kalshi-weather-markets

Name: kalshi-weather-markets
Author: agiprolabs

60installs53forks

How to install kalshi-weather-markets?

agiprolabs/claude-trading-skills/kalshi-weather-markets

$npx -y skills add agiprolabs/claude-trading-skills --skill kalshi-weather-markets

Installs into the current project.

Summary

Daily temperature high/low bracket and threshold contracts on Kalshi — contract structure, forecast→P(YES) map, settlement rules, cross-venue divergences, and weather-specific pitfalls

Files · 1

View on GitHub

SKILL.md

1	# Kalshi Weather Markets — Daily Temperature High/Low
2
3	Kalshi lists daily high and low temperature options for ~20 US cities as binary contracts that settle YES ($1.00) or NO ($0.00). This skill covers the market structure, the forecast-to-probability map, exact settlement mechanics, and hard-won pitfalls. It builds on the exchange layer — for Kalshi API mechanics (host, auth, orders, order book, candlesticks) see the `kalshi-api` skill; for strategy, sizing, and backtesting see `prediction-market-strategy`.
4
5	## Contract Types
6
7	### Brackets — `B<center>`
8
9	A bracket ticker `B<center>` is a 2°F-wide, both-ends-inclusive window.
10
11	- `B74.5` covers the two integers {74, 75}°F.
12	- YES iff the settled temperature is exactly 74 or 75.
13	- Brackets in one event are mutually exclusive and (with two open tail markets) collectively exhaustive.
14	- Their YES prices sum to the overround (fair = 1.0; > 1.0 = aggregate overpricing).
15
16	### Thresholds — `T<strike>`
17
18	A threshold ticker `T<strike>` is a one-sided binary.
19
20	- `greater` → YES iff `cli >= strike + 1`
21	- `less` → YES iff `cli <= strike - 1`
22	- Critical: `strike_type` (`"greater"` / `"less"`) is not inferable from the ticker. Read it from the API `strike_type` field every time.
23
24	### Ticker Format
25
26	```
27	KXHIGH<CITY>-<YYMONDD>-B<center> # bracket high
28	KXLOW<CITY>-<YYMONDD>-T<strike> # threshold low
29	```
30
31	The date is encoded in the ticker, not derivable from `close_time`.
32	`KXHIGHNY-26JUN21` settles 2026-06-21 LST. `close_time` is next-day UTC (~00:59 ET). Joining on `close_time` off-by-ones every label — use the ticker date.
33
34	---
35
36	## Forecast → P(YES)
37
38	Given a forecast distribution `N(μ, σ)` for the day's extreme, apply the half-integer continuity correction (mandatory — settlement is on integers, not a continuous scale):
39
40	```
41	# Bracket B<center>, covering integers {floor, cap}
42	P(YES) = Φ((cap + 0.5 − μ) / σ) − Φ((floor − 0.5 − μ) / σ)
43
44	# Threshold "greater":
45	P(YES) = 1 − Φ((T + 0.5 − μ) / σ)
46
47	# Threshold "less":
48	P(YES) = Φ((T − 0.5 − μ) / σ)
49
50	Φ(x) = 0.5 · (1 + erf(x / √2)) # stdlib only, no scipy needed
51	```
52
53	The `±0.5` shift is not optional. Dropping it biases every bracket. Treating 2°F brackets as 1°F half-open windows produced a +1640% phantom backtest in one project.
54
55	See `scripts/weather_brackets.py` for runnable implementations of all four functions.
56
57	---
58
59	## Deriving (μ, σ) from Ensemble Quantiles
60
61	```
62	sigma_raw = max((p90 − p10) / 2.56, 0.5) · sigma_scale · sigma_mult
63	mu = p50 # or nowcast-blended (see forecasting.md)
64	sigma = max(sigma_raw, 0.1) # hard floor against degeneracy
65	```
66
67	The 2.56 divisor is the 10th–90th percentile span of a standard normal (2 × 1.28σ).
68
69	---
70
71	## CLI-Space Bias Correction
72
73	The settlement value (NWS CLI integer °F, LST day) is not the same as raw ASOS/METAR hourly max/min — CLI applies QC, backup-station fallback, and LST aggregation. Shift μ before computing P(YES):
74
75	```
76	mu_cli = mu_metar + bias_city_season # bias = oracle_extreme − asos_extreme, fit per city + season
77	```
78
79	Fit `bias_max` / `bias_min` as seasonal (circular) curves per city. Skipping this systematically misprices every bracket for cities with a structural CLI/METAR gap.
80
81	---
82
83	## Settlement Rules
84
85	### Kalshi
86
87	- Source: NWS Climatological Report (CLI) — the official daily climate summary issued by each WFO.
88	- Fallback: IEM ASOS daily download matches CLI 100% and is available programmatically.
89	- Window: LST (Local Standard Time), no DST adjustment. The day runs midnight-to-midnight LST year-round.
90	- Value: Integer °F maximum (HIGH) or minimum (LOW) temperature for that LST day.
91	- Bracket: YES iff `cli ∈ {floor, cap}` (both ends inclusive).
92	- Threshold greater: YES iff `cli >= strike + 1`.
93	- Threshold less: YES iff `cli <= strike - 1`.
94
95	### Settlement-Source References
96
97	> Read each market's own rulebook before scoring or trading. Settlement source, station, and day-window are per-market contract terms that can change.
98
99	\| Resource \| URL \|
100	\|----------\|-----\|
101	\| Kalshi market rules / Rulebook \| <https://docs.kalshi.com> (per-market "Rulebook") \|
102	\| NWS Climatological Report (CLI) \| <https://www.weather.gov/wrh/Climate> \|
103	\| IEM ASOS daily download \| <https://mesonet.agron.iastate.edu/request/daily.phtml> \|
104	\| Polymarket resolution (WU) \| <https://www.wunderground.com> \|
105	\| Polymarket disputes (UMA) \| <https://docs.uma.xyz> \|
106
107	---
108
109	## Cross-Venue Divergence
110
111	The same metro on the same date can settle to different values across venues — both because of the station and the DST window in spring/fall.
112
113	\| Axis \| Kalshi \| Polymarket \|
114	\|------\|--------\|------------\|
115	\| Source \| NWS CLI / IEM ASOS \| Weather U

