wayback backtest

Run an SPY-benchmarked backtest on the WAYBACK historical event dataset (2+ years, 13K events) instead of the recent live event dataset (2 months, 1.7K events).

liveverifiedunknown~5s typical

mcpprobedboolsaisignals

$0.0020

per call

plus $0.001 platform fee per run

Open MCP server; Cracked bills routing only

Provider Upstream docs

$Set up https://cracked-ai-pearl.vercel.app/SKILL.md, then use Cracked to run mcp-ai-boolsai-signals /wayback_backtest for me.

Input

Field	Type	Description
group_by	string (change_type \| key_path \| key_name \| parent_path \| domain)	Dimension to slice on default "key_path"
min_n	integer	Minimum sample size default 20
horizon_days	integer	Forward-return window default 7
top_k	integer	default 15
since	string	YYYY-MM-DD lower bound on event date (default: when prices start)
exclude_noise	boolean	Filter out is_meta_noise=1 events default true

Call it

curl

curl https://cracked-ai-pearl.vercel.app/v1/run \
  -H "Authorization: Bearer ck_live_..." -H "content-type: application/json" \
  -d '{"provider":"mcp-ai-boolsai-signals","endpoint":"/wayback_backtest","input":{"group_by":"key_path","min_n":20,"horizon_days":7,"top_k":15,"exclude_noise":true}}'

cli

npx cracked-ai run -p mcp-ai-boolsai-signals -e /wayback_backtest -i '{"group_by":"key_path","min_n":20,"horizon_days":7,"top_k":15,"exclude_noise":true}'

mcp

run_tool({ provider: "mcp-ai-boolsai-signals", endpoint: "/wayback_backtest", input: {"group_by":"key_path","min_n":20,"horizon_days":7,"top_k":15,"exclude_noise":true} })

Try it

Runs against your signed-in workspace balance. Sign in if you have not.