For example, how many times a month would you expect a 2 block reorg? What about for deeper reorgs?
1
-
Tezos doesn't have reorgs happen that often. So, 0 per month is the expectation. The protocol does not allow reorganizations longer than 5 cycles ago. It's an extremely rare phenomenon in Tezos. I believe the new tenderbake protocol will introduce 2-block finality.– utdrmacJun 21 at 17:37
Add a comment
|
Basically reorgs do not seem to happen on mainnet, as far as I know.
However, theoretically they can happen. Using the analysis from this blog post, for an attacker having 20% of the total active stake, we get that currently (with Emmy+) 2-block reorgs may a priori happen around 8 times per month. With the new protocol Granada, which upgrades the consensus algorithm to Emmy*, such reorgs would only happen once every 90 years.
The raw data: probability of a fork of length 2 per block
- in Emmy+ is 0.000183514; therefore #reorgs per month = 0.000183514 * 43200 = 7.92, as we can assume one block per minute, thus 43200 blocks per month;
- in Emmy* is 1.0888e-08; therefore 1 reorg happens once every 1 / (1.0888e-08 * 2 * 60 * 24 * 356) years = 89.58 years, where we assume 2 blocks per minute.
