kalshi-crypto-index-markets

Name: kalshi-crypto-index-markets
Author: agiprolabs

57installs53forks

How to install kalshi-crypto-index-markets?

agiprolabs/claude-trading-skills/kalshi-crypto-index-markets

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

Installs into the current project.

Summary

Kalshi daily and hourly range markets on crypto prices (BTC, ETH) and equity indices (S&P 500, Nasdaq-100) — bracket structure, Gaussian P(YES) modeling on price/vol, close-offset decision timing, longshot-sell edge with honest evidence bounds

Files · 1

View on GitHub

SKILL.md

1	# Kalshi Crypto & Index Range Markets
2
3	Kalshi lists daily (and for crypto, hourly) range bracket markets on the level of four liquid underlyings: S&P 500, Nasdaq-100, Bitcoin, and Ethereum. The math is the same partition-and-Gaussian framework as weather brackets — the variable is just price/return and σ comes from the underlying's realized or implied volatility, not a temperature model.
4
5	> Cross-references: for Kalshi API mechanics see `kalshi-api`; for the strategy, sizing, and backtesting framework see `prediction-market-strategy`; for the temperature counterpart see `kalshi-weather-markets`; for the shared bracket/overround formulas see `prediction-markets/references/brackets-and-settlement.md`.
6
7	---
8
9	## Series & Market-Type Mapping
10
11	```python
12	SERIES_MARKET = {
13	"KXINX": "index", # S&P 500 index level
14	"KXNASDAQ100": "index", # Nasdaq-100 index level
15	"KXBTC": "crypto", # Bitcoin price (USD)
16	"KXETH": "crypto", # Ethereum price (USD)
17	}
18	```
19
20	- `market_type = "index"` — daily range only; underlying is the index points level at the official market close.
21	- `market_type = "crypto"` — daily and hourly range markets; BTC and ETH each have multiple hourly events running in parallel with the daily.
22
23	Cadence is higher than weather: crypto hourlies open and settle throughout the day; daily markets open the prior session and settle at the reference close.
24
25	---
26
27	## Bracket Structure
28
29	Each event is a mutually exclusive, collectively exhaustive partition of the underlying's possible values at settlement:
30
31	- An interior bracket `B<center>` covers a contiguous price band (floor to cap, both-ends-inclusive). Bracket widths are set per-market — read the event's market list; do not assume a fixed width.
32	- Two open-tail markets anchor the partition: a `less-than` (below the lowest bracket floor) and a `greater-than` (above the highest bracket cap).
33	- At settlement exactly one bracket is YES; all others are NO.
34
35	The overround (sum of all YES prices) is typically > 1.0. The excess is concentrated in the cheap tails — the same favorite–longshot bias seen in weather markets. See `prediction-markets/references/brackets-and-settlement.md` for the overround formula.
36
37	---
38
39	## Gaussian P(YES) on Price/Vol
40
41	Given a price forecast distribution `N(μ, σ)` for the underlying at settlement, with bracket covering `[floor, cap]`:
42
43	```
44	# interior bracket
45	P(YES) = Φ((cap − μ) / σ) − Φ((floor − μ) / σ)
46
47	# open-tail, "greater than cap"
48	P(YES) = 1 − Φ((cap − μ) / σ)
49
50	# open-tail, "less than floor"
51	P(YES) = Φ((floor − μ) / σ)
52	```
53
54	`Φ` is the standard normal CDF. Unlike temperature brackets (which settle on integers), price brackets settle on a continuous reference price — the half-integer continuity correction used for weather is not applicable here. Do not add ±0.5.
55
56	### Sourcing σ
57
58	- Realized vol: rolling 5- or 20-day realized volatility of the underlying, scaled to the settlement horizon.
59	- Implied vol: the ATM IV from listed options (SPX/NDX for index, CME/Deribit for crypto) is a forward-looking σ that already embeds the market's distributional view for the horizon.
60	- Hourly markets: for BTC/ETH hourlies, σ is the 1-hour IV or intraday realized vol. Σ shrinks rapidly — even small absolute moves land in a different bracket. Sensitivity is high; prefer IV-derived σ.
61
62	For daily markets, a simple log-return diffusion gives `σ_daily ≈ σ_annual / √252` for index, or the equivalent annualized vol / √365 for crypto. Express in price units (not %) before inserting into the formula.
63
64	---
65
66	## Decision Timing — Close-Offset, Not a City Hour
67
68	This is the key difference from weather markets.
69
70	Weather settles on a daily extreme that occurs at some unknown intraday time. The decision book is read near the likely peak/trough (a city-local hour).
71
72	Crypto and index markets settle at a fixed reference close:
73
74	- Daily S&P 500 / Nasdaq-100: the official market close (4:00 PM ET).
75	- Daily BTC/ETH: a defined reference close published in the market's rulebook.
76	- Hourly BTC/ETH: the close of each hour interval per the rulebook.
77
78	The practical convention used in production:
79
80	```python
81	DECISION_OFFSET_MINUTES = 120 # read book ~2h before settlement close
82	decision_ts = settlement_close_ts - timedelta(minutes=DECISION_OFFSET_MINUTES)
83	```
84
85	This offset balances information freshness (IV and order-book signal) against the risk of being front-run by news that drops in the final window. Tune per-series based on your fill-rate observations.
86
87	---
88
89	## Settlement — Verify the Rulebook Reference
90
91	Settlement is on Kalshi's own `result` field. Do not re-derive from an external feed. The exact reference pric

