PaperlayerView exact revision on arXiv

Source-linked AI summary

On-Chain Archaeology of Bitcoin Oracles: Evidence of Use under Limited Observability

Giulio Caldarelli

arXiv:2610.11439v1cs.CRcs.CYcs.IR

TL;DR

Bitcoin oracle use is difficult to compare because activity is unevenly documented and contract designs disclose different evidence. This study combines transaction, ledger, and public DLC records, finding that both observed activity and its interpretation depend on design and source preservation.

  • Problem

    The scale, distribution, and applications of Bitcoin oracle use remain unevenly documented, and different designs make evidence of activity difficult to observe and interpret.

    Source: §1 · p. 2

  • Method

    The study combines a Counterparty betting census, full-chain analysis, searches for documented service keys, and DLC records recovered from an archived explorer and public Nostr relays.

    Source: §1 · p. 2

  • Results

    Counterparty activity was concentrated in a small number of services and applications, while measures across systems depend on public-key roles, transaction disclosures, and available external records.

    Source: §1 · p. 2

  • Takeaways & Limitations

    The findings distinguish indicators of oracle-information provision from indicators of consumption and highlight preserving event descriptions and service documentation alongside transaction records.

    Source: §1 · p. 2

    Topic shares depend partly on update frequency, and the recurring Deribit price series strongly shapes the Suredbits archive distribution.

    Source: §5.1 · p. 20

Abstract

from arXiv · show

Before Ethereum made the "oracle problem" a household term, Bitcoin already had oracles serving as feeds, key-release services, federated signers, and arbiters that carried real value on the main chain. This study traces their use and the changing evidence of oracle activity from early days through July 2026. We combine a complete census of Counterparty betting (1,149 bets), analysis of the full Bitcoin chain through block 958,628, and searches for documented keys from Reality Keys, Orisi, Bitrated, and Oraclize in an 854-million-row public-key index. We also recover DLC oracle records from an archived explorer and live Nostr relays. Two results emerge. First, early contracts remain on-chain, but many event descriptions have disappeared, and protocol encoding and API limitations complicate access to the surviving record. However, for modern DLCs, public oracle announcements can survive even when the contracts using them cannot be identified on-chain. In the script classes examined, the share of spends that reveal no script peaks at 81.9% in 2024 after excluding spends containing inscription data. Second, public registries can give a misleading picture of oracle use. In Counterparty, 95% of pre-2018 sources declaring an oracle fee were never bet on. In Bitrated, 0.1% of archived keys appear on-chain overall, compared with 10 of 19 keys captured in 2014. Sport dominates Counterparty's matched volume, while a daily price series dominates the archived DLC announcements. These findings show how protocol design and data preservation shape the historical record of Bitcoin oracle use.

1. Introduction

Bitcoin oracle research has documented designs and trust trade-offs, but the scale, concentration, and applications of recorded use remain unevenly known. This study compares activity records across early services and DLCs while examining how design and source availability shape what can be observed.

1. Introduction

  • The study asks what available records reveal about Bitcoin oracle use and how contract design and historical sources shape its measurement.The stated dimensions are scale, concentration, and applications.Source: §1 · p. 2
  • The analysis combines a Counterparty betting census, Bitcoin-chain analysis, documented-key searches, and recovered DLC records from an archived explorer and Nostr relays.It also uses historical documentation and archived websites to interpret the records.Source: §1 · p. 2
  • Counterparty activity was concentrated in a small number of services and applications, with modest fee revenues and limited matching of offered value, although most matched contracts settled.Across systems, activity measures depend on public-key roles, transaction disclosures, and available external records.Source: §1 · p. 2
  • DLC announcements and attestations extend the public record of oracle-supplied information beyond bets and settlements visible in Counterparty.The paper also highlights preserving event descriptions and service documentation alongside transaction records.Source: §1 · p. 2

2. Background and related work

Prior work examines oracle trust, security, and implementation, as well as Bitcoin designs ranging from signing services to prediction markets and privacy-preserving DLCs. Transaction histories and external documentation can each leave gaps in reconstructing the applications and events behind oracle activity.

