Will Natural gas price on Oct 21, 2026 at 5:00 PM EDT close above $3.215?

Settling above $3.215 by 5pm EDT Oct 21 requires an ~8.5% move from the last observed close of $2.963 (Oct 20), against a five-session losing streak and a negatively-skewed return distribution — a low-probability tail event, not the base case.

Own objection: Natural gas is prone to sharp upside squeezes on cold-weather repricing; kurtosis of 46.5 means extreme moves (in either direction) are far more common than normal-distribution math implies, so an 8.5% rally may not be as rare in this series as the skew-adjusted estimate suggests, especially if the HDD forecast data (which failed to load) would have shown an incoming cold snap.

Settling above 3.215 USD/MMBtu by 5pm EDT Oct 21 requires an intraday rally of roughly +8.5% from the last confirmed close (2.963 on Oct 20), which is near the extreme tail of this contract's observed daily move distribution. Absent a major cold-shock catalyst materializing same-day, this is a low-probability outcome.

Own objection: Natural gas is a famously fat-tailed, jump-prone market (skew 0.26, kurtosis 0.44 in d3; observed max move already equals the needed threshold), and d4's sharp jump to +1.667 (its second-highest reading in 25 observations) could signal a genuine cold-forecast revision that triggers an outsized demand-driven rally. If this is a true HDD-anomaly spike rather than noise, the tail probability could be materially higher than the unconditional 0.8% base rate, perhaps 2-3x higher.

Base rate for an 8.5%+ single-day up-move in NYMEX natural gas futures, derived from the empirical daily log-return distribution, is approximately 5%. Given the fresh six-session downtrend (2.963 last close, down from 3.283 a week prior) and no statistically significant mean drift in returns (bootstrap CI straddles zero), I see no fundamental or technical reason to adjust this upward; if anything the negative return skew argues for adjusting slightly below the naive normal-approximation estimate of ~7.6%. I settle at P(YES)=0.05, reflecting that the market would need an outsized, catalyst-driven reversal within a single session to breach 3.215 by 5PM EDT 10/21.

Own objection: Natural gas is notoriously prone to violent short-covering rallies or weather-driven repricing (e.g., a surprise cold-snap forecast update or bullish EIA storage miss) that can produce single-day moves well in excess of 8.5%, and the kurtosis of 24.4 in s1 confirms fat tails exist in BOTH directions historically (max observed return in d2 was +28.9%, larger in magnitude than the worst single-day drop of -47.5% but still showing large up-moves do occur). If such a catalyst hits before 5PM EDT, this estimate is too low and the market could gap straight through 3.215 intraday regardless of the prior week's downtrend, since futures markets reprice very fast on weather model updates.

Yesterday's (2026-10-20) close was 2.963 (d1/d4), and the threshold for YES is 3.215 — a required one-day log-move of +8.17% (ln(3.215/2.963)=0.0817). This far exceeds the typical daily volatility observed in the return series (d2 sd=2.61%, d3 sd=5.27%), and the 5-day trend has been persistently negative (3.283→3.198→3.236→3.116→3.058→2.963), i.e., -11.6% cumulative over 5 sessions. Under a fat-tailed empirical distribution (d3, kurtosis=25.2), the 95th percentile of daily returns is only +5.55% — below the +8.17% needed. This places the required move beyond the empirical 95th percentile, even accounting for heavy tails.

Own objection: Natural gas is notorious for discrete, large single-day jumps driven by weather forecast revisions (cold snap forecasts), storage report surprises, or pipeline/supply disruptions — events not captured by the smooth historical return distribution. The kurtosis of 25.2 in d3 and the extreme tail values (min -47.5%, max +28.9%) confirm the series has experienced moves far larger than needed here, meaning a 3x-normal-vol event is not actually rare for this asset class; if a cold-weather catalyst hits on this specific day, an 8%+ intraday reversal is plausible despite the recent downtrend.

No market quote data (bid/ask implied probability, open interest, or funding) was provided for this Kalshi contract, so I cannot directly observe what the market has priced in; I infer the implied probability should track closely with the statistical likelihood of an 8.5% one-day rally from the last close of 2.963, given the strong, persistent downtrend. That statistical likelihood is low (~2-5%) given observed return distributions, and absent evidence of a fresh bullish catalyst (e.g., cold weather alert) in the dataset, I default close to the base rate with a small downward adjustment for the accelerating multi-day downtrend (d1, d4) offset slightly by a small upward adjustment acknowledging gas's known capacity for outsized tail moves (as evidenced by the sample's own 9.06% max return).

