cURL
curl -sS -X POST https://smartapex.uk/api/tools/deflated-sharpe/run \
-H 'Content-Type: application/json' \
-d '{ "returns": [ 0.01, -0.002, 0.004, 0.006, -0.001 ], "n_trials": 12, "trial_sharpe_variance": 0.04}'Apex exposes a permissionless DATA ONLY deflated-sharpe wrapper. Send returns and trial count, get Sharpe, PSR, MinTRL, DSR, conventions, and a no-source-release receipt. No signup, no keys, no trading, no wallet access.
Read https://smartapex.uk/llms.txt. Install or call the Apex MCP server at https://smartapex.uk/api/mcp. Run apex_run_deflated_sharpe with returns=[0.01,-0.002,0.004,0.006,-0.001], n_trials=12, trial_sharpe_variance=0.04. Explain whether the backtest looks overfit. Do not request raw Apex source code.
This path is designed for AI agents: read the LLM wiki, use MCP, run the bounded diagnostic, then explain the result without asking for Apex source.
curl -sS -X POST https://smartapex.uk/api/tools/deflated-sharpe/run \
-H 'Content-Type: application/json' \
-d '{ "returns": [ 0.01, -0.002, 0.004, 0.006, -0.001 ], "n_trials": 12, "trial_sharpe_variance": 0.04}'{
"returns": [
0.01,
-0.002,
0.004,
0.006,
-0.001
],
"n_trials": 12,
"trial_sharpe_variance": 0.04
}The sample returns a WATCH-style result: the surface Sharpe looks interesting, but multiple-testing and short track record evidence must be considered.
Trying many strategies and keeping the best one makes ordinary Sharpe and p-values look better than they are.
MinTRL shows how much evidence is needed before a Sharpe-like result becomes more credible.
Apex returns a verification receipt with hashes and boundaries, but never releases the private wrapper source.
{
"mcpServers": {
"apex": {
"url": "https://smartapex.uk/api/mcp"
}
}
}DATA ONLY. Read-only. No orders. No private balances. No wallet access. No source package. Reviews are optional for the permissionless tier.