Sea-Tac Airport · KSEA
Seattle, Seattle-Tacoma International Airport
As of 12:05 am · Wednesday, August 12
Clear
66.20°F
station reports whole-degree readings (±1°)
↑76.00°
↓58.00°
Trend
62.10°
by wed 3:05 am
Raw local trend ±1.41°/hr
-2.02°/hr
Applied, after damping
68%
How much of the raw trend the model trusts, based on proximity to sunrise/peak-heat hour and current cloud/wind. Low near those turning points, since the trend is about to reverse.
Uncertainty band
57.60°–66.60° (±9.00°)
Widens when local pressure is falling (unsettled weather approaching), and independently when the marine push or offshore flow signals below are elevated - separate from the trend confidence above.
Widened: offshore flow signal rising; noisy local trend fit.
Last 6h + projection
6h ago
wed 3:05 am
Conditions · last 6h
6h ago
now
The same station readings the model itself runs on - what the trend fit and the cross-station gradient network are built from. History only; the dashed projection above is temperature-specific.
Estimated hottest/coldest time of day
↑
76.00°
projected for today's peak-heat hour · est. peak ~4:06 pm
NWS hourly forecast - too far out for the local trend to be worth anything (local trend 66.2° · NWS 76.0°)
↓
58.00°
today's overnight low, almost here · est. trough ~wed 6:02 am
NWS hourly forecast - too far out for the local trend to be worth anything (local trend 59.2° · NWS 58.0°)
Confidence, from similar past days
↑74%based on 5 days with marine push steady, offshore flow risingMAE 1.64°F, bias -1.32°F
↓86%based on all 21 finalized days · no specific match yet (2 of 5 needed)MAE 0.41°F, bias +0.21°F
Measured from finalized days whose marine-push, offshore-flow and pressure-gradient state at prediction time matched today's, not one average across every day regardless of conditions (calibration_log.py). Falls back to the all-days figure until a specific combination has 5 matching days behind it - the sample count above always says which one you are looking at.
Actual so far today
↑
66.20°
recorded at 12:00 am
station reported 19°C (whole degree), so this is ±0.9°F
↓
64.40°
recorded at 12:10 am
station reported 18°C (whole degree), so this is ±0.9°F
Yesterday
↑
≈78–80°
recorded at 3:40 pm
↓
55.0°
recorded at 5:53 am
from the observation stream; NWS's final report normally lands ~1:25 am and isn't out yet
Conditions
Wind
4.60 mph
Cloud cover
0%
Pressure
steady
Trend confidence
68%
Marine push
steady
via KHQM + KORS
Offshore flow
rising
via KELN
Trend confidence is how much the model trusts its own short-term extrapolation right now - not confidence in the readings above. It's deliberately low near sunrise or the day's peak-heat hour, since the trend is about to reverse there regardless of how strong it looks right before. Marine push (coastal/strait pressure gradient) and offshore flow (interior gap-wind gradient) are two distinguishable patterns behind Seattle temperature swings - the latter is the mechanism behind most real heat events, not just solar heating. Both are still first-pass signals, not yet calibrated against real outcomes (see calibration log).
Tomorrow's high
80.00°
~4:00 pm · reference confidence 72%*
From the NWS hourly forecast (HRRR model) - real atmospheric dynamics, not this tool's own trend/persistence guess. Reference confidence is measured from the last 20 finalized next-day predictions using this same source (calibration_log.py) - not a placeholder.
Tomorrow's low
58.00°
~6:00 am · reference confidence 79%*
From the NWS hourly forecast (HRRR model) - real atmospheric dynamics, not this tool's own trend/persistence guess. Reference confidence is measured from the last 20 finalized next-day predictions using this same source (calibration_log.py) - not a placeholder.
Kalshi · today's high market
Estimated high: 76.00°F
71° or below
72° to 73°
74° to 75°
76° to 77°
78° to 79°
80° or above
Highlighted bracket is the one this tool's own estimate falls into - not a probability call, just a quick eyeball check against the market's pricing.
