Rebased metric

A useful metric for Tezos is by how much some tez held in the genesis block grow over time through block rewards, endorsement rewards, fees etc. To make this precise, I would suggest the idealized definition:

The index starts with 1 tez in the genesis block. On each block, it accrues a fraction of the total reward for the block (including block reward, endorsements, nonce revelation tips, transaction fees, and denunciation) determined by the ratio of the index to the total delegated amount at the time of the snapshot used to determine baking rights. These rewards accrue to the index as they would in the protocol.

The index is an approximation of how 1 tez at genesis grows from baking. It ignores the randomness of the block allocation, the coarseness of rolls because it’s an idealized metric that measures an average. For pragmatic reasons, it also ignores any transaction fees paid off-band or baker extractable value which are hard to quantify.

Using this index, one can compute the “rebased supply” which is the total supply of tez divided by the index.

By that metric, the “rebased supply” of tez has steadily gone down since the genesis block. Graphs showing the evolution of the rebased supply could be a good educational tool to explain that, up to a simple change of denomination, the network is effectively deflationary, not inflationary. That denomination would be a rebased tez (that’s index * 1 tez) and a price graph of a “rebased” tez in some other currency can illustrate that phenomenon as well. Note that so-called “market cap” is unaffected, as (rebased supply) * (rebased price) = (supply / index) * (index * price) = supply * price.

I’m using the word “rebase” because it has come to get some popularity in the defi world. I would prefer to say “change of numéraire” which has more cachet but if the goal is education, it’s better to go with something that’s more likely to ring a bell.

In general this can be calculated by block explorers and indexers. I tend to think that non-consensus critical metrics have no place in the protocol (because they are not needed and can be computed by external tools), and this may apply to this one as well, although the fact that it’s super easy to track and just a single number might mean it’s worth making an exception.