2. Background and related work

  • Oracle research examines trust, security, and implementation choices, including the vulnerabilities and assumptions involved in obtaining and reporting external information.Bitcoin designs include federated signing, authenticated data, key release, and Counterparty broadcasts with protocol-based settlement.Source: §2 · p. 3
  • Prediction-market work explores decentralized trading and arbitration, while DLCs let parties use oracle signatures without publishing contract terms.DLC privacy also makes their use difficult to identify from transactions (Dryja, 2017).Source: §2 · p. 3
  • Transaction analyses connect records to known services or metadata, but reconstructing oracle contracts may also require event descriptions published outside the blockchain.Online sources can disappear or change, making recorded transactions harder to link to their events.Source: §2 · pp. 3–4

3. Data and methods

The methods combine Counterparty, full-chain, and DLC publication records with historical documentation to measure oracle activity and assess how records can be observed and interpreted.

3. Data and methods

  • The study uses blockchain transactions, the Counterparty ledger, and public DLC oracle records, supplemented by historical documentation and archived websites.The three data sources support analyses of transaction and ledger activity, and oracle announcements and attestations.Source: §3 · p. 4, Table 1 · p. 4

3.1 The Counterparty ledger

  • The Counterparty dataset contains 1,149 bets, including 1,097 protocol-valid submissions, and 397 matches reconstructed from ledger states.Subjects were classified using broadcast text, feed URLs, archived catalogues, and contemporary documentation; declared deadlines were compared with the 2014 World Cup calendar.Source: §3.1 · pp. 4–5, Figure 1 · p. 5

3.2 The Bitcoin blockchain

  • The Bitcoin analysis covers blocks 0–958,628 and indexes 854,077,402 public-key occurrences for exact matching against documented service keys.It records revealed script classes and defines an annual opacity index, with a second version excluding script-path spends containing the Ordinals envelope marker.Source: §3.2 · pp. 5–6

3.3 DLC oracle publications

  • DLC records were collected from the archived Suredbits explorer and Nostr relays, yielding 911 distinct production-domain announcements and 219 distinct Nostr messages.The Nostr messages comprise 167 announcements and 52 attestations; Suredbits category shares were also recalculated without the recurring Deribit-BTC series.Source: §3.3 · pp. 6–7

3.4 Scope and interpretation

  • The study includes Bitrated arbitration but excludes wallet-security-only co-signers and custodial gambling platforms; OpenBazaar was not examined because moderator keys were unavailable.It compares activity, value, concentration, and application domains, and considers how design and historical sources affect identification and interpretation.Source: §3.4 · p. 7

4. Results

The results document oracle activity across Counterparty betting, transactions associated with identifiable Bitcoin services, and DLC oracle publications.

4. Results

  • The results cover Counterparty betting, transactions associated with identifiable Bitcoin services, and DLC oracle publications.Source: §4 · p. 8

4.1. Counterparty Betting activity and market structure

Counterparty betting was brief, sport-heavy, and concentrated among a few services. Later broadcasts continued without new bets, while access to records and event descriptions constrained reconstruction.

4.1.1 Activity, applications and concentration

  • 1,105 of 1,149 recorded bets (96.2%) arrived in 2014; June alone contributed 465 (40.5%), and none appeared after 12 June 2017 through July 2026.Figure 2 tracks monthly bet submissions.Source: §4.1.1 · p. 8, Figure 2 · p. 9
  • Sports accounted for 90.5% of matched XCP value, followed by price feeds at 8.3% and gambling feeds at 1.2%.Figure 3 reports shares of matched XCP by feed category; percentages refer to matched XCP value.Source: Figure 3 · p. 9, §4.1.1 · p. 9
  • The 82 matched xbet.io feeds represented 87.6% of matched value, and xcpfeeds.info accounted for 10.6%, together totaling 98.2%.Figure 4 presents matched escrow by operator and a Lorenz curve across 98 matched feeds.Source: Figure 4 · p. 10, §4.1.1 · p. 10

