Will WTI oil on Oct 13, 2026 close above $91.49?

WTI spot (last print $92.63, d1) sits $1.14 (1.23%) above the $91.49 strike with 3 trading days left to settlement (Fri Oct 9, Mon Oct 12, Tue Oct 13). Daily return volatility (d2: sd=3.15%/day) is large relative to that buffer, and mean daily drift is statistically indistinguishable from zero (s3: bootstrap CI95 for mean return = [-0.59%, +0.80%], straddling zero). Treating the 3-day move as a mean-zero random walk with sd≈5.46% (3.15%×√3), the probability that cumulative return stays above the required -1.24% threshold is ≈61%, which I discount slightly to 58% given (a) elevated realized volatility this week (price round-tripped from 92.87→88.28→92.63 in six sessions) and (b) the Oct 8 price pop (+4.8%) occurred on unusually thin volume (d5: 78,190 vs. 14-day mean 297,800), raising the chance of a partial reversal as liquidity normalizes."

Own objection: The random-walk/no-drift model ignores that the Oct 8 +4.8% rally may reflect a real fundamental catalyst (e.g., supply disruption, OPEC+ action, geopolitical event) not visible in this price-and-volume dataset alone. If that move was information-driven rather than noise, the new $92.63 level could be a stable higher plateau rather than a transient spike prone to reversion, which would push the true YES probability meaningfully above my 58% estimate. Conversely, if it was a short-covering/thin-liquidity artifact, my estimate could be too high — the data cannot distinguish between these two stories, which is the central source of my only moderate (0.45) confidence."

WTI (d1) closed Oct 8 at $92.63, 1.23% above the $91.49 Kalshi strike. The 5-day bootstrap return distribution (s3) implies roughly 61% probability the cumulative 3-trading-day decline into Oct 13 settlement stays shallower than -1.23%, consistent with a stationary, mean-reverting price process (ADF p=0.002, s5) whose 21-day mean (95.07) sits above the current spot. Recent two-day momentum (88.28 -> 92.63) and a positive, if noisy, correlation with a co-moving factor (d4, s4) are additional modest supports. However, elevated 5-day downside tail risk (p_loss=0.438, VaR5=-13.9%) and a sudden volume collapse on the final observed day introduce meaningful uncertainty that prevents higher conviction than roughly 60%.

Own objection: The bootstrap distribution (s3) has p_loss=0.438 (44% chance of any negative 5-day return) and a 95% CI lower bound of -13.9%, meaning the tail risk of a large adverse move (e.g., a surprise OPEC+ supply announcement, demand shock, or risk-off dollar rally) is far from negligible and could easily erase the 1.23% cushion. Additionally, the sharp single-day volume collapse (d5: 78k vs ~315k average) could reflect either early holiday-thin liquidity or a specific supply/demand catalyst not captured in the price series, meaning the historical correlation and stationarity evidence (s4, s5) may not hold if a regime shift is underway that the 21-day window cannot detect. Given n=21 for several key stats and strong autocorrelation (0.88) explicitly flagged as making p-values optimistic, the true confidence in the "stay above 91.49" call should be treated as closer to a coin flip than the bootstrap model's point estimate suggests.

Using d2's empirical daily log-return distribution (mean 0.1175%, sd 3.208%) scaled to a 4-trading-day horizon (mean~0.47%, sd~6.42% under iid assumption), and solving for P(cumulative return > ln(91.49/92.63) = -1.239%), a normal approximation gives P(YES)~60.5%. Adjusting downward for the documented negative skew and fat left tail in d2 (skew -0.72, kurtosis 4.78 per s2), and for the fact that spot traded below 91.49 on three of the last four sessions (Oct 5-7) before the Oct 8 rebound, I discount this to ~0.56. The ADF test (s7, p=0.00225) rejecting a unit root in the 21-day price level is consistent with range-bound/mean-reverting behavior in the ~88-96 band, which cuts both ways - it caps upside but also limits sustained downside, modestly supporting the YES case given price is currently above the threshold.

Own objection: The d2 return series has significant negative skew (-0.72) and excess kurtosis (4.78) — large down-moves happen more often than a normal distribution implies. My normal approximation likely overstates P(YES) by not weighting the fat left tail enough, and the realized path (89.43, 89.44, 88.28 on Oct 5-7, all below the 91.49 threshold) shows the market has already traded below this level three times in the last week before bouncing back on Oct 8. That bounce (+4.81% in one day, d2 last value) is itself a high-variance event — if it partially mean-reverts (which the -0.4548 correlation in s5 weakly suggests, though p=0.044 is fragile at n=20 with autocorrelation inflating significance), spot could easily retest sub-91.49 territory within 4 trading days without any new shock, just continuation of the Oct 5-7 regime rather than the Oct 8 spike being the new baseline.

Base rate for a random-walk asset to stay above a threshold that sits only 1.25% below the current price over a ~3-4 session horizon is roughly 60% given near-zero drift and negligible return autocorrelation observed in both daily (d2) and 4-day (d4/d5) windows. I update only modestly from this base rate: the recent realized choppiness (6.4-point range in 6 sessions) and left-skewed/fat-tailed daily return distribution (s2) slightly increase downside tail risk versus a naive symmetric model, but do not shift the point estimate meaningfully since the cushion (1.25% buffer) and lack of directional momentum dominate. No historical-analog dataset specific to named 2024-2025 OPEC+/geopolitical episodes was queried this session, so that color is flagged as untested domain knowledge, not a quantitative input.

