pith. sign in

arxiv: 2606.00071 · v1 · pith:EMOCBEIUnew · submitted 2026-05-20 · 💱 q-fin.GN · cs.CE· cs.DC· econ.GN· q-fin.EC

Bitcoin Price Prediction: Peer-Reviewed Evidence and Social Media Discourse

Carlos Baquero This is my paper

Pith reviewed 2026-06-30 17:47 UTC · model grok-4.3

classification 💱 q-fin.GN cs.CEcs.DCecon.GNq-fin.EC
keywords Bitcoin price predictionnaive baselineout-of-sample testingpeer-reviewed studiessocial media discourseevaluation methodologystock-to-flowpower law
0
0 comments X

The pith

No peer-reviewed Bitcoin price prediction model has shown robust superiority over the naive baseline at short-to-medium horizons.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper surveys peer-reviewed studies on Bitcoin price prediction and contrasts them with discussions on social media. It concludes that at horizons of one to six months, none of the models demonstrate consistent outperformance compared to simply assuming the price stays the same as today, across different market conditions. This finding is important because it questions the value of developing increasingly complex prediction techniques if they fail basic out-of-sample tests. The analysis also notes that while daily predictability appears real, it may not hold for hourly or monthly periods and could be eroded by transaction costs. Social media critiques about statistical issues like overfitting are highlighted as valid but not yet incorporated into academic standards.

Core claim

At short-to-medium horizons, no peer-reviewed study has shown robust superiority over the naive baseline across multiple market regimes. Daily predictability is real but does not extend to hourly or monthly horizons, and may not survive transaction costs. The stock-to-flow model has failed formal out-of-sample testing, and Metcalfe's Law valuations have been challenged as spurious. The Bitcoin price power law, while empirically compelling, has not been subjected to formal distributional tests. Social media practitioners raise valid statistical critiques that the academic literature has not formalized.

What carries the argument

The categorization of papers by evaluation methodology, which assesses whether studies performed genuine out-of-sample testing against the naive baseline without post-hoc adjustments.

If this is right

  • Daily predictability exists but does not extend reliably to hourly or monthly horizons.
  • The stock-to-flow model fails formal out-of-sample testing.
  • Metcalfe's Law valuations are challenged as spurious.
  • The Bitcoin price power law lacks formal distributional testing.
  • Social media critiques on OLS violations, backtest overfitting, and spurious regressions remain unformalized in academic work.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Adopting the proposed standards could reduce the number of published models but increase their reliability.
  • Investors may benefit more from focusing on transaction costs and risk management than on new forecasting models.
  • Similar baseline dominance might appear in predictions for other volatile assets if the same evaluation methods are applied.
  • The gap between academic and social media critiques suggests a need for hybrid review processes that incorporate practitioner statistical concerns.

Load-bearing premise

The survey's categorization of papers by evaluation methodology accurately captures whether each study performed genuine out-of-sample testing against the naive baseline without post-hoc adjustments.

What would settle it

Publication of a peer-reviewed study that uses walk-forward evaluation, multi-regime holdout windows, naive baseline comparison, zero in hyperparameter grids, and Diebold-Mariano testing to show statistically significant outperformance across multiple market regimes would falsify the central claim.

read the original abstract

Bitcoin price prediction has attracted hundreds of academic papers and continuous social media debate, yet the field lacks consensus on even basic questions: can any model beat a naive "today's price" baseline at horizons of one to six months? We survey the peer-reviewed landscape, categorize papers by evaluation methodology, and contrast academic findings with informal but substantive discourse on X/Twitter. The picture that emerges is sobering. At short-to-medium horizons, no peer-reviewed study has shown robust superiority over the naive baseline across multiple market regimes. Daily predictability is real but does not extend to hourly or monthly horizons, and may not survive transaction costs. The stock-to-flow model has failed formal out-of-sample testing, and Metcalfe's Law valuations have been challenged as spurious. The Bitcoin price power law, while empirically compelling, has not been subjected to formal distributional tests. Meanwhile, social media practitioners raise valid statistical critiques -- ordinary least squares (OLS) violations, backtest overfitting, spurious regressions -- that the academic literature has not formalized. We identify open research directions and propose concrete methodological standards for future work -- walk-forward evaluation, multi-regime holdout windows, naive baseline comparison, inclusion of zero in hyperparameter grids, and Diebold-Mariano significance testing -- arguing that the field's primary need is not more models but better evaluation.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The manuscript is a literature survey of peer-reviewed Bitcoin price prediction studies. It categorizes papers by evaluation methodology, concludes that no study has demonstrated robust superiority over the naive 'today's price' baseline at 1-6 month horizons across multiple regimes, notes that daily predictability may not survive costs or extend to other horizons, critiques specific models (stock-to-flow, Metcalfe's Law, power law), contrasts findings with X/Twitter discourse on statistical issues, and proposes methodological standards including walk-forward evaluation, multi-regime holdouts, naive baseline comparison, and Diebold-Mariano tests.

