysts, which materially shifts the implied probability by several points.
- No market-implied price/odds for this specific strike were supplied to any analyst, so crowding/mispricing relative to an actual tradable quote cannot be assessed -- this is a fair-value estimate, not a confirmed edge over the market.
- Left-tail/skew risk: two near-9% single-day moves occurred within the last 10 observations, meaning the true downside tail may be fatter than the bootstrap or normal-approximation methods capture.
| Commodities Analyst (A) | 0.18 | Highest domain relevance (1.0 commodities) and solid calibration (brier 0.2695, n=82). Uses normal-approx vol scaling independent in method from the bootstrap crowd, but shares the same underlying price/vol series as everyone else -- partial duplication. |
| Weather Analyst (B) | 0.13 | Best raw calibration brier (0.2679) but lowest domain relevance to commodities (weather=1.0, commodities=0.7). Method (bootstrap of 4-day forward returns) is duplicated by Quant Researcher D using near-identical inputs -- downweighted for redundancy. |
| Markets Analyst (C) | 0.16 | Good calibration (brier 0.2649, highest accuracy 0.554 in panel) and adds a distinct framing (driftless random walk + explicit OOS failure of the mean-reversion insight). Moderate independence from bootstrap-only analysts. |
| Quant Researcher (D) | 0.19 | Broadest cross-domain relevance and largest calibration sample (n=101). Bootstrap method is the most explicitly quantified (median +0.83%, 25th pct -3.59%), but overlaps heavily with Weather Analyst's bootstrap -- counted as one evidence line, weight set slightly above B due to larger n and tighter method description. |
| Positioning Analyst (E) | 0.15 | Lowest stated confidence (0.35) and explicitly flags absence of market-implied pricing, limiting actionability. Adds independent value by naming the insight's diagnostic failures (DW=0.20 autocorrelation, Cook's D=0.131 outlier flag) not cited elsewhere. Calibration brier 0.2717 is the weakest in the panel. |
| Causal Skeptic (F) | 0.19 | Best calibration brier in panel (0.2618, n=105) and the most epistemically conservative framing (near-random-walk, no causal driver identified). Provides the cleanest independent falsification of the mean-reversion insight, anchoring the group away from overconfidence. |
- If WTI (or the referenced Dec26 contract) settles below $89.99 on 2026-11-10, this forecast (P=0.58 for YES) is falsified in that instance.
- If the mean-reversion insight (ins_01M6T125W16K1ZG16374JR) is retested with a new out-of-sample sample between now and Nov10 and shows p<0.05 with hit rate >55%, the 'no reliable directional edge' premise underlying this synthesis is falsified and the probability should be revised.
- If realized daily volatility over Nov7-10 exceeds 5%/day (roughly 1.5x the 3.5% trailing estimate used), the vol-scaling base rate underlying this estimate is invalidated and should be recomputed.
The seats
Commodities Analyst56%
58% Unconditional probability that a crude oil settlement 2 trading days out lands above a strike that is ~1.1% below the last observed close, assuming roughly symmetric daily returns with historical vol.
The return distribution is left-skewed (skew=-0.77) with fat tails (kurtosis=3.87, s2), meaning large negative one-day moves (like the -8.66% on Nov4) are more probable than a normal model implies. A single bad headline (OPEC+ supply surprise, demand data miss, geopolitical de-escalation lowering risk premium) could easily push the Dec WTI contract through the 89.99 strike within just two sessions, and the market's recent realized vol regime (large swings Oct30-Nov4) argues for using a wider, fatter-tailed distribution than my point estimate assumes.
- Last close 90.99 on 2026-11-06, only ~1.1% above the 89.99 strike; prior days show a sharp drop to 88.12 (Nov4) followed by partial recovery to 90.52, 90.99.
- Daily log-return sd ≈ 3.5% (s3 CI [2.97%,4.00%]), implying ~4.95% 2-day sd — a 1.1% cushion is well within one standard move, so normal-approx P(stay above)≈0.59.
- Re-test of the mean-reversion insight (insight:ins_01M6T125W16K1ZG16374JR) today gives beta=-0.189, CI includes 0, p=0.273 — the rebound signal is NOT significant now, so I down-weight its earlier OOS +2.47% claim.