4.1.2 Matching, settlement and fees

  • Only 2,625 of 7,158 XCP offered were matched (36.7%); 387 of 397 matches (97.5%) were resolved.Expired or canceled matches returned committed funds through the protocol's refund mechanism.Source: §4.1.2 · p. 10
  • Of 2,965 pre-2018 sources declaring a non-zero fee, only 153 (5.2%) were referenced by a bet; total fees were 23.7 XCP.Figure 5 relates broadcast sources to betting activity.Source: §4.1.2 · p. 11, Figure 5 · p. 11

4.1.3 Broadcast activity and observability

  • From 2022 to July 2026, 77,108 of 80,066 broadcasts (96.3%) contained BTNS commands, while numeric feed publications continued without new bets.BTNS activity peaked at 51,220 messages in 2023; the period also included 1,123 numeric feed publications from 129 sources.Source: §4.1.3 · p. 11
  • The July 2026 Wayback search found no captures for 130 xbet.io feed URLs, 14 betxcp.com URLs, or 11 xcpfeeds.info endpoints; direct database queries recovered 1,149 bets when Core's /v2/bets endpoint returned none.Reconstructing the history required both underlying protocol records and external descriptions of events.Source: §4.1.3 · p. 12, Figure 6 · p. 12

4.2 Bitcoin blockchain analysis

The chain analysis links documented oracle keys to selected transactions while showing how transaction formats, shared scripts, and spending paths constrain identification. Taproot key-path spending sharply reduces script disclosure in the selected population.

4.2.1 Scripts and protocol identification

  • The analysis found 450,015,065 spends revealing scripts in the examined categories, including 316,060,091 multisignature scripts; data-carrier placeholders limited recurrence as an oracle identifier.The public-key index held 11,606,109 distinct values with at least five occurrences, while its three most frequent values each appeared 6,375,188 times.Source: §4.2.1 · pp. 12–13
  • Counterparty’s formats shifted from bare multisignature in 2015 to OP_RETURN in 2017, when 99.94% of ledger transactions used it; a literal CNTRPRTY search found only 333 of 3,146,510 ledger transactions.ARC4 obfuscation and multiple output formats complicate marker-based recognition; matching transaction identifiers located 2,961,933 records, or 94.13%.Source: §4.2.1 · p. 13

4.2.2 Transactions associated with oracle services

  • Documented-key matching identified transactions for Orisi, Oraclize, and Bitrated, while none of the recovered Reality Keys matched the multisignature index.Orisi’s seven matched keys appeared in 17 transactions; Oraclize’s key appeared in 106, and 11 Bitrated profile keys appeared in 56 transactions.Source: §4.2.2 · pp. 13–14, Table 3 · p. 14, Figure 7 · p. 14

4.2.3 Script disclosure and limits of attribution

  • Excluding inscription-related spends, Taproot key-path shares reached 81.9% in 2024, versus 48.6% including them; the selected population also includes ordinary payments.The series was 76.2% in 2025 and 70.1% in the partial 2026 record after exclusion.Source: §4.2.3 · pp. 14–15, Table 4 · p. 15, Figure 8 · p. 15
  • Among 45,860,478 spent 2-of-2 P2WSH outputs, the two Lightning indicators identified 772,269 spends (1.68%), leaving 98.32% unattributed.The two methods identified 482,165 publicly announced channels and 290,104 further spends; 197,739 were identified by both.Source: §4.2.3 · p. 15, Table 5 · p. 16, Figure 9 · p. 16

4.3 DLC oracle publications

The recovered DLC records show distinct publication patterns across the Suredbits archive and Nostr relays. Topic comparisons vary with the recurring Deribit-BTC series and with the mix of tests and delivery-related messages.

