Current tick grids for up to 200 markets of one provider

The same resolution as GET /v1/markets/tick-size for up to 200 (contract_id, asset_id) pairs of ONE provider in one round trip. Polymarket token ids are read from the metadata cache in a single batch. Rate-limit bucket: LIGHT, flat cost of 1 unit. Not cached (Cache-Control: no-store).

results is positional — results[i] answers items[i] — and each entry is either a tick grid, the predictfun unsupported payload, or a per-item error carrying the same code the single route would have answered with (not_found, tick_unavailable, cache_cold, upstream), or timeout / cancelled for items not started before the 15s batch bound or the caller left. A batch where every item failed is still a 200; only request-level problems (bad body, unknown provider, an empty contract_id, more than 200 items, or a failed token prefetch) fail the whole call.