- Walk-forward OOS excess return is only 0.17% with p=0.806 — essentially no edge from the mean-reversion rule in the current regime.
If the Dec WTI contract settlement price on 2026-11-10 is at or below 89.99 USD/Bbl, this thesis is falsified; conversely, if volatility collapses and price settles comfortably above (e.g., >92) that would support but not uniquely confirm this reasoning.
- Settlement reflects the December WTI contract price path from the current 90.99 close with no major scheduled catalyst (OPEC meeting, EIA inventory shock) known to me before as_of.
- Daily log-return distribution (s2/s3) is a reasonable proxy for the next 2 trading sessions despite the recent elevated-vol regime.
- Nov7-Nov8 are non-trading weekend days, so only Nov9 and Nov10 settlement prices matter for the path to resolution.
- The library mean-reversion insight's failed re-test today is treated as evidence of no current edge, not as evidence of reversal in the opposite direction.
debate skipped: panel unanimous within 0.04
one observation has outsized influence (Cook's D=0.131) · residual autocorrelation (DW=0.20); HAC lags=10 applied
Weather Analyst59%
Loading…
Public Markets Analyst56%
Loading…
Quantitative Researcher60%
Loading…
Market Positioning Analyst57%
Loading…
Causal Skeptic56%
Loading…
Claims and evidence
The debate
| Commod | Weathe | Public | Quanti | Market | Causal | |
|---|---|---|---|---|---|---|
| Commodities | ||||||
| Weather | ||||||
| Public Marke | ||||||
| Quantitative | ||||||
| Market Posit | ||||||
| Causal Skept |
WTI settled at 90.99 on 2026-11-06, giving only a ~1.1% cushion above the 89.99 strike with two trading sessions (Nov9, Nov10) remaining before settlement. Given realized daily volatility near 3.5%, this cushion is thin but the base case (random walk) still favors staying above the strike more often than not.
WTI (Dec26, proxied by d1) closed 1.10% above the $89.99 strike at 90.99 on Nov 6, and has shown 2 consecutive days of recovery after an Nov4 shock low of 88.12. A bootstrap of the 4-day forward return distribution (s3) implies roughly 63% probability of staying above the strike through Nov10 settlement, driven by a slightly positive median drift and the threshold sitting between the 25th and 50th percentile of simulated outcomes. I discount this somewhat for fat-tailed, left-skewed downside risk given two near-9% single-day moves occurred within the last 10 observations, and because the one library insight suggesting mean-reversion support is not statistically robust (insignificant regression, near-zero walk-forward OOS excess). Net: slightly favor YES but with real tail risk.
WTI (D1 proxy) closed at 90.99 on 2026-11-06, only 1.1% above the 89.99 strike, after a highly volatile week (104.1 -> 88.12 -> 90.99). With daily realized vol near 3.1% (s2) and no significant directional edge from the library's momentum/reversal insight (OOS retest failed: p=0.806, hit rate 0.455), the 4-day outcome is close to a coin-flip. I lean slightly toward YES (0.56) purely because the current price already sits above the strike and a driftless-random-walk prior preserves that starting edge, but the large volatility relative to the thin 1.1% buffer and recent downside skew keep this estimate close to 50/50 with low confidence.
WTI settled at 90.99 on 2026-11-06, giving a 1.10% buffer above the 89.99 strike. The settlement horizon spans only 2 trading sessions (Nov 9-10). Bootstrapped and volatility-adjusted analysis puts the probability of staying above the strike moderately above 50%, tempered by elevated recent volatility and negative return skew.
The question is a near-the-money binary with spot (90.99 on 11/06) sitting only ~1.1% above the 89.99 strike, against a realized daily return sd of ~3.5% (s2) over the trailing 20 sessions. A pure random-walk diffusion over the ~3 remaining trading sessions to Nov 10 gives roughly a 55-60% chance of settling above strike, with no reliable directional edge from the library's mean-reversion insight because its live re-test failed (i1_wf OOS hit rate 0.455, p=0.806; i1_reg beta -0.189, p=0.273, DW=0.20 flags severe residual autocorrelation and Cook's D=0.131 flags outlier distortion). The insight's own retest today is evidence against relying on it, even though the headline historical stats (beta -0.331, hit 0.68) looked promising before. No market-implied quote was supplied in the data, so I cannot measure crowding directly; this estimate is a fair-value anchor, not a trade signal, and the lack of edge is itself a reason for caution.
