Pith. sign in

REVIEW 3 major objections 5 minor 30 references

Option Pricing Using Ensemble Learning

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Ensemble learning—specifically LGBM, NGBoost, and XGBoost—consistently outperforms other tested models in pricing CSI 300 index options.

desk verdict The paper's central model comparison is invalid as reported because the training set contains future data relative to the test window, so the ensemble advantage may be leakage rather than genuine predictive skill. read the letter →

arxiv 2506.05799 v1 pith:HAYUOGGO submitted 2025-06-06 cs.LG

Zeyuan Li , Qingdao Huang This is my paper
classification cs.LG MSC 91G2068T05
keywords OptionpricingEnsemblelearningGradientboostingLGBMNGBoostXGBoostBlack-ScholesmodelNoiserobustness
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tries to establish that ensemble learning models—gradient-boosted tree methods LGBM, NGBoost, and XGBoost in particular—price options more accurately than classical machine learning models and the Black–Scholes baseline. Using CSI 300 index option data, the authors run four experiments covering input configurations, moneyness (the spot-to-strike price ratio), sliding-window local features, and noise robustness. In the two main accuracy experiments, the boosting methods consistently post the lowest root-mean-squared error, outperforming MLP, LSTM, genetic algorithms, random forest, DeepForest, CatBoost, Black–Scholes, and Black–Scholes–Merton. The paper also introduces a hybrid tuning strategy that transfers hyperparameters across experiments and a scoring mechanism that uses the Black–Scholes model as a theoretical anchor, arguing this makes the evaluation more financially grounded. If correct, the results suggest gradient boosting is a practical substitute for neural-network approaches in nonparametric option pricing.

What carries the argument

The load-bearing machinery is the family of gradient-boosted tree ensembles—XGBoost, LightGBM, and NGBoost—combined with two methodological innovations. The hybrid tuning strategy restricts hyperparameter search to one sub-experiment and transfers the chosen settings to all other experiments, which is meant to test generalization rather than per-experiment optimization. The evaluation mechanism defines a 'score rate' that compares each model's error to both the best error and the Black–Scholes error, together with a weighted scheme that up-weights sub-experiments using Black–Scholes inputs; higher scores indicate better performance, and the mechanism is designed to let financial theory participate in model evaluation. The experiments also use a training set that mixes 2020 and 2021 data to create what the paper calls a noise-controlled training set.

What would settle it

Re-run the input and moneyness experiments with a strictly chronological split—train on January–August 2020 only, test on September–December 2020—and compare RMSE rankings. If LGBM, NGBoost, or XGBoost lose their lead, or if the denoised-training-set effect in the noise experiment reverses, the central claim of ensemble superiority in these setups would be contradicted.

Watch

Extended reading notes

Core claim

In this study, the central discovery is that ensemble methods, especially LGBM, NGBoost, and XGBoost, consistently outperform other models in both major experiments, as measured by RMSE on CSI 300 index options. The paper argues that the structural flexibility and regularization of gradient-boosted trees give them an advantage over MLP and LSTM neural networks, genetic algorithms, and the classical Black–Scholes model. It further claims that a novel experimental strategy—tuning hyperparameters only in an initial sub-experiment and inheriting them elsewhere—improves robustness and mirrors how financial practitioners reuse working models, and that its scoring and weighted evaluation mechanism, which anchors comparisons to Black–Scholes, shows the theoretical model contributes positively to the experiment. The noise and sliding-window experiments are presented as evidence that flexibility and noise robustness are not simply opposed, but the conclusions there are explicitly left open.

Load-bearing premise

The ranking of models rests on treating a training set that includes 2021 data as a valid noise-controlled basis for predicting a 2020 test period; if that temporal mixing introduces look-ahead information, the reported robustness and accuracy results would need to be reassessed.