Own objection: Natural gas is notoriously capable of violent single-day moves on short-notice weather forecast revisions (cold blasts, storage draw surprises) — the empirical max return in the sample (+9.06%) already exceeds the 8.5% threshold needed, meaning a move of this exact magnitude has happened at least once in the lookback window, so dismissing it as negligible-probability tail risk could understate true frequency, especially heading into shoulder-season volatility where forecast-driven repricing is common.

None of the causal mechanisms offered in the data (mean-reversion regression, event-count effects, storage-flow correlation) are robust enough to justify predicting a breach of 3.215; the regression is run on a non-stationary series (spurious-regression risk), the event-count variable shows no significant effect versus a random-date placebo, and the storage correlation is non-significant with inflated apparent precision due to high serial correlation. Stripped of these shaky causal stories, the question reduces to: can price move +8.5% in one session against a 6-day declining trend? Historical daily volatility (sd≈2.7%) says this is a ~3-sigma event, i.e., rare.

Own objection: None of the supplied data includes same-day (10/21) weather, pipeline, or storage-report news, which are the actual proximate drivers of single-day natural gas price spikes; a cold-snap forecast or supply disruption announced after 12:06 UTC today could independently drive an 8%+ intraday spike regardless of what the historical return distribution says, and my base-rate-from-sd approach assumes a roughly normal/stationary return process that natural gas (a commodity with well-documented fat-tailed, weather-driven jump risk) is known to violate — the sample kurtosis of only 0.33 here may simply reflect a calm recent period rather than the true tail risk.

debate skipped: panel unanimous within 0.03

debate skipped: panel unanimous within 0.03

debate skipped: panel unanimous within 0.03

debate skipped: panel unanimous within 0.03

debate skipped: panel unanimous within 0.03

debate skipped: panel unanimous within 0.03

Base rate: unconditional empirical frequency of an 8%+ one-day up-move in NYMEX natural gas is roughly 3-5% across the sampled windows (95th percentile up-moves cluster at +3.7% to +5.6%, below the +8.2-8.5% needed). Starting point before any update: ~0.05 (midpoint of raw empirical tail estimates). Update: conditioning on the current 5-6 session accelerating downtrend (-11.6% cumulative) and negative return skew (more downside tail mass than upside), I adjust down to 0.04, matching the deterministic pool. I did not adjust further despite the causal_skeptic's valid point that none of the mean-reversion/event-count causal stories hold up, because that argument is symmetric — it removes a reason to go up, but doesn't independently push probability down further; it just removes noise from analysts who might have argued for upside correction. Duplicated evidence collapsed: five of six analysts (commodities, weather, markets, quant, positioning) all cite the same core fact pattern — last close $2.963, six-session decline from $3.283, and a measured daily-return distribution with negative skew/high kurtosis whose 95th percentile falls short of the +8.2-8.5% required move. This is one evidentiary line (price series + return distribution), not five. Only the causal_skeptic's ADF unit-root test and placebo-check on causal regressions constitute genuinely independent evidence. Tradeability: This is a low-probability, well-identified tail bet with converging analyst estimates (0.03-0.06) and a clean falsification date (Oct 21, 5pm EDT). It IS tradeable as a "NO" / sell-YES position on a binary contract at $3.215 strike, but only if the market-implied price diverges meaningfully from ~0.04 — no quote data was available to any analyst, so I cannot confirm an edge exists right now. Recommended expression: before execution, check the live Kalshi/CME-equivalent implied probability; if it trades above ~0.08-0.10 (i.e., market overpricing the tail event, perhaps due to headline cold-weather risk), sell YES / buy NO with a defined stop given the binary's one-day horizon and hard resolution time. Do not size this as a directional futures trade — the edge, if any, is specifically in the binary-contract mispricing versus the ~0.04 model probability, not a view on natural gas fundamentals broadly. If no mispricing versus ~0.04 is observable, this is not tradeable and should be passed.

Consensus 4% · interval 2% to 8%