Bid/ask are the current highest buy offer and lowest sell offer for that bracket (last-price is the most recent trade, which can lag). A wide gap between them usually means a thin, less reliable market.
KXHIGHTSEA-26AUG12
Kalshi · today's high market volume
6,945
contracts traded, last 24h
Dollar volume (our estimate)
≈$1,444 est.
Contract count matches Kalshi's own "Volume" figure; the dollar amount is our own estimate (contracts × that hour's average price) since Kalshi doesn't publish a dollar total directly. Applies to all four volume charts on this page (today's/tomorrow's high/low).
Kalshi · today's low market
Estimated low: 58.00°F
51° or below
52° to 53°
54° to 55°
56° to 57°
58° to 59°
60° or above
Trades stay manual for now - this is a side-by-side view for a human to review, not an automated signal.
KXLOWTSEA-26AUG12
Kalshi · today's low market volume
1,476
contracts traded, last 24h
Dollar volume (our estimate)
≈$381 est.
Kalshi · tomorrow's high market
Market not open yet.
Highlighted bracket is the one this tool's tomorrow-high estimate falls into.
no open market
Kalshi · tomorrow's high market volume
—
contracts traded, last 24h
Dollar volume (our estimate)
—
Volume unavailable.
Kalshi · tomorrow's low market
Market not open yet.
Highlighted bracket is the one this tool's tomorrow-low estimate falls into.
no open market
Kalshi · tomorrow's low market volume
—
contracts traded, last 24h
Dollar volume (our estimate)
—
Volume unavailable.
Next-day forecast · track record
| Forecast for | High est. | Actual | Err | Low est. | Actual | Err |
|---|---|---|---|---|---|---|
| 08-13 | 80.0 | — | pending | 58.0 | — | pending |
| 08-12 | 76.0 | 66.2 | pending | 58.0 | 64.4 | pending |
| 08-11 | 78.0 | 79.0 | -1.0 | 57.0 | 55.0 | +2.0 |
| 08-10 | 78.0 | 73.9 | +4.1 | 56.0 | 55.0 | +1.0 |
| 08-09 | 79.0 | 77.0 | +2.0 | 57.0 | 55.0 | +2.0 |
| 08-08 | 82.0 | 82.9 | -0.9 | 59.0 | 57.2 | +1.8 |
| 08-07 | 87.0 | 86.0 | +1.0 | 59.0 | 59.0 | +0.0 |
| 08-06 | 85.0 | 87.8 | -2.8 | 59.0 | 59.0 | +0.0 |
High: MAE 1.74°F, bias +0.03°F over 21 days · Low: MAE 1.16°F, bias +0.35°F over 21 days. Bias is signed forecast minus actual, so positive means the forecast ran warm. Days still in progress show pending - they are excluded from the averages, because scoring against a partial day's high is what made these numbers look twice as bad as they are.
Weekly performance · high (6 of 7 days)
| Date | Predicted | Night before | 1h before pred. | NWS 1h before | Actual | 1h before actual | Temp Δ | Time Δ | Kalshi peak vol. | Kalshi implied |
|---|---|---|---|---|---|---|---|---|---|---|
| 08-11 | no data | |||||||||
| 08-10 | 4:07pm · 78.0° | 78.0° (+4.0) | 69.8° | 76.0° | 4:53pm · 74.0° | 71.1° | +4.00 | -46 min | 3,904 @ 6:15pm | 75° (75° or below) |
| 08-09 | 4:08pm · 78.0° | 79.0° (+2.0) | 75.2° | 78.0° | 3:42pm · 77.0° | 75.2° | +1.00 | +26 min ✓ | 7,127 @ 4:45pm | 78° (78° to 79°) |
| 08-08 | 4:09pm · 83.0° | 82.0° (-1.0) | 78.8° | 83.0° | 4:53pm · 83.0° | 81.0° | +0.00 | -44 min | 7,669 @ 4:00pm | 84° (83° to 84°) |
| 08-07 | 4:10pm · 86.0° | 87.0° (+1.0) | 82.4° | 86.0° | 4:57pm · 86.0° | 84.2° | +0.00 | -48 min | 3,971 @ 3:30pm | 86° (85° to 86°) |
| 08-06 | 4:11pm · 86.6° | 85.0° (-3.0) | 84.2° | 87.0° | 5:47pm · 88.0° | 86.0° | -1.40 | -96 min | 21,711 @ 4:45pm | 88° (87° to 88°) |
| 08-05 | 4:11pm · 86.0° | 88.0° (+1.0) | 84.2° | 82.0° | 5:20pm · 87.0° | 86.0° | -1.00 | -69 min | 20,436 @ 9:30pm | 87° (87° or below) |
Trailing 7 days, most recent first. Predictions only appear once calibration_log.py has logged them (recent addition) - actual/Kalshi columns work back further.
