pith. sign in

arxiv: 2606.31475 · v1 · pith:TWGL4CFHnew · submitted 2026-06-30 · 💱 q-fin.ST · nlin.PS

Real-time identification of the onset of financial rogue waves

Rosie Hayward , Orla Lennon , Fabio Biancalana This is my paper

Pith reviewed 2026-07-01 02:32 UTC · model grok-4.3

classification 💱 q-fin.ST nlin.PS
keywords financial volatilityrogue wavesSchrödinger equationKerr nonlinearityeigenvalue gradientVIX indexextreme eventsAnderson localization
0
0 comments X

The pith

A Kerr-nonlinear Schrödinger model applied to volatility data detects the onset of most financial rogue-wave peaks via the gradient of its minimum eigenvalue.

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

The paper tries to establish that financial volatility indices exhibit rogue-wave-like peaks with similar statistics to optical and hydrodynamical systems. By solving a Schrödinger equation whose potential follows a Kerr nonlinearity shape over moving time windows on indices such as the VIX, the approach shows Anderson localisation before peaks and a reliable spike in the numerical gradient of the minimum eigenvalue at extreme-event onset. If correct, this yields real-time detection of seven out of eight major VIX peaks and replicates at the same rate on out-of-sample VXO and VSTOXX tests. A sympathetic reader would care because volatility extremes remain hard to flag near their start, and the method supplies an early-warning signal drawn from a physics analogy.

Core claim

By treating financial volatility within a moving time window as solutions to a Schrödinger equation whose potential follows a Kerr nonlinearity, the system exhibits Anderson localisation approaching rogue peaks in the VIX, and the numerical gradient of the minimum eigenvalue spikes reliably at the onset of an extreme event. This approach detects all but one of the VIX's major peaks in real-time simulation, and replicates the result on out-of-sample tests for the VXO and VSTOXX indices at 87.5 percent success rate.

What carries the argument

The numerical gradient of the minimum eigenvalue of the Kerr-nonlinear Schrödinger system, which spikes at the onset of extreme volatility events.

If this is right

  • The method identifies seven out of eight major VIX peaks given a reasonable amount of history.
  • Out-of-sample tests on the VXO and VSTOXX indices each detect all but one major peak.
  • The approach adapts to simulate real-time data arrival for practical monitoring.
  • It supplies a candidate tool for identifying potential financial crises before full development.

Where Pith is reading between the lines

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

  • The same eigenvalue-gradient indicator could be tested on non-financial time series that display extreme events, such as climate or network-load records.
  • If the detection rate holds, the window length or exact shape of the Kerr potential could be varied to check whether false-positive rates change.
  • The underlying analogy might link financial volatility to other systems already described by nonlinear Schrödinger equations, opening cross-domain comparisons of localisation before extremes.

Load-bearing premise

The statistical properties of financial volatility indices are captured by a Schrödinger equation with a Kerr nonlinearity potential so that the minimum-eigenvalue gradient indicates onset without needing post-hoc peak labels.

What would settle it

A concrete falsifier would be if, on longer histories or additional volatility indices, the gradient fails to spike before more than one in eight major peaks or generates frequent false positives unrelated to known extremes.

Figures

Figures reproduced from arXiv: 2606.31475 by Fabio Biancalana, Orla Lennon, Rosie Hayward.

Figure 1
Figure 1. Figure 1: a) VIX highs (light grey dashed), envelope wave (black), with extreme event peaks marked by red dots (> 2.5 SWH) and red crosses (> 2 SWH). A horizontal line marks where the VIX exceeds 35. b) Frequency distribution of raw heights (values at the peak locations (> 2.5 SWH) ) of the VIX highs, with a vertical orange line marking a height of 35. c) Distribution of the prominences of the peaks (> 2.5 SWH) foun… view at source ↗
Figure 2
Figure 2. Figure 2: a) Frequency density plot of minimum eigenvector widths at differing distances from the next extreme peak. In all panels, blue shows the distribution of values found between 0 and 5 days from a peak, orange between 5 and 30 days, and green for more than 30 days. b) Density plot of the skewness of the potential, c) density plot of the minimum eigenvector distance, and d) density plot of the kurtosis of the … view at source ↗
Figure 3
Figure 3. Figure 3: a) Minimum eigenvalue gradient (black dots) for increasing number of days before the next extreme peak. The red dashed line marks the cut-off of 74, beyond which spikes are considered a reliable indicator of an oncoming peak. b) The gradient of the squared envelope wave (black dots) for an increasing number of days before the next peak. The same red line with the cut-off is shown. c) The envelope wave of t… view at source ↗
Figure 4
Figure 4. Figure 4: a) The VIX (grey) and the maximum eigenvalue gradient found in the last ten days (blue). b) The envelope wave (grey), rogue wave peaks (red dots (2.5×SWH) or crosses (2×SWH)), and the rogue wave warning indicator (red vertical lines, except in the ten days following a peak where they are blue, dashed). 30 20 10 0 10 20 30 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 Pearson cross correlation VXO VSTOXX VIX [PI… view at source ↗
Figure 5
Figure 5. Figure 5: Pearson cross correlation between the maximum eigenvalue gradient for each volatility index and the height of the volatility index. VSTOXX The out-of-sample test results for the VSTOXX index are displayed in figure 7. Here, the threshold chosen based on the results of repeated trials within the shaded in-sample period is equal to 300. The results for this index are very precise (>70%), with the 2016 peak m… view at source ↗
Figure 6
Figure 6. Figure 6: The envelope wave of the VXO index (grey line), the in-sample area (grey shading), rogue wave peaks (red dots (2.5×SWH) or crosses (>2×SWH), the rogue wave warning indicator (red vertical lines, except for the ten days following an extreme peak where it is displayed using blue dashed lines), and the out of sample test threshold (grey dashed). 2000 2004 2008 2012 2016 2020 2024 20 0 20 40 60 vstoxx envelope… view at source ↗
Figure 7
Figure 7. Figure 7: The envelope wave of the VSTOXX index (grey line), the in-sample area (grey shading), rogue wave peaks (red dots (2.5×SWH) or crosses (>2×SWH), the rogue wave warning indicator (red vertical lines, except for the ten days following an extreme peak where it is displayed using blue dashed lines), and the out of sample test threshold (grey dashed). Condition % (> 2.5 SWH) % (> 2 SWH) Peaks detected 87.5 88.9 … view at source ↗
read the original abstract

Extreme events in financial systems, often captured by indicators such as volatility, remain difficult to identify close to their onset. Volatility shares many statistical properties with other natural, complex systems which experience extreme events, which we explore in this manuscript. We extend the analogy between rogue waves in optical and hydrodynamical systems to financial volatility by identifying rogue-wave-like peaks with similar statistical properties. We use a Schr\"odinger equation where the potential follows the shape of a Kerr nonlinearity to examine the properties of financial volatility indices within a moving time window. We see evidence of Anderson localisation as a rogue peak approaches in the VIX, and show that the numerical gradient of the system's minimum eigenvalue reliably spikes at the onset of an extreme event. We adapt our methodology to simulate the real-time arrival of data, and show that all but one of the VIX's major peaks can be detected given a reasonable amount of history. We then perform two out-of-sample tests, one for the VXO index, and one for the VSTOXX index, and successfully replicate our initial results, identifying all but one major peak (87.5% or 7/8) in both cases. This method of analysis shows considerable promise as a tool for identifying potential financial crises, aiding in their mitigation.

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

3 major / 2 minor

Summary. The manuscript extends the rogue-wave analogy from optics to financial volatility by modeling VIX, VXO and VSTOXX time series inside sliding windows with a Schrödinger equation whose potential follows a Kerr nonlinearity. It reports that the numerical gradient of the minimum eigenvalue spikes at the onset of extreme events, detects 7/8 major VIX peaks in a real-time simulation, and replicates the result out-of-sample on VXO and VSTOXX at the same 87.5 % rate. Evidence of Anderson localisation is also claimed as peaks approach.

Significance. If the eigenvalue-gradient indicator proves robust and non-circular, the work would supply a concrete, physics-motivated early-warning statistic for volatility spikes that could be implemented with modest computational cost. The out-of-sample replication on two additional indices is a positive feature. However, the absence of any derivation linking the Kerr term to a stochastic-volatility SDE or market-microstructure model leaves open the possibility that the reported spikes are an artifact of the chosen nonlinearity rather than a reflection of underlying financial dynamics.

major comments (3)
  1. [Introduction / Methods] The manuscript motivates the Kerr-nonlinear Schrödinger equation solely by statistical resemblance to optical rogue waves (Introduction). No mapping is supplied from any SDE for stochastic volatility or from any market-microstructure model to the nonlinear term; consequently the claim that the minimum-eigenvalue gradient constitutes a genuine leading indicator (rather than a numerical consequence of the chosen potential inside a sliding window) rests on an untested analogy.
  2. [Results] The definition of 'major peaks', the numerical threshold applied to the eigenvalue gradient, and the precise rule for declaring a detection in the real-time simulation are not stated. Without these operational details it is impossible to assess whether the reported 87.5 % (7/8) detection rate on VIX and the identical rate on the two out-of-sample indices reflect genuine predictive power or post-hoc tuning.
  3. [Methods] No sensitivity analysis is reported for window length, discretization of the Schrödinger operator, or the functional form used to shape the Kerr potential. These choices directly affect the eigenvalue gradient; their omission leaves the central empirical claim vulnerable to the criticism that any sufficiently nonlinear potential would produce comparable spikes.
minor comments (2)
  1. [Abstract] The abstract states that 'all but one of the VIX's major peaks can be detected given a reasonable amount of history' yet does not quantify what constitutes a 'reasonable amount of history' or report the corresponding lag distribution.
  2. [Figures] Figure captions and axis labels should explicitly state the units of the eigenvalue gradient and the time scale of the moving window so that readers can reproduce the spike detection without additional assumptions.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the thoughtful comments and the recognition of the out-of-sample tests. We provide point-by-point responses below. Where clarifications are needed, we will revise the manuscript accordingly. We maintain that the nonlinear Schrödinger framework offers a novel perspective on volatility dynamics based on established analogies in complex systems.

read point-by-point responses

  1. Referee: [Introduction / Methods] The manuscript motivates the Kerr-nonlinear Schrödinger equation solely by statistical resemblance to optical rogue waves (Introduction). No mapping is supplied from any SDE for stochastic volatility or from any market-microstructure model to the nonlinear term; consequently the claim that the minimum-eigenvalue gradient constitutes a genuine leading indicator (rather than a numerical consequence of the chosen potential inside a sliding window) rests on an untested analogy.

    Authors: The motivation indeed stems from the observed statistical similarities between financial volatility and rogue waves in optics, as detailed in the Introduction. We do not claim a direct derivation from an SDE; rather, the model serves as an exploratory tool to detect patterns in the data. The consistent performance across VIX, VXO, and VSTOXX provides empirical support. To address the concern, we will expand the discussion to explicitly note the analogy-based approach and suggest avenues for linking to stochastic volatility models in future work. revision: partial

  2. Referee: [Results] The definition of 'major peaks', the numerical threshold applied to the eigenvalue gradient, and the precise rule for declaring a detection in the real-time simulation are not stated. Without these operational details it is impossible to assess whether the reported 87.5 % (7/8) detection rate on VIX and the identical rate on the two out-of-sample indices reflect genuine predictive power or post-hoc tuning.

    Authors: We agree that these details are crucial for evaluating the results. In the revised version, we will clearly define major peaks (e.g., as local maxima exceeding 2 standard deviations above the mean in the volatility series), specify the gradient threshold used for detection (based on a percentile of historical values), and outline the exact real-time simulation protocol, including the amount of history required. revision: yes

  3. Referee: [Methods] No sensitivity analysis is reported for window length, discretization of the Schrödinger operator, or the functional form used to shape the Kerr potential. These choices directly affect the eigenvalue gradient; their omission leaves the central empirical claim vulnerable to the criticism that any sufficiently nonlinear potential would produce comparable spikes.

    Authors: We recognize the importance of robustness checks. While the Kerr form is chosen to parallel the optical case, we will include a sensitivity analysis in an appendix demonstrating that the detection performance is stable across a range of window lengths (e.g., 50-200 days) and discretization steps. We will also discuss why alternative potential forms were not pursued, to maintain consistency with the physical analogy. revision: partial

Circularity Check

0 steps flagged

No significant circularity; derivation is self-contained empirical application

full rationale

The paper motivates the Schrödinger equation with Kerr-shaped potential by statistical analogy to optical rogue waves and applies it to volatility time series in sliding windows to compute the gradient of the minimum eigenvalue as an indicator. This indicator is then tested for spikes preceding identified peaks in VIX (in-sample) and replicated on VXO/VSTOXX (out-of-sample), yielding an 87.5% detection rate. No quoted step shows a fitted parameter or threshold being renamed as a prediction, no self-citation chain justifies a uniqueness claim, and the potential form is an explicit modeling choice rather than a self-definitional reduction. The detection performance is an empirical outcome on external data, not forced by construction from the inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the modeling assumption that financial volatility obeys the same nonlinear wave equation used for optical rogue waves; no free parameters or invented entities are explicitly named in the abstract.

axioms (1)
  • domain assumption Financial volatility indices share statistical properties with rogue waves in optical and hydrodynamical systems and can therefore be modeled by a Schrödinger equation with Kerr nonlinearity potential.
    This is the explicit analogy stated in the abstract that justifies applying the physics model.

pith-pipeline@v0.9.1-grok · 5757 in / 1270 out tokens · 56830 ms · 2026-07-01T02:32:00.301230+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

29 extracted references · 20 canonical work pages · 1 internal anchor

  1. [1]

    Guerrero, F. G. & Garcia-Baños, A. Multiplicative processes as a source of fat-tail distributions.Heliyon6, e04266, DOI: 10.1016/j.heliyon.2020.e04266 (2020)

    work page doi:10.1016/j.heliyon.2020.e04266 2020

  2. [2]

    & Bouchaud, J.-P

    Sornette, D., Johansen, A. & Bouchaud, J.-P. Stock Market Crashes, Precursors and Replicas.J. Phys. I France6, 167–175, DOI: 10.1051/jp1:1996135 (1996)

    work page doi:10.1051/jp1:1996135 1996

  3. [3]

    & Ahn, K

    Lee, G., Jeong, M., Park, T. & Ahn, K. More than ex- post fitting: log-periodic power law and its AI-based classification.Humanit. Soc Sci Commun12, 1664, DOI: 10.1057/s41599-025-05920-7 (2025)

    work page doi:10.1057/s41599-025-05920-7 2025

  4. [4]

    M., Genty, G., Mussot, A., Chabchoub, A

    Dudley, J. M., Genty, G., Mussot, A., Chabchoub, A. & Dias, F. Rogue waves and analogies in optics and oceanography.Nat Rev Phys1, 675–689, DOI: 10. 1038/s42254-019-0100-0 (2019)

    2019

  5. [5]

    A Possible Freak Wave Event Measured at the Draupner Jacket January 1 1995.Actes de colloques-IFREMER39(2004)

    Haver, S. A Possible Freak Wave Event Measured at the Draupner Jacket January 1 1995.Actes de colloques-IFREMER39(2004)

    1995

  6. [6]

    & Jalali, B

    Solli, D., Ropers, C., Koonath, P. & Jalali, B. Optical rogue waves.Nature450, 1054–7, DOI: 10.1038/ nature06402 (2008)

    2008

  7. [7]

    & Richter, C

    Gutenberg, B. & Richter, C. F. Frequency of earth- quakes in California.Bull. Seismol. Soc. Am.34, 185– 88, DOI: 10.1785/BSSA0340040185 (1944)

    work page doi:10.1785/bssa0340040185 1944

  8. [8]

    & Haken, H

    Hutt, A. & Haken, H. (eds.)Synergetics(Springer US, New York, NY , 2020)

    2020

  9. [9]

    Liu, Y .et al.The statistical properties of the volatility of price fluctuations.Phys. Rev. E60, 1390–1400, DOI: 10.1103/PhysRevE.60.1390 (1999). ArXiv:cond- mat/9903369

    work page doi:10.1103/physreve.60.1390 1999

  10. [10]

    Global Omori law decay of triggered earthquakes: Large aftershocks outside the classi- cal aftershock zone.J

    Parsons, T. Global Omori law decay of triggered earthquakes: Large aftershocks outside the classi- cal aftershock zone.J. Geophys. Res.107, DOI: 10.1029/2001JB000646 (2002)

    work page doi:10.1029/2001jb000646 2002

  11. [11]

    Studies of the limit order book around large price changes

    Toth, B., Kertesz, J. & Farmer, J. D. Studies of the limit order book around large price changes.Eur . Phys. J. B71, 499–510, DOI: 10.1140/epjb/e2009-00297-9 (2009). ArXiv:0901.0495 [q-fin]

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1140/epjb/e2009-00297-9 2009

  12. [12]

    & Mantegna, R

    Lillo, F. & Mantegna, R. N. Power-law relaxation in a complex system: Omori law after a financial mar- ket crash.Phys. Rev. E68, 016119, DOI: 10.1103/ PhysRevE.68.016119 (2003)

    2003

  13. [13]

    & Serva, M

    Pasquini, M. & Serva, M. Multiscale behaviour of volatility autocorrelations in a financial market.Econ. Lett.65, 275–279, DOI: 10.1016/S0165-1765(99) 00159-7 (1999)

    work page doi:10.1016/s0165-1765(99 1999

  14. [14]

    & Dodge, Y

    Ghashghaie, S., Breymann, W., Peinke, J., Talkner, P. & Dodge, Y . Turbulent cascades in foreign exchange mar- kets.Nature381, 767–770, DOI: 10.1038/381767a0 (1996)

    work page doi:10.1038/381767a0 1996

  15. [15]

    Ivancevic, V . G. Adaptive-Wave Alternative for the Black-Scholes Option Pricing Model.Cogn. Comput. 2, 17–30, DOI: 10.1007/s12559-009-9031-x (2010)

    work page doi:10.1007/s12559-009-9031-x 2010

  16. [16]

    Financial Rogue Waves.Commun

    Yan, Z. Financial Rogue Waves.Commun. Theor . Phys. 54, 947, DOI: 10.1088/0253-6102/54/5/31 (2010)

    work page doi:10.1088/0253-6102/54/5/31 2010

  17. [17]

    & Bouchaud, J.- P

    Aubrun, C., Morel, R., Benzaquen, M. & Bouchaud, J.- P. Identifying new classes of financial price jumps with wavelets.Proc. Natl. Acad. Sci.122, e2409156121, DOI: 10.1073/pnas.2409156121 (2025)

    work page doi:10.1073/pnas.2409156121 2025

  18. [18]

    Zakharov, V . E. Stability of periodic waves of finite am- plitude on the surface of a deep fluid.J Appl Mech Tech Phys9, 190–194, DOI: 10.1007/BF00913182 (1972)

    work page doi:10.1007/bf00913182 1972

  19. [19]

    Y ., Garmire, E

    Chiao, R. Y ., Garmire, E. & Townes, C. H. Self- Trapping of Optical Beams.Phys. Rev. Lett.13, 479– 482, DOI: 10.1103/PhysRevLett.13.479 (1964)

    work page doi:10.1103/physrevlett.13.479 1964

  20. [20]

    Nonlinear Schrödinger equations

    Kato, T. Nonlinear Schrödinger equations. In Holden, H. & Jensen, A. (eds.)Schrödinger Operators, 218– 263 (Springer Berlin Heidelberg, Berlin, Heidelberg, 2005)

    2005

  21. [21]

    Optics Express19(10), 9315–9329 (2011).https://doi.org/10.1364/OE

    Saleh, M. F., Conti, C. & Biancalana, F. Anderson localisation and optical-event horizons in rogue-soliton generation.Opt. Express25, 5457, DOI: 10.1364/OE. 25.005457 (2017)

    work page doi:10.1364/oe 2017

  22. [22]

    & Falcon, E

    Ricard, G., Novkoski, F. & Falcon, E. Effects of nonlinearity on Anderson localization of surface grav- ity waves.Nat Commun15, 5726, DOI: 10.1038/ s41467-024-49575-5 (2024)

    2024

  23. [23]

    V olatility Index® Methodology: Cboe V olatility In- dex® (2024)

    2024

  24. [24]

    & Shuttlewood, T

    Seegopaul, H. & Shuttlewood, T. VSTOXX 101: Un- derstanding Europe’s volatility benchmark (2024)

    2024

  25. [25]

    Discussion & Debate: Rogue Waves – Towards a Unifying Concept?

    Akhmediev, N. & Pelinovsky, E. Editorial – Introduc- tory remarks on “Discussion & Debate: Rogue Waves – Towards a Unifying Concept?”.Eur . Phys. J. Spec. Top. 185, 1–4, DOI: 10.1140/epjst/e2010-01233-0 (2010)

    work page doi:10.1140/epjst/e2010-01233-0 2010

  26. [26]

    Dysthe, K., Krogstad, H. E. & Müller, P. Oceanic Rogue Waves.Annu. Rev. Fluid Mech.40, 287–310, DOI: 10.1146/annurev.fluid.40.111406.102203 (2008). 10/19

    work page doi:10.1146/annurev.fluid.40.111406.102203 2008

  27. [27]

    D., Srokosz, M., Moat, B

    Cattrell, A. D., Srokosz, M., Moat, B. I. & Marsh, R. Can Rogue Waves Be Predicted Using Characteristic Wave Parameters?J. Geophys. Res. Ocean.123, 5624– 5636, DOI: 10.1029/2018JC013958 (2018). _eprint: https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1029/2018JC013958

    work page doi:10.1029/2018jc013958 2018

  28. [28]

    Transverse Anderson localization of light: a tutorial.Adv

    Mafi, A. Transverse Anderson localization of light: a tutorial.Adv. Opt. Photon., AOP7, 459–515, DOI: 10.1364/AOP.7.000459 (2015)

    work page doi:10.1364/aop.7.000459 2015

  29. [29]

    lose weight

    Degiannakis, S., Filis, G. & Hassani, H. Forecasting global stock market implied volatility indices.J. Empir. Finance46, 111–129, DOI: 10.1016/j.jempfin.2017.12. 008 (2018). 11/19 Supplementary results 12/19 0 100 200 300 Days since peak (13/10/2008) 0 2 4 6 8 10 12N(t) a) R^2=0.978 8.46 8.47 8.48 8.49 8.50 8.51 8.52 ln(date index) 3.2 3.4 3.6 3.8 4.0 4.2...

    work page doi:10.1016/j.jempfin.2017.12 2017