Significance. If the survey methodology and categorization prove comprehensive and consistent upon detailed inspection, the negative result would be significant for quantitative finance by documenting a lack of progress against a simple benchmark and by supplying concrete, actionable standards for future work. The explicit contrast with practitioner critiques and the call for falsifiable evaluation practices are strengths that could help redirect the field.

major comments (1)
  1. [Abstract / Survey methodology section] The central claim ('no peer-reviewed study has shown robust superiority over the naive baseline across multiple market regimes') is load-bearing on the survey's paper-selection and categorization process. The abstract and manuscript supply no explicit search string, database list, inclusion/exclusion criteria, or per-paper classification table, so the mapping from raw literature to the negative conclusion cannot be audited or replicated. This directly affects the verifiability of the result (see § on survey methodology and the skeptic concern on categorization).
minor comments (1)
  1. [Abstract] The abstract states the negative result clearly but would benefit from a parenthetical note on the approximate number of papers reviewed and the time window of the literature search.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for highlighting the importance of transparent survey methodology to support the central claim. We address the comment below and will revise accordingly to improve verifiability.

read point-by-point responses

  1. Referee: [Abstract / Survey methodology section] The central claim ('no peer-reviewed study has shown robust superiority over the naive baseline across multiple market regimes') is load-bearing on the survey's paper-selection and categorization process. The abstract and manuscript supply no explicit search string, database list, inclusion/exclusion criteria, or per-paper classification table, so the mapping from raw literature to the negative conclusion cannot be audited or replicated. This directly affects the verifiability of the result (see § on survey methodology and the skeptic concern on categorization).

    Authors: We agree that explicit documentation of the survey process is necessary for the claim to be fully auditable and replicable. The current manuscript describes the categorization approach at a high level but omits the precise search parameters, databases, and a complete classification table. In the revised version we will expand the survey methodology section to specify the search strings (e.g., combinations of 'Bitcoin price prediction', 'cryptocurrency forecasting', and 'out-of-sample evaluation'), the databases queried (Google Scholar, Web of Science, SSRN, arXiv), the time window, inclusion criteria (peer-reviewed empirical studies with quantitative out-of-sample tests at 1-6 month horizons), exclusion criteria (non-peer-reviewed preprints, purely theoretical papers, studies without baseline comparisons), and an appendix table that lists each surveyed paper together with its evaluation method, horizon, regime coverage, and reason for classification relative to the naive baseline. This addition will directly address replicability concerns and allow readers to inspect the mapping from literature to conclusion. revision: yes

Circularity Check

0 steps flagged

Literature survey with no derivations or self-referential reductions

full rationale

This is a literature survey paper whose central claim rests on categorization and summary of external peer-reviewed studies. No equations, fitted parameters, predictions, or derivations are present that could reduce to the paper's own inputs by construction. The taxonomy of evaluation methodologies is an empirical classification of outside work, not a self-definitional or fitted-input step. No self-citations are load-bearing for any uniqueness theorem or ansatz. The paper is self-contained against external benchmarks in the sense that its conclusions are falsifiable by re-examination of the cited literature.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim depends on the assumption that the reviewed papers constitute a representative sample and that their evaluation methodologies were correctly classified as lacking robust out-of-sample testing.

axioms (2)
  • domain assumption The naive 'today's price' baseline is the appropriate comparator for assessing predictive skill at 1-6 month horizons.
    Invoked when stating that no study shows superiority over this baseline.
  • domain assumption Peer-reviewed studies can be exhaustively categorized by evaluation methodology from their published descriptions.
    Required for the claim that none have demonstrated robust superiority.

pith-pipeline@v0.9.1-grok · 5771 in / 1197 out tokens · 22587 ms · 2026-06-30T17:47:40.980175+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

36 extracted references · 31 canonical work pages

  1. [1]

