Looking for average TXN cost per UTXO

Got into a flame war earlier today with someone complaining about TXN cost looking at BTC stats. I was explaining that the “average TXN cost” includes a lot of exchange traffic that are massive 50-in / 50-out TXNs. Then it occurred to me that a better stat would be TXN cost / UTXO. So a 50-in / 50-out would have 100 UTXOs (50 spent, 50 created). Using that logic, the “cost per UTXO” would be:

all_fees_in_block / (all_utxos_spent_in_block + all_utxos_created_in_block)

I might dust off my python IDE and give it a shot, but was curious if anyone is already compiling this stat. Obviously I would ignore the [coinbase TXN](

View Reddit by brianddkView Source

Leave a Reply

Your email address will not be published. Required fields are marked *

GIPHY App Key not set. Please check settings


  1. These might be of interest to readers here:



    For comparison to blockchains that don’t use UTXOs but rather a single account per user, one difference between the two designs is that for a merchant, if they want to identify individual transactions then they can’t simply request payment to their account, since they won’t know which payee is which (unless they request off-chain the payee’s account number), so will instead request payment to a unique account for that transaction/customer and then later would need to, one by one (since there is no way to do a bulk transfer) pay for additional transactions if they want to merge their account balance, rather than manage many separate wallets/accounts. Two alternatives do exist; asking the payee to call some contract method, incurring considerably more expense and block space usage than a simple transfer, or using the memo field to include some extra identifier (I’m not sure how well supported this is by wallets).

    UTXOs were a great design choice.

  2. I don’t know about the availability of these statistics, but aren’t you immediately claiming that all fees are actually half as much by counting a simple 1 to 1 transaction as two transactions? Also, wouldn’t you be counting each half transaction twice by counting it both where it’s created and where it’s spent? Maybe I’m misunderstanding.

  3. No that wouldn’t work for two reasons. First of all Inputs (or consumed UTXOS) are not a great stat to use. Alice paying Bob might consume 1 UTXO or 10, depending on Alices UTXO-Set. However both are a single „payment-event“.

    Secondly most payments create two UTXOs. One for the recipient and one for change.

    So while it’s a good idea, it won’t work that way

  4. The fees are also dependent on the address format, e.g., [native segwit vs legacy]( Why make up a new metric when there is already an established one? Every transaction pays for the space it uses on the Blockchain in Satoshis per vByte.

    At the moment you can make transactions for 1 sat/vByte ([](, so a simple transaction is <$0.10. Sounds pretty cheap to me.

    edit: I think I misinterpreted what you wanted to achieve. Ignore me.

What do you think?

Crypto CEOs and Their Testimonies During Today's Congressional Hearing

Crypto CEOs and Their Testimonies During Today’s Congressional Hearing

Macro Guru Raoul Pal Owns XRP, Says Crypto Asset Has Phenomenal Risk-Reward Despite Ripple Lawsuit

Macro Guru Raoul Pal Owns XRP, Says Crypto Asset Has Phenomenal Risk-Reward Despite Ripple Lawsuit