Weekly performance · low (6 of 7 days)
| Date | Predicted | Night before | 1h before pred. | NWS 1h before | Actual | 1h before actual | Temp Δ | Time Δ | Kalshi peak vol. | Kalshi implied |
|---|---|---|---|---|---|---|---|---|---|---|
| 08-11 | no data | |||||||||
| 08-10 | 5:59am · 55.0° | 56.0° (+1.0) | 55.4° | 56.0° | 3:53am · 55.0° | 55.9° | +0.00 | +126 min | 892 @ 3:15pm | 54° (54° to 55°) |
| 08-09 | 5:58am · 55.9° | 57.0° (+2.0) | 55.4° | 57.0° | 6:23am · 55.0° | 55.4° | +0.90 | -25 min | 2,474 @ 4:00am | 56° (55° to 56°) |
| 08-08 | 5:56am · 57.2° | 59.0° (+1.0) | 57.2° | 59.0° | 5:42am · 58.0° | 59.0° | -0.80 | +14 min ✓ | 1,789 @ 4:30am | 58° (58° to 59°) |
| 08-07 | 5:55am · 58.9° | 59.0° (+0.0) | 60.8° | 59.0° | 4:15am · 59.0° | 60.8° | -0.10 | +100 min | 800 @ 6:15pm | 58° (57° to 58°) |
| 08-06 | 5:54am · 60.2° | 59.0° (+0.0) | 62.1° | 60.0° | 6:30am · 59.0° | 60.8° | +1.20 | -36 min | 2,542 @ 6:15am | 58° (58° to 59°) |
| 08-05 | 5:52am · 59.5° | 59.0° (+0.0) | 61.0° | 58.0° | 5:37am · 59.0° | 62.6° | +0.50 | +14 min | 1,890 @ 5:30am | 58° (58° to 59°) |
Monthly performance · August 2026
High
Temp bias (signed)-0.09°F (10 days)
Temp MAE1.17°F
NWS bias (signed)-0.60°F (10 days)
NWS MAE1.40°F
model's MAE is 0.23°F lower (better) than NWS's this month.
Peak-time bias (signed)-41 min
Peak-time MAE49 min
Mean Kalshi peak volume9,025 contracts
Model agreed w/ Kalshi favourite (at prediction time)90% (10 days)
Low
Temp bias (signed)+0.45°F (10 days)
Temp MAE0.69°F
NWS bias (signed)+0.50°F (10 days)
NWS MAE1.10°F
model's MAE is 0.41°F lower (better) than NWS's this month.
Peak-time bias (signed)-50 min
Peak-time MAE176 min
Mean Kalshi peak volume1,626 contracts
Model agreed w/ Kalshi favourite (at prediction time)56% (9 days)
Bias/MAE use the same (predicted − actual) sign convention as the backtest above. Positive bias means the model's prediction ran warm/high or late; negative means it ran cool/low or early.
Paper trading · strategies compared
Trailing 7 days
Same-day · edge-triggered (1hr leg)
Insufficient data - only 2 bet(s) so far (need at least 5 to treat win rate/band coverage as meaningful).
Same-day · unconditional (high + low)
No simulated bets placed yet in this window.