4.3.1 The Suredbits explorer archive

  • The Suredbits archive yielded 911 distinct production-domain announcements, 673 (73.9%) linked to recovered attestations, with 649 (71.2%) concerning price and finance.The Deribit-BTC series contributed 597 announcements (65.5%); without it, sport was the largest category at 54.1% of the remaining 314 announcements.Source: §4.3.1 · pp. 16–17, Figure 10 · p. 17

4.3.2 Nostr relay publications

  • Nostr relays returned 167 announcements and 52 attestations; tests or demonstrations accounted for 95 announcements (56.9%), while delivery or return outcomes accounted for 33 (19.8%).The decoder parsed 165 announcement and 49 attestation payloads; all 214 decoded messages had an embedded oracle key matching the Nostr author key.Source: §4.3.2 · pp. 17–18

4.3.3 Comparison across sources

  • Across collections, sport represented 83.5% of Counterparty bets with identified feeds and 90.5% of matched XCP value, while Suredbits’ price-and-finance share fell from 71.2% to 16.6% when its recurring series was excluded.Counterparty contributes bet counts and matched value, whereas the DLC columns describe announcement counts.Source: §4.3.1 · p. 17, §4.3.3 · p. 18, Table 6 · p. 18

5. Discussion

The discussion distinguishes oracle availability, contract participation, and settlement, and emphasizes that historical interpretation depends on preserved service context. It also argues that contract design and publication records affect what researchers can measure.

5.1 Patterns of oracle use

  • Only 36.7% of offered Counterparty value was matched, but 97.5% of matches settled; two services accounted for 98.2% of matched value.The findings separate feed availability, participation, and settlement, and the recorded fees indicate modest realized rewards for oracle provision.Source: §5.1 · p. 19

5.2 Interpreting the historical record

  • Service keys, protocol versions, event descriptions, and service interfaces are useful records to preserve alongside transaction and publication data for later interpretation.Recovered keys connect transactions to implementations, while archived websites and developer accounts help explain their roles and operating conditions.Source: §5.2 · pp. 20–21

5.3 Contract design and the measurement of use

  • Shared 2-of-2 scripts alone cannot identify the remaining activity as oracle contracts, while Taproot key-path spending withholds scripts and reached an 81.9% share in 2024 after inscription-related spends were excluded.DLC announcements and attestations document oracle publications, but comparisons should distinguish those records from identifiable contract transactions and committed value.Source: §5.3 · pp. 21–22

6. Conclusion

The study reconstructs Bitcoin oracle activity from early services through later DLC publications, finding that contract design and data preservation shape the evidence available for measuring use. It recommends linking public DLC records to voluntarily shared contract records to study which events are used in funded contracts.

6. Conclusion

  • The Counterparty census documents 1,149 bets from 2014–2017, with matched value dominated by two services; key matching identified transactions associated with Orisi, Oraclize, and Bitrated.Suredbits and Nostr records extend the evidence to later oracle publications, while sport and financial prices recur alongside testing and experimentation.Source: §6 · p. 22
  • Linking DLC announcements and attestations to contract records voluntarily shared by providers and participants could show which events appear in funded contracts and how committed value is distributed.The conclusion also identifies additional services, publication channels, and more detailed event descriptions as directions for further historical comparison.Source: §6 · p. 22

Appendix A. Currency conversion references and calculations

Appendix A gives illustrative USD conversions to indicate the economic scale of recorded activity, while keeping XCP and BTC as the primary monetary measures.

Appendix A. Currency conversion references and calculations

  • USD equivalents indicate the economic scale of recorded activity; XCP and BTC amounts remain the primary monetary measures.Source: Appendix A · p. 25
  • Counterparty aggregates use a fixed illustrative rate of USD 2.40 per XCP, close to the USD 2.3646 price recorded on 13 July 2014.Table A1 presents the resulting equivalents, rounded as reported in the main text.Source: Appendix A · p. 25, Table A1 · p. 25
  • Orisi’s 0.02037468 BTC in previous outputs converts to USD 12.44, reported as approximately USD 12.The total comprises 0.019 BTC from June–July 2014 and 0.00137468 BTC from October 2014, converted at date-specific illustrative reference prices.Source: Appendix A · p. 26
Loading 2610.11439v1…