BIP Idea: incrementalrelayfee in feefilter?

[Matthew Husák]

Hi all! 

Core 30.0 defaults both minrelaytxfee and incrementalrelayfee to 0.1 sat/vB now, but lots of nodes still run 1 sat/vB for incrementalrelayfee (or other settings).

Problem 1: You send RBF tx → your node accepts → peers reject replacements → dead end.
Problem 2: You can't see what min RBF fee peers have.

The obvious fix is to manually connect to peers with matching incrementalrelayfee, which is impossible because this policy isn't exposed over P2P. BIP133 feefilter only signals minrelaytxfee.

Proposal: Extend feefilter with incrementalrelayfee field (parallel to minrelaytxfee). Peers can then filter out un-bumpable RBF invs, reducing spam and improving propagation.

Benefits:

- spam protection: your node stops spamming peers with un-bumpable RBFs

- transparency: you can see what RBF policy other nodes use, thus you can find miner node which you can add as manual peer

- pure policy change, no consensus risk

WDYT guys?

Happy to prototype or draft a BIP if there's interest.

Cheers,

Matt.

[Murch]

Hi Matt,

The feerate policy change was rolled out not just with the Bitcoin Core
30.0 release, but also backported to Bitcoin Core 29.1. According to
e.g., Clark Moody’s dashboard, that means that over 30% of listening
nodes already use the lower incremental feerate (assuming they use the
default value). Per a quick calculation even non-listening nodes with
only eight connections should have an over 94% chance to have at least
one peer that accepts and rebroadcasts such a replacement
transaction—and rising.

Meanwhile, very few wallets so far support creating transactions with
lower feerates or implement replacements with the lower
incrementalrelayfee. Presumably the wallet adoption will continue to
trail the node policy adoption in this regard, exactly due to the
current unreliable state.

I am not convinced that special handling for the incrementalrelayfee is
necessary at this time, and either way, a more general approach such as
the recently proposed BIP 153: SENDTEMPLATE (PR:
https://github.com/bitcoin/bips/pull/1937) would feel like a bigger bang
for the buck in that regard.

Best,
Murch

On 2026-01-06 17:36, Matthew Husák wrote:
> Hi all!
>
> Core 30.0 defaults both minrelaytxfee and incrementalrelayfee to 0.1
> sat/vB now, but lots of nodes still run 1 sat/vB for
> incrementalrelayfee (or other settings).
>

> *Problem 1*: You send RBF tx → your node accepts → peers reject
> replacements → dead end.
> *Problem 2*: You can't see what min RBF fee peers have.

>
> The obvious fix is to manually connect to peers with matching
> incrementalrelayfee, which is impossible because this policy isn't
> exposed over P2P. BIP133 feefilter only signals minrelaytxfee.
>

> *Proposal*: Extend feefilter with incrementalrelayfee field (parallel

> to minrelaytxfee). Peers can then filter out un-bumpable RBF invs,
> reducing spam and improving propagation.
>

> *Benefits*:

> - spam protection: your node stops spamming peers with un-bumpable RBFs
> - transparency: you can see what RBF policy other nodes use, thus you
> can find miner node which you can add as manual peer
> - pure policy change, no consensus risk
>

> *WDYT guys?*

>
> Happy to prototype or draft a BIP if there's interest.
>
> Cheers,
> Matt.

> --
> You received this message because you are subscribed to the Google
> Groups "Bitcoin Development Mailing List" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to bitcoindev+...@googlegroups.com.
> To view this discussion visit
> https://groups.google.com/d/msgid/bitcoindev/d3ddaf1c-44c9-4ebd-98fe-88e03a84891en%40googlegroups.com
> <https://groups.google.com/d/msgid/bitcoindev/d3ddaf1c-44c9-4ebd-98fe-88e03a84891en%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Matthew Husák]

Hi Murch!

thanks for such quick response.

I think that we've been one of the first wallets to support such lower fees, so we faced the initial "pain" to find peers with the lower minrelaytxfee. That kinda made me feel that this has to change, because when we solved the peers config (with our desired minrelaytxfee setting) we realized that many nodes (before the update)  forgot to change the minincrementalfee policy and since you can't check for peer's minincrementalfee it makes the adoption even harder, because it's almost impossible to find peers on your own (you could sign a RBF tx bumped up to fee under 1sat/vb and try to send it to different peers, but thats far from ideal :-D).

I'm not sure if lowering the defaults will ever happen again, but I believe that you should be able to see under what conditions other nodes relay, so that you don't spam them with tx they don't want and so that the adoption is easier.

I haven’t seen BIP 153 before, so I’ll read it and try to understand it. That said, I have two ideas: the simplest is to expose  incrementalrelayfee , and the second (I’d love your opinion on this) is adding the ability to configure the node to actively seek peers meeting specific  minrelaytxfee  and  minincrementalfee  thresholds—since AFAIK it currently settles for ~10 outbound nodes of any config, this would let it target at least 2 high-quality peers with those minimums during discovery.

(or I could be totally off with my ideas :-D but I hope its not that case)

Thanks!

Matt

Dne středa 7. ledna 2026 v 20:54:11 UTC+1 uživatel Murch napsal: