A threshold signature is a digital signature that may have been created by an authorized subset of the private keys which were previously used to create the corresponding public key. Threshold signatures can be verified using only a single public key and a single signature.

Different algorithms exist for creating threshold signatures, but perhaps the simplest of these is a slight extension of a typical algorithm for creating a scriptless multisignature. This is easiest explained with an example: participants A, B, and C want to receive funds that can be spent by any three of them. They cooperate to create an ordinary multisignature public key for receiving the funds, then they each take the extra step of deriving two secret shares from their private key—one share for each of the other two participants. The shares are created so that any two shares can reconstitute the private key that created it. Each participant gives a different one of their secret shares to each other participant, so each participant ends up having to store their own private key plus one share for each other participant. Each participant is then able to verify that the shares they received were correctly derived (i.e. not fake) and are unique from the shares received by the other participants.

Later, A and B decide they want to create a signature without C. A and B share with each other the two shares they each have for C. This allows them to reconstitute C’s private key. A and B also have their own private keys, giving them all three of the keys necessary to create a multisignature.

The example above only covers the simplest of threshold signature algorithms. Other algorithms exist that reduce conceptual simplicity but provide additional features, such as reducing the number of steps required or providing increased resistance to communication problems.

Comparison to multisig scripts

Bitcoin’s Script language (including the proposed tapscript modified alternative) allows providing a threshold k of signatures for a group of n keys, commonly called k-of-n multisig. This requires providing k signatures and n public keys in any onchain transactions.

By comparison, threshold signatures only require a single public key and single signature, no matter how many participants are involved. This can significantly reduce the size of transactions, correspondingly reducing the cost of their transaction fee. It also increases their privacy: nobody can tell which of the parties signed (or even that multiple parties needed to sign in the first place).

Signer privacy does create a problem for schemes that want third-party auditability of which parties signed. Auditing can be implemented using an independent system (e.g. having all communication between participants go through a logging server). It can also sometimes be implemented using clever constructions, such as for a 2-of-3 threshold scheme in taproot where the usual signers (A and B) can use a 2-of-2 multisignature keypath spend and the two alternatives (A and C, or B and C) can be their own 2-of-2 multisignatures in known positions in the merkle tree for scriptpath spending. By looking at the spend, the participants can determine exactly which two parties signed.

Terminology

The following table summarizes the differences between threshold signature and related terms.

Term Private keys Messages
(e.g. tx inputs)		 Published pubkeys Signatures Signers required Notes
Scripted multisig m 1 m k where k<=m k Uses Bitcoin Script multisig opcodes
Scriptless multisignatures m 1 1 1 m Indistinguishable onchain from single-sig
Threshold signature m 1 1 1 k where k<=m Indistinguishable onchain from single-sig

Optech newsletter and website mentions

2024

2023

2022

2020

2019

2018

See also

Previous Topic:
Testnet

Next Topic:
Time warp

Edit page
Report Issue