    Regime switching forecasting for cryptocurrencies

    Ilgar Agakishiev, Wolfgang Karl H\" a rdle, and Delia Becker. Regime switching forecasting for cryptocurrencies. Digital Finance, 7: 0 107--131, 2025. doi:10.1007/s42521-024-00123-2

    work page doi:10.1007/s42521-024-00123-2 2025

  2. [2]

    When are statistical forecast gains economically relevant? Evidence from Bitcoin returns

    Rehan Arain and Stephen Snudden. When are statistical forecast gains economically relevant? Evidence from Bitcoin returns. Journal of Forecasting, 2025. doi:10.1002/for.70077

    work page doi:10.1002/for.70077 2025

  3. [3]

    Bailey, Jonathan M

    David H. Bailey, Jonathan M. Borwein, Marcos L\' o pez de Prado, and Qiji Jim Zhu. Pseudo-mathematics and financial charlatanism: The effects of backtest overfitting on out-of-sample performance. Notices of the American Mathematical Society, 61 0 (5): 0 458--471, 2014. https://www.ams.org/notices/201405/rnoti-p458.pdf

    2014

  4. [4]

    The naive--power law blend as a robust baseline for Bitcoin price forecasting

    Carlos Baquero and Daniel Tinoco. The naive--power law blend as a robust baseline for Bitcoin price forecasting. Zenodo preprint (submitted to Digital Finance), 2026. doi:10.5281/zenodo.19558174

    work page doi:10.5281/zenodo.19558174 2026

  5. [5]

    Baur and Thomas Dimpfl

    Dirk G. Baur and Thomas Dimpfl. The volatility of Bitcoin and its role as a medium of exchange and a store of value. Empirical Economics, 61 0 (5): 0 2663--2683, 2021. doi:10.1007/s00181-020-01990-5

    work page doi:10.1007/s00181-020-01990-5 2021

  6. [6]

    Forecasting Bitcoin returns: Econometric time series analysis vs.\ machine learning

    Timo Berger and Jana Koubov\' a . Forecasting Bitcoin returns: Econometric time series analysis vs.\ machine learning. Journal of Forecasting, 43 0 (7): 0 2904--2916, 2024. doi:10.1002/for.3165

    work page doi:10.1002/for.3165 2024

  7. [7]

    Broido and Aaron Clauset

    Anna D. Broido and Aaron Clauset. Scale-free networks are rare. Nature Communications, 10: 0 1017, 2019. doi:10.1038/s41467-019-08746-5

    work page doi:10.1038/s41467-019-08746-5 2019

  8. [8]

    Bitcoin 's natural long-term power-law corridor of growth

    Harold Christopher Burger. Bitcoin 's natural long-term power-law corridor of growth. https://hcburger.com/blog/powerlaw/, 2019. Accessed: 2026-04-11

    2019

  9. [9]

    Machine learning and the cross-section of cryptocurrency returns

    Nusret Cakici, Syed Jawad Hussain Shahzad, Barbara B e dowska-S\' o jka, and Adam Zaremba. Machine learning and the cross-section of cryptocurrency returns. International Review of Financial Analysis, 94: 0 103244, 2024. doi:10.1016/j.irfa.2024.103244

    work page doi:10.1016/j.irfa.2024.103244 2024

  10. [10]

    Campbell, Andrew W

    John Y. Campbell, Andrew W. Lo, and A. Craig MacKinlay. The Econometrics of Financial Markets. Princeton University Press, 1997

    1997

  11. [11]

    Aaron Clauset, Cosma Rohilla Shalizi, and Mark E. J. Newman. Power-law distributions in empirical data. SIAM Review, 51 0 (4): 0 661--703, 2009. doi:10.1137/070710111

    work page doi:10.1137/070710111 2009

  12. [12]

    On the intraday behavior of Bitcoin

    Giacomo De Nicola. On the intraday behavior of Bitcoin . Ledger, 6, 2021. doi:10.5195/ledger.2021.213

    work page doi:10.5195/ledger.2021.213 2021

  13. [13]

    Diebold and Roberto S

    Francis X. Diebold and Roberto S. Mariano. Comparing predictive accuracy. Journal of Business & Economic Statistics, 13 0 (3): 0 253--263, 1995. doi:10.1080/07350015.1995.10524599

    work page doi:10.1080/07350015.1995.10524599 1995

  14. [14]

