V3/TRUC miner adoption

[Peter Todd]

FYI I ran a short experiment with my Alice OpenTimestamps calendar to see what
% of hash power was mining nVersion=3 (TRUC) transactions. I was upgrading the
Bitcoin node on the Alice calendar to v29.0 and was curious.

The following transactions were broadcast with nVersion=3 over the course of
about 20 hours:

27582aa5aaeb8c5f3772b9a892ec5bb6cf81a5c3e41a6b4c68c932809ca897ba
f039fa1b056b1446c822d2a96a41a11bd0f53511753d98141c1a4f6387a2c1c3
b409e329e3d9ab18703278de602f7275323e72c0dc4ff5807b7b8493aa5f3afd
e91e0f0604a037d24f89cef2f4e4697880f5a9a5305cb3f321e86fc620835d1b
647345eb10b10965e60dde320d107edaadeb7477a6ccff84005c147a8e06a5e1
c9a708aab754f5eb9ff1d5e8c27ce3131a869f2aafc42db5367728531d9b49df
f31d5aad9a1cb6b63ecaacb5f26e20e429eb4f342880a225c50508da11ae2ff4
f34821dbeb7aefca30a9cab404f6466ef087730afdc88608a06aeb0e23b95b24
210673e41dd0bcd9243d04e5af7d58772492b0206f6e11d4ee79e4324bc9ba2c
9bb821cb44e136c8b69e4eed47571b97ca27f76ad6be9a08286c582025bef232
f2f66f8fe4da2d08a1c6ffbfeae1ad1c74dbecd44da36fb81ef09642bf54b75a
5033d1e0af974d71d26107f564c3c7299848fd3de5dbace2902dd7a3ee808e20

...every transaction was mined by either Foundry USA (28% hashpower), F2Pool
(10%), or MARA (8%).

Notably, Antpool does not appear to mine V3 transactions - it missed multiple
opportunities at high fee-rates. Presumably that's true for Antpool proxies as
well.

I don't know if Foundry USA or F2Pool is actually following the TRUC
standardness rules; I suspect MARA is not based on my discussions with them
about how their backend works.

It's possible that poor propagation is the issue here. But up-to-date Libre
Relay nodes also relay V3 transactions, and they're currently quite well
connected. I'm sure Antpool runs enough nodes that if they accepted V3, they'd
be getting them.

This was just a short test so there's a good chance I missed some smaller
miners who would have mined V3/TRUC transactions; given the lack of hash power
mining V3 I'm going to switch it back later today to avoid delaying timestamps
unnecessarily.

If this information was useful to you, feel free to send some sats to pay for
OTS transaction fees:

https://geyser.fund/project/opentimestamps
https://alice.btc.calendar.opentimestamps.org/

--
https://petertodd.org 'peter'[:-1]@petertodd.org

[Peter Todd]

On Wed, Jun 04, 2025 at 11:35:37AM +0000, Peter Todd wrote:
> FYI I ran a short experiment with my Alice OpenTimestamps calendar to see what
> % of hash power was mining nVersion=3 (TRUC) transactions. I was upgrading the
> Bitcoin node on the Alice calendar to v29.0 and was curious.
>
> The following transactions were broadcast with nVersion=3 over the course of
> about 20 hours:
>
> 27582aa5aaeb8c5f3772b9a892ec5bb6cf81a5c3e41a6b4c68c932809ca897ba
> f039fa1b056b1446c822d2a96a41a11bd0f53511753d98141c1a4f6387a2c1c3
> b409e329e3d9ab18703278de602f7275323e72c0dc4ff5807b7b8493aa5f3afd
> e91e0f0604a037d24f89cef2f4e4697880f5a9a5305cb3f321e86fc620835d1b
> 647345eb10b10965e60dde320d107edaadeb7477a6ccff84005c147a8e06a5e1
> c9a708aab754f5eb9ff1d5e8c27ce3131a869f2aafc42db5367728531d9b49df
> f31d5aad9a1cb6b63ecaacb5f26e20e429eb4f342880a225c50508da11ae2ff4
> f34821dbeb7aefca30a9cab404f6466ef087730afdc88608a06aeb0e23b95b24
> 210673e41dd0bcd9243d04e5af7d58772492b0206f6e11d4ee79e4324bc9ba2c
> 9bb821cb44e136c8b69e4eed47571b97ca27f76ad6be9a08286c582025bef232
> f2f66f8fe4da2d08a1c6ffbfeae1ad1c74dbecd44da36fb81ef09642bf54b75a
> 5033d1e0af974d71d26107f564c3c7299848fd3de5dbace2902dd7a3ee808e20
>
> ...every transaction was mined by either Foundry USA (28% hashpower), F2Pool
> (10%), or MARA (8%).
>
> Notably, Antpool does not appear to mine V3 transactions - it missed multiple
> opportunities at high fee-rates. Presumably that's true for Antpool proxies as
> well.

Right after I sent that email AntPool mined multiple V3 transactions:

774ed3c3760a859bf1cc4f6b11829807b19cacbe3045348694f1b6deff13720b
a1f5d87ad33fb38b2bf950ae712cbcefb0f91f8e951c4f6bb2e825af88ae25a9
a4c3637cddb10ab3b99a7a27c5356952a1a756ccfa4cf06569d8de9c351415a9
1449ab8fb32e3c7664276f8c6e320e478ff44e1599c44d3dfb6ec6607c45ef26

Mining pools often split up their block template creation across multiple
nodes, so it's quite possible that I just got unlucky. Or maybe someone at
AntPool noticed my email and upgraded one or more of their nodes.

Based on the available data all I can say for sure is that ViaBTC, SpiderPool,
Binance Pool, BTC.com, and Luxor seem to *not* be mining V3:

a4c3637cddb10ab3b99a7a27c5356952a1a756ccfa4cf06569d8de9c351415a9
e6469be8344aabe4758fc420279e764e784f6371f9df0a8fb1e463809dd765cb
35a49bf34f5af21d295a1d19d5d557251f6caa02e640457c73f921ddd9186ecb

It'd be interesting to compare these findings to the accusations that many
other pools are just a proxy for AntPool... I've got better things to do with
my time right now. But I'll leave the experiment running for a bit longer. :)