Editorial extensions

If this is right

  • Practitioners pricing index options can expect gradient-boosted tree ensembles (LGBM, NGBoost, XGBoost) to give lower RMSE than MLP, LSTM, GA, and Black–Scholes on similar data.
  • The score-rate mechanism provides a template for evaluations in which a theoretical finance model is used as a reference point, not just an accuracy benchmark.
  • The transfer of hyperparameters across experiments, if accepted, reduces the computational cost of tuning while testing generalization.
  • The sliding-window experiments imply that the benefit of local-feature extraction depends on moneyness: it helps OTM options more than ITM or ATM ones.
  • The noise experiments single out MLP as robust to the paper's denoised training set, while LGBM and XGBoost show consistent performance.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the training set includes data from September–December 2021 while the test set is September–December 2020, a standard chronological split would need to be run to confirm that the reported rankings are not inflated by look-ahead information; this is an extension the paper does not make.
  • The ensemble advantage may transfer to other option markets (e.g., S&P 500 or FTSE 100) and other derivatives, but the paper only examines CSI 300 index options, so the breadth of the claim is untested.
  • The proposed score-rate formulas could be adapted to other asset-pricing tasks where a closed-form benchmark exists.
  • The paper's tentative flexibility/noise trade-off suggests a testable hypothesis: models that capture local information well should degrade more under noise; a dedicated study varying window length and noise level jointly could sharpen this.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The manuscript compares ensemble learning methods (LGBM, XGBoost, NGBoost, CatBoost, DeepForest, RF) with classical ML models (MLP, LSTM, GA) and Black-Scholes/BSM baselines for pricing CSI 300 index options. It reports four experiments: input-feature sets, moneyness subsets, sliding-window on/off, and a noise-robustness comparison. The authors claim that gradient boosting ensembles (LGBM, NGBoost, XGBoost) consistently outperform the other models, and they propose a 'parameter transfer' experimental strategy plus a scoring/weighted-evaluation mechanism intended to embed financial theory into model assessment.

Significance. If the empirical results were valid, the paper would provide a useful benchmark comparison of tree-based ensembles against neural networks and genetic algorithms on real Chinese index options, and the proposed theory-weighted evaluation mechanism might be a reproducible evaluation template. The study is transparent in reporting full RMSE/MSE tables and the data source, and it addresses practically relevant dimensions such as moneyness and local-feature extraction. However, the significance is substantially limited by a temporal leakage in the experimental design and by internal inconsistencies in the proposed scoring formulas; these issues affect the main empirical claims rather than only their presentation.

major comments (3)
  1. [Section 4.1, Data and data processing method] The training set is stated to include September-December 2021 data while the test set is September-December 2020, and the 2021 data is described as 'denoised information' without any denoising procedure being defined. This is a look-ahead bias: models trained on observations after the test period can use information that would not be available in a real out-of-sample pricing exercise. Because this split underlies Tables 2, 4, 6, and 7, the reported RMSE rankings and the noise-robustness comparison do not constitute a valid evaluation of predictive performance. The noise experiment in Table 7 compares the baseline training set against the future-augmented set, so it measures training-set composition rather than any denoising method. The experiments need to be rerun with a causal chronological split, or the authors must provide a detailed, leakage-free justification for why future data may be used to train a model evaluated on an earlier period.
  2. [Section 3.2, Eqs. (4)-(5)] The definitions of the Score Rate are internally inconsistent. The text says e is 'the smallest numerical error observed within that sub-experiment,' but Tables 3 and 5 report different score rates for every model, which is only possible if e is the evaluated model's own error. The later phrase 'unless the numerical error of the evaluated model is greater than that of the BS model' also implies e is model-specific. As written, Eqs. (4)-(5) yield a single scalar per sub-experiment and cannot reproduce the tables. Please rewrite the definitions with explicit notation such as e_i for model i, state which quantity is the denominator in each formula, and give the weighted aggregation formula used across sub-experiments.
  3. [Section 5, Discussion (BS-model inference) and Section 6, Conclusion] The claim that a lower Score Rate when the BS model is included 'implies that the BS model contributes positively' is not established. For a fixed model error e, Score_BS < Score_NP is equivalent to E_BS < E_NP, where E_BS and E_NP are the denominators in Eqs. (4) and (5). Thus the comparison only reflects the relative size of the largest errors used as denominators; it does not measure the BS model's causal contribution to the experiment. The corresponding conclusion in Section 6 should be removed or supported by a formal argument that rules out this alternative explanation.
minor comments (5)
  1. [Section 3.1] The sentence 'the parameter settings of the remaining experiments in Table 6 are directly inherited from In1' appears to reference the wrong table; the input experiment sub-experiments are in Table 2, so the intended reference is likely Table 2 or a general statement about all later experiments.
  2. [Section 4.2] There is a typo in 'feild' which should be 'field.'
  3. [References] Reference [15] contains an apparent typo in the author name 'Ivas,cu'; the comma should be removed or replaced with the proper diacritic.
  4. [Table 6 versus Tables 2 and 4] The RMSE magnitudes in Table 6 are orders of magnitude smaller than those in Tables 2 and 4; please clarify whether the output variable is scaled differently in the sliding-window experiments or whether a different subset of data is used, so readers can compare results across tables.
  5. [Section 2.2, Eq. (1)] In Eq. (1), the notation f(x_i) under the derivative is ambiguous; the subscript f_{m-1}(x) in the evaluation point suggests the derivative should be written with an explicit model index, e.g., ∂L(y_i, f(x_i))/∂f(x_i) evaluated at f = f_{m-1}(x).

