Backtest VaR and expected shortfall forecasts

Use this when you have VaR or Expected-Shortfall forecasts and need to know whether reality breached them more often or deeper than your claimed tail level allo

liveverifiedunknown~5s typical

mcpprobedmcp

$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-com-alphaassay-mcp /assay_var_es for me.

Input

Field	Type	Description
returns*	array	Realised per-period returns (the outcomes your risk model was forecasting for).
var_forecasts*	array	Your model's ex-ante VaR forecasts (positive loss thresholds at level alpha) to backtest f
es_forecasts	string	Optional matching expected-shortfall forecasts; enables the joint (VaR, ES) e-process tail
alpha	string	Tail level, e.g. 0.05 for a 95% VaR (0.001-0.25). Sets the Basel traffic-light expectation default 0.05
benchmark_var_forecasts	string	Optional naive/benchmark VaR forecasts (e.g. a rolling historical quantile) for the Diebol
benchmark_es_forecasts	string	Optional benchmark ES forecasts; switches the EPA test to the joint FZ0 loss.
api_key	string	API key for a paid MCP check. Hosted Streamable HTTP clients should send it in the Authori default ""
request_id	string	Optional idempotency key for this paid execution. Retry the same request_id with the same

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-com-alphaassay-mcp","endpoint":"/assay_var_es","input":{"returns":[],"var_forecasts":[],"alpha":0.05,"api_key":""}}'

cli

npx cracked-ai run -p mcp-com-alphaassay-mcp -e /assay_var_es -i '{"returns":[],"var_forecasts":[],"alpha":0.05,"api_key":""}'

mcp

run_tool({ provider: "mcp-com-alphaassay-mcp", endpoint: "/assay_var_es", input: {"returns":[],"var_forecasts":[],"alpha":0.05,"api_key":""} })

Try it

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