# Structural Break Real-Time: may infer use prior completed-series state? **URL:** https://forum.crunchdao.com/t/structural-break-real-time-may-infer-use-prior-completed-series-state/1186 **Category:** ADIA Lab **Created:** [August 7, 2026, 6:51am UTC](https://forum.crunchdao.com/t/structural-break-real-time-may-infer-use-prior-completed-series-state/1186 "2026-08-07T06:51:07Z") **Posts on this page:** 1 **Showing post:** 4
### Author: ![enzo](https://forum.crunchdao.com/user_avatar/forum.crunchdao.com/enzo/32/15_2.png) [@enzo](https://forum.crunchdao.com/u/enzo) #### Post date: [August 13, 2026, 6:15pm UTC](https://forum.crunchdao.com/t/structural-break-real-time-may-infer-use-prior-completed-series-state/1186/4 "2026-08-13T18:15:34Z")
Hello Ionut, I actually wrote a long explanation here: [Leaderboard comparability after the June 8 real-time data-access fix — were pre-fix scores rescored? - #2 by enzo](https://forum.crunchdao.com/t/leaderboard-comparability-after-the-june-8-real-time-data-access-fix-were-pre-fix-scores-rescored/1188/2#p-2369-the-sharing-of-state-2) In short, trying to persist state across time series will likely result in your code not being deterministic based on when it starts. - Depending on the number of `INFER_PARALLELISM`, if you run the full dataset, the runner will start at regular intervals. - However, during the determinism dataset, those offsets will be unique. Here is a visual representation: ![image](https://forum.crunchdao.com/uploads/default/original/2X/d/dfd9f2ca2980bf9e513fe4dc1b545c31f4082523.png) (Feedbacks are welcome as I am not the best designer.) There seems to be a general misunderstanding of this mechanism. I should add it to the documentation. * * * P.S.: It’s only 10%, not 30%. I recently found out that the documentation was outdated, but the change happened a month after the competition started. --- _[View the full topic](https://forum.crunchdao.com/t/structural-break-real-time-may-infer-use-prior-completed-series-state/1186)._