    Eugene F. Fama. Efficient capital markets: A review of theory and empirical work. The Journal of Finance, 25 0 (2): 0 383--417, 1970. doi:10.2307/2325486

    work page doi:10.2307/2325486 1970

  15. [15]

    Forecasting Bitcoin with technical analysis: A not-so-random forest? International Journal of Forecasting, 39 0 (1): 0 1--17, 2023

    Nikola Gradojevic, Dragan Kukolj, Robert Adcock, and Vladimir Djakovic. Forecasting Bitcoin with technical analysis: A not-so-random forest? International Journal of Forecasting, 39 0 (1): 0 1--17, 2023. doi:10.1016/j.ijforecast.2021.08.001

    work page doi:10.1016/j.ijforecast.2021.08.001 2023

  16. [16]

    Cross-cryptocurrency return predictability

    Li Guo, Bo Sang, Jun Tu, and Yu Wang. Cross-cryptocurrency return predictability. Journal of Economic Dynamics and Control, 163: 0 104863, 2024. doi:10.1016/j.jedc.2024.104863

    work page doi:10.1016/j.jedc.2024.104863 2024

  17. [17]

    Deep learning and NLP in cryptocurrency forecasting: Integrating financial, blockchain, and social media data

    Vincent Gurgul, Stefan Lessmann, and Wolfgang Karl H\" a rdle. Deep learning and NLP in cryptocurrency forecasting: Integrating financial, blockchain, and social media data. International Journal of Forecasting, 41: 0 1666--1695, 2025. doi:10.1016/j.ijforecast.2025.02.007

    work page doi:10.1016/j.ijforecast.2025.02.007 2025

  18. [18]

    Hansen, Asger Lunde, and James M

    Peter R. Hansen, Asger Lunde, and James M. Nason. The model confidence set. Econometrica, 79 0 (2): 0 453--497, 2011. doi:10.3982/ECTA5771

    work page doi:10.3982/ecta5771 2011

  19. [19]

    Volatility estimation for Bitcoin : A comparison of GARCH models

    Paraskevi Katsiampa. Volatility estimation for Bitcoin : A comparison of GARCH models. Economics Letters, 158: 0 3--6, 2017. doi:10.1016/j.econlet.2017.06.023

    work page doi:10.1016/j.econlet.2017.06.023 2017

  20. [20]

    Bitcoin price direction forecasting and market variables

    Taegyum Kim, Hyeontae Jo, Woohyuk Choi, and Bong-Gyu Jang. Bitcoin price direction forecasting and market variables. Journal of Futures Markets, 45 0 (10): 0 1579--1600, 2025. doi:10.1002/fut.70010

    work page doi:10.1002/fut.70010 2025

  21. [21]

    o se, Yunus Emre G\

    Nezir K\" o se, Yunus Emre G\" u r, and Emre \" U nal. Deep learning and machine learning insights into the global economic drivers of the Bitcoin price. Journal of Forecasting, 44 0 (5): 0 1666--1698, 2025. doi:10.1002/for.3258

    work page doi:10.1002/for.3258 2025

  22. [22]

    Andrew W. Lo. The adaptive markets hypothesis. The Journal of Portfolio Management, 30 0 (5): 0 15--29, 2004. doi:10.3905/jpm.2004.442611

    work page doi:10.3905/jpm.2004.442611 2004

  23. [23]

    Cryptocurrency forecasting: More evidence of the Meese-Rogoff puzzle

    Nicol\' a s Magner and Nicol\' a s Hardy. Cryptocurrency forecasting: More evidence of the Meese-Rogoff puzzle. Mathematics, 10 0 (13): 0 2338, 2022. doi:10.3390/math10132338

    work page doi:10.3390/math10132338 2022

  24. [24]

    Predicting the price of Bitcoin using machine learning

    Sean McNally, Jason Roche, and Simon Caton. Predicting the price of Bitcoin using machine learning. In 26th Euromicro International Conference on Parallel, Distributed and Network-based Processing (PDP), pages 339--343, 2018. doi:10.1109/PDP2018.2018.00060

    work page doi:10.1109/pdp2018.2018.00060 2018

  25. [25]

    Meese and Kenneth Rogoff

    Richard A. Meese and Kenneth Rogoff. Empirical exchange rate models of the seventies: Do they fit out of sample? Journal of International Economics, 14 0 (1--2): 0 3--24, 1983. doi:10.1016/0022-1996(83)90017-X

    work page doi:10.1016/0022-1996(83)90017-x 1983

  26. [26]