Circularity Check

1 steps flagged · score 4.0 of 10

Partial circularity in one supporting evaluation claim: the 'BS contributes positively' result is forced by the score-rate definition; the central ensemble comparison is not circular.

  1. self definitional [Section 3.2, Equations (4)-(5); Section 5 Discussion, BS-model paragraph; Tables 3 and 5]
    "Score Rate= EBS −e EBS ×100%(4) ... Score Rate= E−e E ×100%(5) ... a lower overall Score Rate in the presence of the BS model (i.e., using EBS ) suggests that the performance gap among models has been reduced. This implies that the inclusion of the BS model contributes positively to the overall experiment. ... The necessity of incorporating the BS model is clearly demonstrated, as the Score Rate is significantly lower when it is included compared to when it is not."

    The 'demonstration' is a tautology of the chosen denominators. For any sub-experiment with e>0, if the BS error E_BS is below the largest ML error E, then (E_BS-e)/E_BS < (E-e)/E purely by algebra; no model outputs or data are needed. Thus the observed 'lower Score Rate with BS' and the inference that BS 'contributes positively' are built into the definition of the two score rates. The conclusion is equivalent to the inequality E_BS < E, which is already visible in the RMSE tables, so the evaluation mechanism cannot provide independent empirical evidence for that conclusion. The metric was constructed so that any reference model with intermediate error would automatically produce the claimed effect.

full rationale

The paper's central claim—that LGBM, NGBoost, and XGBoost consistently outperform other models in the input and moneyness experiments—rests on RMSE comparisons against external baselines and other ML models. No fitted parameter is renamed as a prediction, and the parameter-transfer strategy is a stated design choice rather than a circular reduction of one experiment's result into another's. The self-citations to Li and Huang [17] for the data-processing strategy and the dividend-rate variable q are methodological inputs, not load-bearing derivations of the ensemble result. The temporal-leakage concern from adding September-December 2021 data to a training set evaluated on September-December 2020 is a serious validity threat, but it is a data-construction flaw, not a circularity. The one genuine circular step is in the evaluation mechanism: the claim that BS 'contributes positively' because the BS-based Score Rate is lower than the non-BS Score Rate is algebraically guaranteed whenever E_BS < E, so that conclusion is definitional rather than empirical. Hence partial circularity in a supporting claim, while the central model comparison retains independent content.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new theoretical entities. The free parameters are the undisclosed hyperparameters, hand-chosen weights, arbitrary moneyness boundaries, and unreported GARCH parameters. The most consequential axiom is the use of future data in the training set, which is an ad hoc assumption that breaks standard temporal validation.

free parameters (4)
  • Model hyperparameters = not listed
    Tuned only on In1 or ALL, then transferred to other experiments; exact values are never disclosed.
  • Evaluation weights = 1, 1, 2, 2, 1, 1
    Hand-assigned weights for the six input sub-experiments to emphasize In3 and In4, which use Black-Scholes-style inputs.
  • Moneyness boundaries = 0.96, 1.04
    Arbitrary cutoffs defining ATM, ITM, and OTM buckets; no justification given.
  • GARCH volatility parameters = not listed
    The volatility feature is computed with a GARCH model, but the fitted parameters are not reported.
assumptions (4)
  • standard math Black-Scholes and Black-Scholes-Merton formulas are valid pricing benchmarks
    Used as the theoretical reference in the evaluation mechanism (Section 2.7) and as baselines in all tables.
  • domain assumption The resset.com dataset is complete and accurately represents CSI 300 options
    All pricing data is purchased from www.resset.com; the paper provides no independent validation of data quality.
  • ad hoc to paper Future 2021 data can serve as denoised training information for a 2020 test set
    Section 4.1 states the training set includes Sep-Dec 2021 data while the test set is Sep-Dec 2020. This assumption introduces look-ahead bias and is core to the noise experiment.
  • domain assumption Sliding window captures local features relevant to option prices
    The sliding window experiments assume that local temporal information improves pricing, cited from prior work but not re-derived here.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Option Pricing Using Ensemble Learning." pith.science (2026). https://pith.science/paper/HAYUOGGO