People who liked this also viewed

6 picks

Type

kalshi-crypto-index-markets

57installs53forks

View on GitHub

How to install kalshi-crypto-index-markets?

agiprolabs/claude-trading-skills/kalshi-crypto-index-markets

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

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Kalshi Crypto & Index Range Markets
2
3	Kalshi lists daily (and for crypto, hourly) range bracket markets on the level of four liquid underlyings: S&P 500, Nasdaq-100, Bitcoin, and Ethereum. The math is the same partition-and-Gaussian framework as weather brackets — the variable is just price/return and σ comes from the underlying's realized or implied volatility, not a temperature model.
4
5	> Cross-references: for Kalshi API mechanics see `kalshi-api`; for the strategy, sizing, and backtesting framework see `prediction-market-strategy`; for the temperature counterpart see `kalshi-weather-markets`; for the shared bracket/overround formulas see `prediction-markets/references/brackets-and-settlement.md`.
6
7	---
8
9	## Series & Market-Type Mapping
10
11	```python
12	SERIES_MARKET = {
13	"KXINX": "index", # S&P 500 index level
14	"KXNASDAQ100": "index", # Nasdaq-100 index level
15	"KXBTC": "crypto", # Bitcoin price (USD)
16	"KXETH": "crypto", # Ethereum price (USD)
17	}
18	```
19
20	- `market_type = "index"` — daily range only; underlying is the index points level at the official market close.
21	- `market_type = "crypto"` — daily and hourly range markets; BTC and ETH each have multiple hourly events running in parallel with the daily.
22
23	Cadence is higher than weather: crypto hourlies open and settle throughout the day; daily markets open the prior session and settle at the reference close.
24
25	---
26
27	## Bracket Structure
28
29	Each event is a mutually exclusive, collectively exhaustive partition of the underlying's possible values at settlement:
30
31	- An interior bracket `B<center>` covers a contiguous price band (floor to cap, both-ends-inclusive). Bracket widths are set per-market — read the event's market list; do not assume a fixed width.
32	- Two open-tail markets anchor the partition: a `less-than` (below the lowest bracket floor) and a `greater-than` (above the highest bracket cap).
33	- At settlement exactly one bracket is YES; all others are NO.
34
35	The overround (sum of all YES prices) is typically > 1.0. The excess is concentrated in the cheap tails — the same favorite–longshot bias seen in weather markets. See `prediction-markets/references/brackets-and-settlement.md` for the overround formula.
36
37	---
38
39	## Gaussian P(YES) on Price/Vol
40
41	Given a price forecast distribution `N(μ, σ)` for the underlying at settlement, with bracket covering `[floor, cap]`:
42
43	```
44	# interior bracket
45	P(YES) = Φ((cap − μ) / σ) − Φ((floor − μ) / σ)
46
47	# open-tail, "greater than cap"
48	P(YES) = 1 − Φ((cap − μ) / σ)
49
50	# open-tail, "less than floor"
51	P(YES) = Φ((floor − μ) / σ)
52	```
53
54	`Φ` is the standard normal CDF. Unlike temperature brackets (which settle on integers), price brackets settle on a continuous reference price — the half-integer continuity correction used for weather is not applicable here. Do not add ±0.5.
55
56	### Sourcing σ
57
58	- Realized vol: rolling 5- or 20-day realized volatility of the underlying, scaled to the settlement horizon.
59	- Implied vol: the ATM IV from listed options (SPX/NDX for index, CME/Deribit for crypto) is a forward-looking σ that already embeds the market's distributional view for the horizon.
60	- Hourly markets: for BTC/ETH hourlies, σ is the 1-hour IV or intraday realized vol. Σ shrinks rapidly — even small absolute moves land in a different bracket. Sensitivity is high; prefer IV-derived σ.
61
62	For daily markets, a simple log-return diffusion gives `σ_daily ≈ σ_annual / √252` for index, or the equivalent annualized vol / √365 for crypto. Express in price units (not %) before inserting into the formula.
63
64	---
65
66	## Decision Timing — Close-Offset, Not a City Hour
67
68	This is the key difference from weather markets.
69
70	Weather settles on a daily extreme that occurs at some unknown intraday time. The decision book is read near the likely peak/trough (a city-local hour).
71
72	Crypto and index markets settle at a fixed reference close:
73
74	- Daily S&P 500 / Nasdaq-100: the official market close (4:00 PM ET).
75	- Daily BTC/ETH: a defined reference close published in the market's rulebook.
76	- Hourly BTC/ETH: the close of each hour interval per the rulebook.
77
78	The practical convention used in production:
79
80	```python
81	DECISION_OFFSET_MINUTES = 120 # read book ~2h before settlement close
82	decision_ts = settlement_close_ts - timedelta(minutes=DECISION_OFFSET_MINUTES)
83	```
84
85	This offset balances information freshness (IV and order-book signal) against the risk of being front-run by news that drops in the final window. Tune per-series based on your fill-rate observations.
86
87	---
88
89	## Settlement — Verify the Rulebook Reference
90
91	Settlement is on Kalshi's own `result` field. Do not re-derive from an external feed. The exact reference pric

People who liked this also viewed

6 picks

Type