freeloading attack + threshold

[dejavukong]

Background

We use the on-chain aggregation to prevent freeloading attacks with the commit/reveal pattern.

Problem Description

For instance, for some proof, let's say proofA, it needs 10 verifications to decide this proof is valid or not.

Since the commit/reveal pattern needs 2 txs per verification per worker nodes, what if in stage commitment, there are exactly 10 commits submitted while in the reveal stage, there are only 8 reveals(can not reach the threshold).

Then what we are gonna do with this proof-verification task? partly re-open or do it right up from the beginning? The pros and cons for these 2 solutions are quite obvious, partly re-open is not perfectly safe because after it is revealed, the freeloaders then have a chance to cheat. do-it-again is safe but expensive for verification node operators