@misc{pith2026250605799,
  author       = {Pith},
  title        = {Pith review of: Option Pricing Using Ensemble Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HAYUOGGO}},
  note         = {Machine review of arXiv:2506.05799}
}
read the original abstract

Ensemble learning is characterized by flexibility, high precision, and refined structure. As a critical component within computational finance, option pricing with machine learning requires both high predictive accuracy and reduced structural complexity-features that align well with the inherent advantages of ensemble learning. This paper investigates the application of ensemble learning to option pricing, and conducts a comparative analysis with classical machine learning models to assess their performance in terms of accuracy, local feature extraction, and robustness to noise. A novel experimental strategy is introduced, leveraging parameter transfer across experiments to improve robustness and realism in financial simulations.Building upon this strategy, an evaluation mechanism is developed that incorporates a scoring strategy and a weighted evaluation strategy explicitly emphasizing the foundational role of financial theory. This mechanism embodies an orderly integration of theoretical finance and computational methods. In addition, the study examines the interaction between sliding window technique and noise, revealing nuanced patterns that suggest a potential connection relevant to ongoing research in machine learning and data science.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

30 extracted references · 26 canonical work pages

  1. [1]

    Improving the pricing of op- tions: A neural network approach

    Anders, U., Korn, O., Schmitt, C., 1998. Improving the pricing of op- tions: A neural network approach. Journal of forecasting 17, 369–388

    work page 1998

  2. [2]

    Black–scholes versus artificial neural networks in pricing ftse 100 options

    Bennell, J., Sutcliffe, C., 2004. Black–scholes versus artificial neural networks in pricing ftse 100 options. Intelligent Systems in Accounting, Finance & Management: International Journal 12, 243–260

    work page 2004

  3. [3]

    The pricing of options and corporate liabilities

    Black, F., Scholes, M., 1973. The pricing of options and corporate liabilities. Journal of political economy 81, 637–654

    1973

  4. [4]

    Random forests

    Breiman, L., 2001. Random forests. Machine learning 45, 5–32

    work page 2001

  5. [5]

    Classifica- tion and regression trees

    Breiman, L., Friedman, J., Olshen, R.A., Stone, C.J., 2017. Classifica- tion and regression trees. Routledge

    work page 2017

  6. [6]

    Introduction to boosted trees

    Chen, T., 2014. Introduction to boosted trees. University of Washington Computer Science 22, 14–40

    work page 2014

  7. [7]

    Xgboost: A scalable tree boosting system, in: Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, pp

    Chen, T., Guestrin, C., 2016. Xgboost: A scalable tree boosting system, in: Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, pp. 785–794

    work page 2016

  8. [8]

    Catboost: gradient boost- ing with categorical features support

    Dorogush, A.V., Ershov, V., Gulin, A., 2018. Catboost: gradient boost- ing with categorical features support. arXiv preprint arXiv:1810.11363

    arXiv 2018

Show all 30 references
  1. [9]

    Ngboost: Natural gradient boosting for probabilistic pre- diction, in: International conference on machine learning, PMLR

    Duan, T., Anand, A., Ding, D.Y., Thai, K.K., Basu, S., Ng, A., Schuler, A., 2020. Ngboost: Natural gradient boosting for probabilistic pre- diction, in: International conference on machine learning, PMLR. pp. 2690–2700

    2020

  2. [10]

    Greedy function approximation: a gradient boost- ing machine

    Friedman, J.H., 2001. Greedy function approximation: a gradient boost- ing machine. Annals of statistics , 1189–1232. 18

    2001

  3. [11]

    3d tensor-based deep learning models for predicting option price, in: 2021 International Conference on Information Science and Communications Technologies (ICISCT), IEEE

    Ge, M., Zhou, S., Luo, S., Tian, B., 2021. 3d tensor-based deep learning models for predicting option price, in: 2021 International Conference on Information Science and Communications Technologies (ICISCT), IEEE. pp. 1–6

    2021

  4. [12]

    Gou, X., He, L., Zhang, Y., Wang, K., Liu, X., Yang, T., Wang, Y., Cui, B., 2020. Sliding sketches: A framework using time zones for data stream processing in sliding windows, in: Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining,...