No tested signal in the library demonstrates a reliable causal driver of WTI's path over the next 4 days; the headline mean-reversion insight fails its own retest (null correlation, sub-chance OOS hit rate, pathological residual autocorrelation from overlapping windows). The question therefore resolves close to a noise-dominated coin flip around the current price, modestly favoring YES only because the threshold sits slightly below the last observed close.
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
Synthesis: six analysts converge tightly (0.56-0.60) but on substantially overlapping evidence -- the same WTI price series (90.99 close, 1.1% cushion to strike), the same ~3-3.5% vol estimate, and the same failed retest of the library's mean-reversion insight. This is effectively 2-3 independent lines of evidence (price/vol diffusion math, bootstrap empirical distribution, and the insight-failure diagnostic), not six. I weight the Causal Skeptic (F) and Quant Researcher (D) most heavily for calibration and methodological distinctness, and downweight the Weather Analyst (B) for domain irrelevance and redundancy with D's bootstrap. The deterministic pool gives 0.57; I set consensus at 0.58, slightly above the pool, because the vol-scaling base rate (random walk, no drift, 1.1% cushion vs ~3.5% daily sd over ~2 sessions -> Phi(0.22)≈0.588) is the most defensible single number and nearly all analysts' independent methods cluster just above it. CI [0.49,0.67] reflects genuine model uncertainty in horizon definition (2 vs 4 trading days) and left-tail fatness not fully captured by bootstrap or normal approximations. Tradeability: this is NOT a high-conviction trade. The edge over a flat 50/50 prior is thin (~8 points) and entirely a function of starting price already being above strike, not a demonstrated causal driver -- every analyst who tested the one available directional signal (mean-reversion) found it fails out-of-sample. No market-implied quote was available to any analyst, so I cannot confirm whether 0.58 diverges from the actual tradable price. Recommended expression, if tradeable instruments exist: only take a position if the market-implied probability for 'WTI > 89.99 on Nov10' is observed below ~0.50 or above ~0.68 (i.e., a >10-point gap from this fair-value estimate), sized small given the thin statistical edge and short 4-day horizon; otherwise, no trade -- the crowding score (0.85) reflects that this analysis offers no edge beyond what any vol-aware observer of the same OHLC series would compute.
Assumptions
- reviewSettlement reflects the December WTI contract price path from the current 90.99 close with no major scheduled catalyst (OPEC meeting, EIA inventory shock) known to me before as_of.
- reviewDaily log-return distribution (s2/s3) is a reasonable proxy for the next 2 trading sessions despite the recent elevated-vol regime.
- reviewNov7-Nov8 are non-trading weekend days, so only Nov9 and Nov10 settlement prices matter for the path to resolution.
- reviewThe library mean-reversion insight's failed re-test today is treated as evidence of no current edge, not as evidence of reversal in the opposite direction.
- reviewReturns are approximately i.i.d. and variance scales with sqrt(time)
- reviewd1 last value (90.99 on 2026-11-06) is the Dec26 WTI contract price relevant to settlement, consistent with Kalshi's underlying
- reviewDaily log-return distribution (d2, sd=2.82%) is a reasonable proxy for the Dec contract's near-term vol despite recent regime shift (Nov4 -8.66% shock)
- reviewNo major scheduled macro/OPEC event between Nov7-10 is assumed beyond what's embedded in recent price action (untested - no news data available)
- review4-trading-day bootstrap (s3) is a fair proxy for the actual ~3 trading day window (Nov7,9,10 given possible holiday Nov 9 is a Monday - treating as normal)
- reviewWTI December 2026 contract settlement tracks the underlying WTI spot/futures series represented in d1 closely enough to use d1 as proxy
- reviewDaily log-return distribution (sd≈0.031) from d2/s2 is representative of the next 4 trading days' vol regime
- reviewNo major scheduled macro/OPEC event known to the author before cutoff that would be a one-sided catalyst in either direction over Nov 7-10
- reviewThe Nov 6 to Nov 10 window corresponds to 2 trading sessions (Nov 9, Nov 10), since Nov 7-8 fall on a weekend.