Own objection: The left-skewed, fat-tailed daily return distribution (s2: skew=-0.70, kurtosis=5.45) means large downside moves are more probable than a symmetric normal model suggests — exactly the kind of tail event (e.g., a surprise demand shock, OPEC+ supply surge, or risk-off macro event) that could push WTI through the 91.49 level even though the median/modal path favors YES. The recent 6-day history itself shows a ~4.6-point intraweek swing (92.87→88.28), proving the instrument is fully capable of a >1.25% move within the remaining 3-4 session window; my estimate could be ~5-8 points too high if I'm underweighting skew/kurtosis relative to the symmetric quantile interpolation I used.

Base rate from spot-vs-strike distance and random-walk assumption is ~54%. I nudge this up to ~0.57 because the level series shows high persistence (autocorr 0.93) and fails to reject a unit root (no detected mean-reversion pull back to the 87-89 range), and the Oct 8 spike coincides with a drop in a positioning/volume proxy (d5) rather than a surge — suggesting the rally wasn't a crowded speculative move prone to immediate unwind. However, I have no direct Kalshi quote for this contract, so I cannot confirm whether this 0.57 is already priced in, underpriced, or overpriced by the market; crowding appears low-to-moderate given the d5 drop, meaning there is plausibly modest tradeable edge if the true market price is near 50/50, but confidence in this read is limited by ambiguous data definitions (d2's implausibly high daily vol, unclear identity of d4/d5).

Own objection: The Oct 7→Oct 8 move (88.28→92.63, +4.9% in one session) is itself evidence of abnormally high short-term volatility — if that single-day jump was driven by a transient headline (e.g., geopolitical supply scare) rather than a durable fundamental shift, mean-reversion of that specific jump within 3-4 days is plausible even without a long-run mean-reverting process, and would pull settlement back below 91.49. My thesis relies on treating the Oct 8 level as the new stable base, but a sharp one-day spike followed immediately by a prior down-week is a pattern more consistent with noise than a regime change, and I have no news/catalyst data to adjudicate which it is.

None of the causal/predictive apparatus in the library (the Dallas-wind-gust → WTI insight and its associated regression) survives basic robustness checks — it fails out-of-sample, is driven by a single influential observation, and has no articulated mechanism connecting Texas wind gusts to global crude settlement at a 5-day horizon. Discarding that signal entirely, the only defensible basis for a forecast is the raw price level itself: WTI closed at 92.63 on 2026-10-08, about 1.2% above the 91.49 threshold, inside a recent 21-day window that has shown 4-5% two-day swings in both directions with no statistically reliable drift (ADF rejects a unit root, consistent with mean-reversion, not trend). Treating the next ~2-3 trading sessions as an undirected walk from a level modestly above threshold justifies only a mild tilt above 50%, not a confident YES call.

Own objection: My estimate leans almost entirely on treating the 5-day move as an undirected random walk calibrated off a 21-day historical sd, but that window itself spans a period of elevated volatility (sd=5.25 on mean=95.07, i.e., unusually wide for WTI), which could overstate the true dispersion going into a shorter, 2-3-trading-day window to Oct 13. If I instead use a shorter realized-vol estimate or if there is a known near-term catalyst (e.g., an OPEC+ meeting or EIA inventory release) not present in this dataset, the true probability could be materially different from 0.55-0.60 in either direction. I have no data to rule this in or out, which is itself the core problem: I am forced to default to a no-information random walk when a real causal driver (supply decision, inventory surprise) could dominate but is invisible to me before as_of.

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 functionally drawing on one core shared dataset (the 5-6 day WTI price/return series) run through closely related bootstrap/normal-approximation machinery, landing in a tight 0.56-0.60 band — this is high crowding and low genuine independence, not six confirmations. I down-weight the headline agreement accordingly rather than treating 6-way consensus as strong evidence. The one substantive, unresolved disagreement — contradictory ADF unit-root tests (B/C: mean-reverting, p=0.00225; E/F: random walk, p=0.421) — is a real methodological fork that the panel did not reconcile, and I resolve it conservatively by defaulting to the random-walk/no-drift framework, since an unreplicated, contradicted stationarity claim should not be allowed to pull the estimate meaningfully above the mechanical distance-to-strike/volatility calculation. This lands me at 0.57, essentially at the deterministic pool (0.58) but very slightly below it, reflecting (a) the fat left tail and negative skew in daily returns (C's s2), (b) the unresolved liquidity signal from the Oct 8 low-volume rally, and (c) a preference for the more assumption-light model given conflicting stationarity tests. Tradeability: directionally this is a near-coinflip-plus-edge situation, not a strong conviction trade. No analyst confirmed an actual quoted Kalshi price for this specific contract — without that, I cannot state an expected dollar return on a position. If the market is quoting YES near 0.50, there is a plausible ~7-8 point modeled edge long YES; if quoted above ~0.60, the trade has negative expected value under this model and should be avoided or faded. Given the thin 1.2% buffer, contradictory stationarity evidence, and small-sample calibration stats underlying every analyst's number, I would size any position small and treat this as a weak-edge, not high-conviction, trade.

Consensus 58% · interval 45% to 68%