Next-day advance · cashed out early
No position has been cashed out early yet. A position is only exited when the bid clears the entry ask by at least 3 cents; otherwise it is carried into settlement and counted in the block below.
Next-day advance · held to settlement
No simulated bets placed yet in this window.
August 2026
Same-day · edge-triggered (1hr leg)
Simulated P&L$-2.50
Win rate0% (5 bets)
Avg. edge at entry17 pts
Within stated uncertainty band80%
On the 2 bet(s) where the model and Kalshi disagreed most: model was closer to the actual outcome 0 time(s), market was closer 2 time(s).
Same-day · unconditional (high + low)
No simulated bets placed yet in this window.
Next-day advance · cashed out early
No position has been cashed out early yet. A position is only exited when the bid clears the entry ask by at least 3 cents; otherwise it is carried into settlement and counted in the block below.
Next-day advance · held to settlement
No simulated bets placed yet in this window.
Four independent decision rules, deliberately never summed into one number - each answers a different question, so a combined P&L would be the average of four experiments rather than the result of any of them. Edge-triggered bets only when the model's probability disagrees with Kalshi by at least 15 points (shown here for the 1hr leg only; the 1hr-vs-2hr comparison is its own experiment and lives in the block below). Unconditional bets the bracket the model's point estimate falls into, every day, no edge required. The two next-day advance rows are the same positions split by how they ended: bought at the ask the evening before, then either sold at the bid early because the price moved our way, or carried into settlement and scored against the actual weather. Simulated only - no real orders, ever.
Paper trading · trailing 7 days
1hr before peak
Insufficient data - only 2 bet(s) so far (need at least 5 to treat win rate/band coverage as meaningful).
2hr before peak
No simulated bets placed yet in this window.
Simulated only - no real orders placed. A bet is logged when the model's own (backtest-calibrated) probability disagrees with Kalshi's price by at least 15 points - independently at ~1hr and ~2hr before the predicted peak, so the two lead times can be compared rather than picking one upfront. Kept separate on purpose - see the two blocks below.
Low market: always-bet-the-model vs. edge-gated
Low - edge-triggered (1hr)
Insufficient data - only 2 bet(s) so far (need at least 5 to treat win rate/band coverage as meaningful).
Low - unconditional (1hr)
No simulated bets placed yet in this window.
Low market only. "Edge-triggered" is the same 15-point-disagreement bet already counted above under "1hr before peak" (isolated here to just the low side, since the block above pools high and low together); "unconditional" instead bets every single day on whichever bracket the model's own point estimate falls into, no disagreement required. Comparing the two answers whether just following the model daily beats only betting when a real mispricing shows up. High doesn't get an unconditional mode - this comparison only applies to low.
Paper trading · August 2026
1hr before peak
Simulated P&L$-2.50
Win rate0% (5 bets)
Avg. edge at entry17 pts
Within stated uncertainty band80%
On the 2 bet(s) where the model and Kalshi disagreed most: model was closer to the actual outcome 0 time(s), market was closer 2 time(s).
2hr before peak
No simulated bets placed yet in this window.
Low market: always-bet-the-model vs. edge-gated
Low - edge-triggered (1hr)
Insufficient data - only 4 bet(s) so far (need at least 5 to treat win rate/band coverage as meaningful).
Low - unconditional (1hr)
No simulated bets placed yet in this window.
Seattle (Fremont) · NWS gridpoint forecast
66.2°
Clear · KBFI, 12:05 am
Wind
3.4 mph
Today, per NWS
↑
74.0°
forecast high ~5:00 pm
↓
58.0°
forecast low ~6:00 am
A separate, closer-to-home comparison point - not part of the KSEA model above, and not blended into it. Sea-Tac Airport runs warmer than Fremont in summer (further from Lake Union/marine moderation), so treat these as two different microclimates. Current conditions are real observations from KBFI (Boeing Field, nearest full ASOS station - still ~5-6mi from Fremont). The high/low is NWS's own HRRR-based forecast for Fremont's actual coordinates, not this tool's in-house trend/gradient-network model - no backtesting or paper trading behind these numbers yet.