    2020

  5. [13]

    Guo, H., Gelfand, S.B., 1992. Classification trees with neural network feature extraction, in: Proceedings 1992 IEEE Computer Society Con- ference on Computer Vision and Pattern Recognition, IEEE Computer Society. pp. 183–184

    1992

  6. [14]

    A nonparametric ap- proachtopricingandhedgingderivativesecuritiesvialearningnetworks

    Hutchinson, J.M., Lo, A.W., Poggio, T., 1994. A nonparametric ap- proachtopricingandhedgingderivativesecuritiesvialearningnetworks. The journal of Finance 49, 851–889

    1994

  7. [15]

    Option pricing using machine learning

    Ivas,cu, C.F., 2021. Option pricing using machine learning. Expert Systems with Applications 163, 113799

    2021

  8. [16]

    Lightgbm: A highly efficient gradient boosting decision tree

    Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q., Liu, T.Y., 2017. Lightgbm: A highly efficient gradient boosting decision tree. Advances in neural information processing systems 30

    2017

  9. [17]

    Option pricing with convolutional kolmogorov- arnold networks

    Li, Z., Huang, Q., 2024. Option pricing with convolutional kolmogorov- arnold networks. arXiv preprint arXiv:2412.01224

    2024 arXiv

  10. [18]

    Spectrum of variable- random trees

    Liu, F.T., Ting, K.M., Yu, Y., Zhou, Z.H., 2008. Spectrum of variable- random trees. Journal of Artificial Intelligence Research 32, 355–384

    2008

  11. [19]

    A neural network model for estimating option prices

    Malliaris, M., Salchenberger, L., 1993. A neural network model for estimating option prices. Applied Intelligence 3, 193–206

    1993

  12. [20]

    New insights and perspectives on the natural gradient method

    Martens, J., 2014. New insights and perspectives on the natural gradient method. arXiv preprint arXiv:1412.1193 . 19

    2014 arXiv

  13. [21]

    A novel stacked generalization ensemble-based hy- brid lgbm-xgb-mlp model for short-term load forecasting

    Massaoudi, M., Refaat, S.S., Chihi, I., Trabelsi, M., Oueslati, F.S., Abu-Rub, H., 2021. A novel stacked generalization ensemble-based hy- brid lgbm-xgb-mlp model for short-term load forecasting. Energy 214, 118874

    2021

  14. [22]

    Theory of rational option pricing

    Merton, R.C., 1973. Theory of rational option pricing. The Bell Journal of Economics and Management Science , 141–183

    1973

  15. [23]

    Classification under streaming emerging new classes: A solution using completely-random trees

    Mu, X., Ting, K.M., Zhou, Z.H., 2017. Classification under streaming emerging new classes: A solution using completely-random trees. IEEE Transactions on Knowledge and Data Engineering 29, 1605–1618

    2017

  16. [24]

    Approximation theory of the mlp model in neural networks

    Pinkus, A., 1999. Approximation theory of the mlp model in neural networks. Acta numerica 8, 143–195

    1999

  17. [25]

    Ensemblelearning

    Polikar, R., 2012. Ensemblelearning. Ensemblemachinelearning: Meth- ods and applications , 1–34

    2012

  18. [26]

    Neural networks for option pricing and hedg- ing: a literature review

    Ruf, J., Wang, W., 2019. Neural networks for option pricing and hedg- ing: a literature review. arXiv preprint arXiv:1911.05620

    2019 arXiv

  19. [27]

    Sajid, S.W., Hasan, M., Rabbi, M.F., Abedin, M.Z., 2023. An ensemble lgbm (light gradient boosting machine) approach for crude oil price pre- diction, in: Novel Financial Applications of Machine Learning and Deep Learning: Algorithms, Product Modeling, and Applications. Spring...

    2023

  20. [28]

    Real option valuation with neural networks

    Taudes, A., Natter, M., Trcka, M., 1998. Real option valuation with neural networks. Intelligent Systems in Accounting, Finance & Manage- ment 7, 43–52

    1998

  21. [29]

    Option price forecasting using neural networks

    Yao, J., Li, Y., Tan, C.L., 2000. Option price forecasting using neural networks. Omega 28, 455–466

    2000

  22. [30]

    Deep forest

    Zhou, Z.H., Feng, J., 2019. Deep forest. National science review 6, 74–86. 20

    2019

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.