Scheduled (Delayed) Transaction Broadcast

The discussion revolves around the comparison between specifying target times in blockchain transactions using block height versus physical time, particularly focusing on their applicability and accuracy within different time frames. The analysis draws from empirical data collected through log parsing, which examined the variance in block production over set periods. It was found that for a three-hour delay, there were significant deviations from the expected average of 18 blocks, with instances where the actual number of blocks produced either significantly exceeded or fell short of this average.

Specifically, the data indicated that there were 199 occasions where more than 24 blocks were generated within a three-hour window and 255 instances where fewer than 12 blocks were produced in the same timeframe since October 24th. This variability highlights the potential challenges in timing transactions accurately when relying on physical time due to the unpredictable nature of block generation rates.

Furthermore, the communication touches on the implications of these findings for transaction scheduling. It suggests that while it is technically feasible to pre-sign transactions for future broadcasting, doing so based on physical time could lead to issues. Transactions might fail to broadcast if attempted too early, absent a mechanism for retrying. This problem is mitigated by using block height as a reference for nLockTime, given the more predictable nature of block production compared to the variable rate of block generation observed.

Therefore, the argument concludes with a recommendation favoring the use of block height over physical time for scheduling transactions. This approach is advocated despite the possibility of setting nLockTime to 0, should the sending wallet support such functionality, due to the inherent advantages in predictability and reliability that block height offers in ensuring transaction timeliness within the blockchain framework.