People who liked this also viewed

6 picks

Type

kalshi-weather-markets

60installs53forks

View on GitHub

How to install kalshi-weather-markets?

agiprolabs/claude-trading-skills/kalshi-weather-markets

$npx -y skills add agiprolabs/claude-trading-skills --skill kalshi-weather-markets

Installs into the current project.

Summary

Daily temperature high/low bracket and threshold contracts on Kalshi — contract structure, forecast→P(YES) map, settlement rules, cross-venue divergences, and weather-specific pitfalls

Files · 1

View on GitHub

SKILL.md

1	# Kalshi Weather Markets — Daily Temperature High/Low
2
3	Kalshi lists daily high and low temperature options for ~20 US cities as binary contracts that settle YES ($1.00) or NO ($0.00). This skill covers the market structure, the forecast-to-probability map, exact settlement mechanics, and hard-won pitfalls. It builds on the exchange layer — for Kalshi API mechanics (host, auth, orders, order book, candlesticks) see the `kalshi-api` skill; for strategy, sizing, and backtesting see `prediction-market-strategy`.
4
5	## Contract Types
6
7	### Brackets — `B<center>`
8
9	A bracket ticker `B<center>` is a 2°F-wide, both-ends-inclusive window.
10
11	- `B74.5` covers the two integers {74, 75}°F.
12	- YES iff the settled temperature is exactly 74 or 75.
13	- Brackets in one event are mutually exclusive and (with two open tail markets) collectively exhaustive.
14	- Their YES prices sum to the overround (fair = 1.0; > 1.0 = aggregate overpricing).
15
16	### Thresholds — `T<strike>`
17
18	A threshold ticker `T<strike>` is a one-sided binary.
19
20	- `greater` → YES iff `cli >= strike + 1`
21	- `less` → YES iff `cli <= strike - 1`
22	- Critical: `strike_type` (`"greater"` / `"less"`) is not inferable from the ticker. Read it from the API `strike_type` field every time.
23
24	### Ticker Format
25
26	```
27	KXHIGH<CITY>-<YYMONDD>-B<center> # bracket high
28	KXLOW<CITY>-<YYMONDD>-T<strike> # threshold low
29	```
30
31	The date is encoded in the ticker, not derivable from `close_time`.
32	`KXHIGHNY-26JUN21` settles 2026-06-21 LST. `close_time` is next-day UTC (~00:59 ET). Joining on `close_time` off-by-ones every label — use the ticker date.
33
34	---
35
36	## Forecast → P(YES)
37
38	Given a forecast distribution `N(μ, σ)` for the day's extreme, apply the half-integer continuity correction (mandatory — settlement is on integers, not a continuous scale):
39
40	```
41	# Bracket B<center>, covering integers {floor, cap}
42	P(YES) = Φ((cap + 0.5 − μ) / σ) − Φ((floor − 0.5 − μ) / σ)
43
44	# Threshold "greater":
45	P(YES) = 1 − Φ((T + 0.5 − μ) / σ)
46
47	# Threshold "less":
48	P(YES) = Φ((T − 0.5 − μ) / σ)
49
50	Φ(x) = 0.5 · (1 + erf(x / √2)) # stdlib only, no scipy needed
51	```
52
