I’d like to contribute a small tutorial for the unchecked volume-clock/event-clock roadmap item.
It would compare a 100 ms physical-time grid with sampling every N successful wait_next_feed(False, timeout) feed wakeups on the same bundled BTCUSDT sample. It would record local timestamps, mid-price, spread in ticks and top-of-book imbalance, and compare sample spacing and feed activity.
I found the replay loop in Fusing Depth Data and the references in Pricing Framework, but no dedicated comparison or competing open PR. I would reuse the bundled raw data and existing conversion/snapshot utilities, without core API changes or new dependencies.
The tutorial would explicitly explain that a wakeup can process several records sharing a local timestamp, so N counts feed batches rather than individual book updates.
Would an event-clock-first tutorial be useful, or would you prefer a brief trade-volume-threshold clock in the same notebook? Is anyone already working on this item?
I’d like to contribute a small tutorial for the unchecked volume-clock/event-clock roadmap item.
It would compare a 100 ms physical-time grid with sampling every N successful
wait_next_feed(False, timeout)feed wakeups on the same bundled BTCUSDT sample. It would record local timestamps, mid-price, spread in ticks and top-of-book imbalance, and compare sample spacing and feed activity.I found the replay loop in Fusing Depth Data and the references in Pricing Framework, but no dedicated comparison or competing open PR. I would reuse the bundled raw data and existing conversion/snapshot utilities, without core API changes or new dependencies.
The tutorial would explicitly explain that a wakeup can process several records sharing a local timestamp, so N counts feed batches rather than individual book updates.
Would an event-clock-first tutorial be useful, or would you prefer a brief trade-volume-threshold clock in the same notebook? Is anyone already working on this item?