Analytics

Analytics endpoints expose market microstructure, portfolio state, and execution-quality metrics used by Terminal and internal SOR analysis.

Tip

Quick start:

GET /v1/analytics/execution for top-level health.
GET /v1/analytics/tca for cost-quality details.
GET /v1/analytics/ab_report for routing-version promotion decisions.

Market Analytics

GET/v1/analytics/market/{symbol}

Returns current market structure for a symbol, including:

NBBO/mid/spread values
aggregate bid/ask depth
bid/ask imbalance
per-venue top-book breakdown

Portfolio Analytics

GET/v1/analytics/portfolio

Returns portfolio summary fields such as:

total position value
unrealized/realized PnL
total fees and net PnL
concentration summary
per-symbol position details

Execution Analytics

GET/v1/analytics/execution

Returns execution-level aggregates grouped into:

orders (submitted/filled/rejected/cancelled, fill and reject rates)
fills (fill count, volume)
latency (submit latency metrics)
execution_quality (slippage/savings/avg execution time)

Venue Quality

GET/v1/analytics/venues

Per-venue routing quality metrics derived from live execution stats:

sent/acked/filled/rejected counts
fill/ack/reject rates
average ack latency
composite quality score
size bucket breakdown (small/medium/large)

TCA Aggregate

GET/v1/analytics/tca

Notional-weighted TCA aggregates from durable order history.

Note

Use TCA endpoints for routing-quality decisions. Route observation telemetry is useful for diagnostics, not promotion decisions.

Query Parameters

Parameter	Description
`symbol`	Optional symbol filter
`since`	ISO8601 lower bound
`until`	ISO8601 upper bound
`venue`	Optional venue filter
`routing_version`	Optional routing version filter (`1`, `2`, `3`)

Output Highlights

weighted averages for IS/spread/fees/impact/savings
weighted percentiles (p50, p75, p95, p99)
total savings and win rate
TCA quality provenance breakdown

Prediction-market fields

Every per-order TCA entry also includes these fields on prediction-market fills (Kalshi / Polymarket). Spot/perp fills leave them at zero:

Field	Meaning
`is_prediction`	`true` when this row came from a binary-outcome fill. Branch on this before interpreting slippage — spot metrics are in price bps, prediction metrics are in probability bps
`arrival_implied_prob_bps`	Arrival mid expressed as implied probability in basis points of the 0–1 range. `0.30 → 3000`
`achieved_implied_prob_bps`	Achieved VWAP as implied probability bps
`probability_slippage_bps`	Side-adjusted shift in implied probability from arrival to achieved. Positive = moved against you. The prediction-market analog of `implementation_shortfall_bps`, but measures shifts in probability rather than price — binary tokens don't have a dollar notional to normalize against

For a YES-token buy at arrival mid 0.30 that fills at VWAP 0.33, probability_slippage_bps = 300 (bought 3 percentage-points of probability more expensively than the reference). For the same flow on a spot instrument, implementation_shortfall_bps would be ~1000bps (3¢ slippage on a $0.30 reference). The two numbers are not comparable.

TCA By Venue

GET/v1/analytics/tca/venues

Per-venue notional-weighted TCA attribution.

Query Parameters

Parameter	Description
`symbol`	Optional symbol filter
`since`	ISO8601 lower bound
`until`	ISO8601 upper bound

TCA Timeseries

GET/v1/analytics/tca/timeseries

Time-bucketed TCA aggregates.

Query Parameters

Parameter	Description
`interval`	`1h`, `4h`, `1d` (hourly default)
`since`	ISO8601 lower bound
`until`	ISO8601 upper bound
`symbol`	Optional symbol filter

TCA Report

GET/v1/analytics/tca/report

Structured report over a fixed period.

Query Parameters

Parameter	Description
`period`	`week` (default) or `month`

TCA Quality Provenance

GET/v1/analytics/tca/quality

Returns TCA coverage and provenance counts (direct, reconstructed, etc.).

A/B Promotion Report

GET/v1/analytics/ab_report

Promotion-readiness report for routing-version experiments.

Query Parameters

Parameter	Description
`days`	Lookback window (`1..90`, default `7`)

Output Highlights

bucketed stats by (routing_version, symbol, side, notional_bucket)
sample sufficiency flags (n >= 30 logic)
bootstrap confidence interval output
promotion_ready decision flag

Tip

Treat promotion_ready as the single rollout signal for V2/V3 promotion. It already incorporates sample floors and bootstrap significance checks.

Route Observations

GET/v1/route_observations

Raw SOR telemetry for diagnostics and model calibration.

Query Parameters

Parameter	Description
`routing_version`	Optional filter
`days`	Lookback window (`1..90`, default `7`)
`limit`	Max rows (`1..1000`, default `100`)

Notes

Includes selected and counterfactual venue rows.
Response also includes observation queue counters (obs_attempted, obs_dropped).