RE: A new puzzle about daily STEEM production (yes, really.)
Missed blocks?
That's it!
I had gone back to my records yesterday and wondered about the number of blocks.
I had recorded from block 76441062 to block 76456642 (15581 blocks). During this process 46941 seconds passed. But with the number of blocks it should have been only 46743 seconds.
With your hint, I looked at the data again and actually found missed blocks in that time. There were exactly 66 blocks missing. I would never have thought that was possible. Thinking about it, it occurs to me that justyy had pointed out something similar recently (He called it Down Time).
If I include these in my calculations of data accrued in the time period, the predicted and calculated values differ by only about 150 STEEM. Given the rounding along the way, that's pretty close.
If a block gets missed, do those STEEM get added to a later block, or are they just gone?
I think the STEEM are gone. There is no hint in the values of a stronger increase of the virtual_supply.
the disabled hive witnesses are helping us by reducing inflation?
You can say that!
On top of that, the witnesses who don't produce blocks are usually not the top witnesses. The witnesses from rank 21 get a higher reward. This higher reward drops, if an "elected witness" do not produce the block. Usually the lower rewards of the top witnesses and the higher rewards of the elected witnesses balance out after a full round. But if the higher rewards are dropped, the inflation is even lower.
The ones that still have valid signing keys are still getting scheduled.
Yes, but until now I thought that the next witness would produce the block if the current one wasn't fast enough or didn't respond at all. That in that case no block is produced at all was not clear to me until now. One more argument to reactivate the mechanism deactivated with HF20, isn't it?
Or should we keep it for the sake of inflation? ;-)
When the witnesses page showed them as disabled
There are different approaches to this. It depends on the frontend. On Steemworld, it only depends on the key. Steemwallet also considers the last generated block. For the blockchain, according to the current HF, they are only disabled if an invalid key was registered.
Phew, that was quite elaborate, but I am very pleased that we could clarify some things here :-)
I didn't even know that had been deactivated! I was wondering why/how those witnesses still had valid signing keys. Now, this makes me wonder if they were making contingency plans for the hive hardfork and subsequent attacks on the blockchain as far back as 2018. I'd be especially suspicious if it turns out that hive has reactivated that mechanism... (but I'm not curious enough to find their fork in github and search it down. ;-)
The inflation is a nice side-effect, but overall, I think I'd prefer not to have the missed blocks. As they said in that conversation, very few stakeholders are going to be monitoring witnesses for missed blocks.
I posted another comment suggesting that Steemit might want to remove one of their witness votes. If that vote gets removed, we should be able to confirm, observationally, if there is a corresponding increase in daily STEEM production.