Quote from: telepatheic on May 21, 2014, 03:23:52 PM

Quote from: ChuckOne on May 21, 2014, 02:11:45 PM

In any case, we would like to increase the uncertainty to 100% after n blocks where n << 1440.

We have approaches reaching from forking the network deliberately to using reverse hashchains to include random numbers into blocks. These random numbers get included into the generation signature and make the forger prediction impossible at a particular block height.

This is essentially copying peercoin's StakeModifier

If I sign a block at height n either the person who gets to sign the block (with first preference) at height n+m is known to me or unknown. If the person is known then there is nothing I can do to affect this. If the person is unknown but at height n+1 the person who gets to sign the block at n+m is known, that implies that something in block n decided who got to sign the block n+m. Since I have complete control over what is included in block n, I get to decide.

This applies to NXT and to peercoin. With NXT, the previous signer gets to choose who gets to sign the next one. However they only have 1 bit of control. They can either choose someone (signing the block) or choose not to sign the block and delegate the choice to the second preference signer who gets to choose someone (signing the block) or choose not to sign the block and delegate the choice to the third preference signer...

I told you, we use reverse hashchains. So, you as a forger are forced to publish a previously publicly unknown number. However, everybody can check if you published the correct number. So, there is no choosing in what you include in your block.

And, btw. who says that everybody should include random numbers into their blocks? Wink