53	The `±0.5` shift is not optional. Dropping it biases every bracket. Treating 2°F brackets as 1°F half-open windows produced a +1640% phantom backtest in one project.
54
55	See `scripts/weather_brackets.py` for runnable implementations of all four functions.
56
57	---
58
59	## Deriving (μ, σ) from Ensemble Quantiles
60
61	```
62	sigma_raw = max((p90 − p10) / 2.56, 0.5) · sigma_scale · sigma_mult
63	mu = p50 # or nowcast-blended (see forecasting.md)
64	sigma = max(sigma_raw, 0.1) # hard floor against degeneracy
65	```
66
67	The 2.56 divisor is the 10th–90th percentile span of a standard normal (2 × 1.28σ).
68
69	---
70
71	## CLI-Space Bias Correction
72
73	The settlement value (NWS CLI integer °F, LST day) is not the same as raw ASOS/METAR hourly max/min — CLI applies QC, backup-station fallback, and LST aggregation. Shift μ before computing P(YES):
74
75	```
76	mu_cli = mu_metar + bias_city_season # bias = oracle_extreme − asos_extreme, fit per city + season
77	```
78
79	Fit `bias_max` / `bias_min` as seasonal (circular) curves per city. Skipping this systematically misprices every bracket for cities with a structural CLI/METAR gap.
80
81	---
82
83	## Settlement Rules
84
85	### Kalshi
86
87	- Source: NWS Climatological Report (CLI) — the official daily climate summary issued by each WFO.
88	- Fallback: IEM ASOS daily download matches CLI 100% and is available programmatically.
89	- Window: LST (Local Standard Time), no DST adjustment. The day runs midnight-to-midnight LST year-round.
90	- Value: Integer °F maximum (HIGH) or minimum (LOW) temperature for that LST day.
91	- Bracket: YES iff `cli ∈ {floor, cap}` (both ends inclusive).
92	- Threshold greater: YES iff `cli >= strike + 1`.
93	- Threshold less: YES iff `cli <= strike - 1`.
94
95	### Settlement-Source References
96
97	> Read each market's own rulebook before scoring or trading. Settlement source, station, and day-window are per-market contract terms that can change.
98
99	\| Resource \| URL \|
100	\|----------\|-----\|
101	\| Kalshi market rules / Rulebook \| <https://docs.kalshi.com> (per-market "Rulebook") \|
102	\| NWS Climatological Report (CLI) \| <https://www.weather.gov/wrh/Climate> \|
103	\| IEM ASOS daily download \| <https://mesonet.agron.iastate.edu/request/daily.phtml> \|
104	\| Polymarket resolution (WU) \| <https://www.wunderground.com> \|
105	\| Polymarket disputes (UMA) \| <https://docs.uma.xyz> \|
106
107	---
108
109	## Cross-Venue Divergence
110
111	The same metro on the same date can settle to different values across venues — both because of the station and the DST window in spring/fall.
112
113	\| Axis \| Kalshi \| Polymarket \|
114	\|------\|--------\|------------\|
115	\| Source \| NWS CLI / IEM ASOS \| Weather U

People who liked this also viewed

6 picks

Type