This metric can be found on the 'Ethereum & ERC20 MVRV' page in Sandata, along with the NVT metric which is described in this article.
To understand the MVRV metrics, we have to establish two term. 'MV' as in 'market value' simply describes the market cap, which is well known when looking at crypto assets. The second part is the 'RV' or 'realized value':
The realized value metric is calculating the acquisition cost of the assets located in a wallet. Imagine that there is a wallet which has 30 tokens and these tokens came into the wallet in 3 transactions:
The realized value of the address is
10 ∗ 5 + 5 ∗ 15 + 15 ∗ 10 = 50 + 75 + 150 = $275
This number gives the value of the tokens for this particular token holder and can be compared to the current market value. If the current price of the token is below $9.1, then the money this holder paid for acquiring these assets are more than their current value, while if the price is over $9.1, the value is greater. That means that if the current price is $10, these tokens are worth $300 and if the holder sells everything, he will get $25 profit.
The interesting part is that we can compute the realized value across the whole network, by summing the realized values of all wallets holding tokens at the moment. This number gives an estimate of the amount of money the users of the network spend to acquire their assets. The definition will be:
If we compare the current market capitalization (market value or MV) to the current realized value, we can get an estimate how overvalued or undervalued the current market cap is. The definition will be:
For example if the current realized value is $4b and the market value is $8b, that means that on average each participant will take 2x profits if he liquidates his tokens. The more this ratio increases, the more people will be willing to sell and take profits.
If the MVRV is below 1, then the market is “undervalued” on average, meaning most people will be realizing losses if they sell their holdings.
Keep in mind that this is in the ideal case and does not account for the addresses with lost private keys. The way to adjust for this is to look at the historical values for the MVRV ratio. As the ratio is approaching historical maximums or minimums, then the possibility of a highly overvalued or undervalued market is much higher.