Not M1. M2. Including long term savings accounts. I use my bitcoins as M2 today. Long term savings account. On a paper wallet 
Not M2. M1. Money on savings accounts are lended by banks to somebody and put to their current accounts. So, M2=M1+IOU. Bitcoin will replace real money, rather than IOUs. So, to calculate (maximal) future value of bitcoins we should use M1 or MB, rather than M2.I think that's right. We are going to see more and more off blockchain type transactions such as me using my CB account to buy something on overstock. This effectively increases the money supply. It doesn't even have to be fraudulent. There is nothing to prevent fractional reserve bitcoin banks if consumers want them.
).M0 are paper bills. They are real money in a sense that one dollar bill cannot be in two places simultaneously. There is no fractional reserve banking with M0. M1 is M0+current accounts. M1 are more-less real money too. The amount of real money is MB, but M1 is a good approximation of it.
The money that sits on saving accounts are not real money, because they exist in two or more places simultaneously. A dollar that sits on your saving account is at the same time lended to somebody else and sits on their current account too. So, money on your saving account are not real money, but just promises to get real money and put them on your current account. We can say that M2 = M1+IOU.
We can make similar definition for bitcoin. Bitcoin's M1 are bitcoins that exist only in one place - in blockchain. BTCM1. Currently there are about 13,000,000 of them.
As you've correctly pointed out, there are also promises of bitcoin, that sits off-chain. So we can define BTCM2 = BTCM1 +BTCIOU.
The whole discussion was what will bitcoin cost if it'll replace fiat?
The suggested formula was: bitcoin price = M2/BTCM1
I disagree with this formula. IMO, correct formula should be:
either bitcoin price = M1/BTCM1.
or bitcoin price = M2/BTCM2.
Problem with the second formula, however is that currently we have no idea how big BTCIOU are going to be.
I see what you are saying, and the second formula is problematic. Maybe it would be more helpful to attack the problem from the other end. Something is only worth what someone is willing to trade for it. If Bitcoin replaces a given market or industry, then if the replacement industry stays the same size, bitcoin is worth whatever the disrupted industry was worth prior to disruption. You could make a partial valuation based on the market share Bitcoin takes from the legacy industry or market, and then divide that number by the number of bitcoins in existence. This value is ADDED to the value Bitcoin has for every other use, so:
The value of the reserve currency used for settlements, peer to peer transactions on the blockchain, remittances, micropayments, titles (colored coins), provably fair voting, PM type store of value, machine to machine (IoT), escrows, and most likely killer apps we haven't even thought of yet. Bitcoin has the potential not only to capture a significant chunk of Gross World Product, but to actually GROW GWP in ways it otherwise wouldn't grow just by increasing financial efficiency.