    Metcalfe 's law after 40 years of Ethernet

    Bob Metcalfe. Metcalfe 's law after 40 years of Ethernet . Computer, 46 0 (12): 0 26--31, 2013. doi:10.1109/MC.2013.374

    work page doi:10.1109/mc.2013.374 2013

  27. [27]

    Exploring the dynamics of Bitcoin 's price: a Bayesian structural time series approach

    Obryan Poyser. Exploring the dynamics of Bitcoin 's price: a Bayesian structural time series approach. Eurasian Economic Review, 9: 0 29--60, 2019. doi:10.1007/s40822-018-0108-2

    work page doi:10.1007/s40822-018-0108-2 2019

  28. [28]

    Quantifying cryptocurrency unpredictability: A comprehensive study of complexity and forecasting

    Francesco Puoti, Fabrizio Pittorino, and Manuel Roveri. Quantifying cryptocurrency unpredictability: A comprehensive study of complexity and forecasting. In Proceedings of the 4th International Conference on AI-ML Systems (AIMLSystems 2024), 2024. doi:10.1145/3703412.3703420

    work page doi:10.1145/3703412.3703420 2024

  29. [29]

    A mechanistic derivation of the Bitcoin price power law: Network adoption dynamics and generalised Metcalfe scaling

    Giovanni Santostasi and Stephen Perrenod. A mechanistic derivation of the Bitcoin price power law: Network adoption dynamics and generalised Metcalfe scaling. Zenodo preprint, 2026

    2026

  30. [30]

    Scott and Hal R

    Steven L. Scott and Hal R. Varian. Predicting the present with Bayesian structural time series. International Journal of Mathematical Modelling and Numerical Optimisation, 5 0 (1--2): 0 4--23, 2014. doi:10.1504/IJMMNO.2014.059942

    work page doi:10.1504/ijmmno.2014.059942 2014

  31. [31]

    The marginal cost of mining, Metcalfe 's law and cryptocurrency value formation: Causal inferences from the instrumental variable approach

    Savva Shanaev, Satish Sharma, Arina Shuraeva, and Binam Ghimire. The marginal cost of mining, Metcalfe 's law and cryptocurrency value formation: Causal inferences from the instrumental variable approach. SSRN working paper 3432431, 2019

    2019

  32. [32]

    Austin Shelton. Bitcoin return prediction: Is it possible via stock-to-flow, Metcalfe 's law, technical analysis, or market sentiment? Journal of Risk and Financial Management, 17 0 (10): 0 443, 2024. doi:10.3390/jrfm17100443

    work page doi:10.3390/jrfm17100443 2024

  33. [33]

    Cryptocurrency competition and market concentration in the presence of network effects

    Konstantinos Stylianou, Leonhard Spiegelberg, Maurice Herlihy, and Nic Carter. Cryptocurrency competition and market concentration in the presence of network effects. Ledger, 6, 2021. doi:10.5195/ledger.2021.226

    work page doi:10.5195/ledger.2021.226 2021

  34. [34]

    Forecasting and trading Bitcoin with machine learning techniques and a hybrid volatility/sentiment leverage

    Mingzhe Wei, Georgios Sermpinis, and Charalampos Stasinakis. Forecasting and trading Bitcoin with machine learning techniques and a hybrid volatility/sentiment leverage. Journal of Forecasting, 42 0 (4): 0 852--871, 2023. doi:10.1002/for.2922

    work page doi:10.1002/for.2922 2023

  35. [35]

    Spencer Wheatley, Didier Sornette, Tobias Huber, Max Reppen, and Robert N. Gantner. Are Bitcoin bubbles predictable? Combining a generalized Metcalfe 's law and the log-periodic power law singularity model. Royal Society Open Science, 6 0 (6): 0 180538, 2019. doi:10.1098/rsos.180538

    work page doi:10.1098/rsos.180538 2019

  36. [36]

    Out-of-sample forecasting of cryptocurrency returns: A comprehensive comparison of predictors and algorithms

    James Yae and George Zhe Tian. Out-of-sample forecasting of cryptocurrency returns: A comprehensive comparison of predictors and algorithms. Physica A: Statistical Mechanics and its Applications, 598: 0 127379, 2022. doi:10.1016/j.physa.2022.127379

    work page doi:10.1016/j.physa.2022.127379 2022