- reviewDaily log returns are approximately uncorrelated at lag 1, so no strong mechanical momentum carries into the next 2 sessions.
- reviewRecent realized volatility (~5%/day over the last 6 sessions) is a better near-term risk estimate than the longer full-sample sd (~3.0-3.5%).
- reviewThe library mean-reversion insight is unreliable for this forecast because its generic retest regression and walk-forward validation both failed to show statistical significance.
- reviewDecember WTI futures settlement on Nov 10 behaves like the front-month series in d1/d2 (no unusual contract-roll basis distortion)
- reviewDaily log-return distribution from s2 (sd 3.52%, skew -1.21) is a reasonable proxy for the next 3 trading days
- reviewNo major scheduled catalyst (OPEC+ meeting, EIA inventory surprise, geopolitical shock) between 11/07 and 11/10 that would invalidate the random-walk diffusion assumption — this is unverified from the data given
- reviewSettlement price moves approximately as a driftless random walk at short (4-day) horizons absent a demonstrated causal driver
- reviewd2/i1_x (5-day rolling log return) and d1 (daily WTI close) are consistent series representing the same underlying asset
- reviewNo scheduled OPEC+ or EIA event between Nov 7-10, 2026 is encoded in the data provided beyond what's reflected in historical volatility
Simulated trade
The portfolio manager proposed no paper trade. Not tradeable at an edge: (1) the ~8-point edge over a flat 50/50 prior is entirely a function of spot already sitting above strike plus a vol-scaling calculation, not a demonstrated causal or informational edge -- any market participant computing the same diffusion math from the same OHLC series arrives at the same number (crowding=0.85); (2) no market-implied probability was available to any analyst, so there is no confirmed mispricing to trade against -- I cannot verify that 0.58 diverges from what options/forward markets already price; (3) statistical quality is middling (0.55) and the analyst pool's apparent 6-way convergence is really 2-3 independent methods, giving false confidence; (4) horizon is short (4d) and transaction costs plus any directional futures exposure would consume most of the thin edge even if it were real. Recommended action: monitor for a market-implied quote on this strike/date; only trade the mispricing (via options/binary) if a >10-point gap from 0.58 appears, sized small (<1% NAV) given the short horizon and lack of causal driver.
Forecasts
| Seat | Stage | P | Base | Resolves | Outcome | Brier |
|---|---|---|---|---|---|---|
| Committee | final | 57% | — | 2026-11-10 | No | 0.329 |
| Commodities Analyst | prior | 80% | 58% | 2026-11-10 | No | 0.640 |
| Commodities Analyst | independent | 56% | 58% | 2026-11-10 | No | 0.314 |
| Commodities Analyst | revised | 56% | 58% | 2026-11-10 | No | 0.314 |
| Weather Analyst | prior | 55% | 62% | 2026-11-10 | No | 0.303 |
| Weather Analyst | independent | 59% | 62% | 2026-11-10 | No | 0.348 |
| Weather Analyst | revised | 59% | 62% | 2026-11-10 | No | 0.348 |
| Public Markets Analyst | prior | 72% | 55% | 2026-11-10 | No | 0.518 |
| Public Markets Analyst | independent | 56% | 55% | 2026-11-10 | No | 0.314 |
| Public Markets Analyst | revised | 56% | 55% | 2026-11-10 | No | 0.314 |
| Quantitative Researcher | prior | 55% | 61% | 2026-11-10 | No | 0.303 |
| Quantitative Researcher | independent | 60% | 61% | 2026-11-10 | No | 0.360 |
| Quantitative Researcher | revised | 60% | 61% | 2026-11-10 | No | 0.360 |
| Market Positioning Analyst | prior | 55% | 50% | 2026-11-10 | No | 0.303 |
| Market Positioning Analyst | independent | 57% | 50% | 2026-11-10 | No | 0.325 |
| Market Positioning Analyst | revised | 57% | 50% | 2026-11-10 | No | 0.325 |
| Causal Skeptic | prior | 50% | 57% | 2026-11-10 | No | 0.250 |
| Causal Skeptic | independent | 56% | 57% | 2026-11-10 | No | 0.314 |
| Causal Skeptic | revised | 56% | 57% | 2026-11-10 | No | 0.314 |