International Journal of Scientific Research in Computer Science, Engineering and Information Technology

ISSN : 2456-3307 (www.ijsrcseit.com)

doi : https://doi.org/10.32628/CSEIT217152

Gamma Scalping using Neural Network with Heikinashi

Transformed Data and its Performance
Nishchal Sharma
Department of Computer Science, Govt. College Kullu, Himachal Pradesh, India

ABSTRACT

Article Info This research paper explores the usefulness of a neural network in portfolio
Volume 7, Issue 1 management using gamma scalping. Since gamma scalping requires both stock
Page Number: 255-259 options position as well as the underlying stock to be added or removed frequently.
Publication Issue : It becomes important to optimize such transactions. The underlying portfolio
January-February-2021 provides enough data points that may be used to calculate precise points of time to
Article History initiate stock addition or removal.
Accepted : 20 Feb 2021 Keywords: Gamma Scalping, Heikinashi data, Neural Network, portfolio
Published : 28 Feb 2021 management.

I. INTRODUCTION However if the underlying create enough gamma, the

institution can book it and overcome the theta loss
Electronic Portfolio management has become a key that may arise from such position. But this requires
method for many investment banks or portfolio precise timing as booking too early may waste future
managers. Although using equity alone may expose gamma gains and booking too late may result from
many such institutions to high risk. So many huge theta losses.
institution create hedges with the help of options.
This is where we use an artificial neutral network to
Gamma scalping is technique where an institution understand the underlying pattern of stock
hedges its open option position with the help of movement and create hot spots where a portfolio
stocks to create a delta neutral position. This method may book its gamma value.
create profits when the underlying moves to either
side. So the institution’s risk for a particular position As it has been shown [1],[2] that neutral network
vanishes. However such position create a theta loss perform extremely well in gauging sentiment and
from the options that are being held by institution. direction of underlying stock. Many researcher [3],[4]
Institution may add or remove stock in order to has demonstrated the usefulness of neural network in
remain delta neutral, this creates a portfolio that is stock selections. The use of Markov chain [4],[5] and
independent from direction but suffers losses with hidden Markov model has provided better signals for
the passage of time. forecasting. Researcher [6] have provided enough

Copyright: © the author(s), publisher and licensee Technoscience Academy. This is an open-access article distributed
under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-
commercial use, distribution, and reproduction in any medium, provided the original work is properly cited
255
 Nishchal Sharma et al Int. J. Sci. Res. Comput. Sci. Eng. Inf. Technol, January-February-2021; 7 (1) : 255-259

evidence the use of SVM and [7] genetic algorithms Lets say that Q represents the entire portfolio with P
to be useful in chaotic systems. being Puts and C being calls. Lets say that n
represents no of lots of both calls and puts.
Better price discoveries and stock market sentiment So Q= n(P+C)
analysis [8],[9] has been studied for many foreign lets say that (P+C) has a combined daily theta value T.
exchanges. Many researchers have used various So net daily theta value of entire portfolio is = nT
method to transform [10],[11],12]stock market data Lets say Gamma generated over a period of Z is G.
and provide better signals using neural network. Net theta loss for Z days= nTZ
Net Gamma Gain is = G
This paper describes a system that provides action Total value of stock present in portfolio = S
points using neural network that allows portfolio For a profitable portfolio we need to have G> nTZ.
managers to book profits in a delta neutral positions. If G>nTZ we can initiate to add or remove stock to
again make entire portfolio delta neutral again. This
II. METHODS AND MATERIAL is because one of the call and puts will have a greater
delta value and entire portfolio will either have a
Here we discuss the basic method of delta neutral positive delta or a negative delta.
positions and Heikinashi Transformation method. Over all value of portfolio is given by
We will use a Heikinashi transformed data to train Q=n(P+C)+S
network and then create a delta neutral portfolio. We measure the percentage gain or loss value of
Portfolio by Qnet over 30 days.
Heikinashi Transformation rules: For a positive delta , we sell stocks to book gamma.
Let Ocurrent, Hcurrent , Lcurrent, Ccurrent represents current For a negative delta , we add stocks to book gamma.
open, high, low, close values. Our system uses a [4,3,3,1] neutral network with 2
hidden layers. The final output is a value between
Let OPrev, HPrev, LPrev, CPrev represents Previous [0,1.0] 0 being low action point and 1being the
day/period open, high, low, close values. highest action point.

Then Heikin-Ashi values (HA) are calculated as:- III. RESULTS AND DISCUSSION
HA-Close = (Ocurrent + Hcurrent +Lcurrent+Ccurrent)/ 4
HA-Open = (HA-OpenPrev + HA-ClosePrev) / 2 Following are the result from the system. As is
HA-High = Maximum of the Hcurrent , HA-Open or HA-Close evident from below diagram, we have a poor success
HA-Low = Minimum of the LCurrent , HA-Open or HA-Close below 30% prediction value. Above 30% prediction
rate we start to see mainly positive results. The
A. Delta neutral position
standard deviation is on the high side, which may
Delta neutral position can be created by buying both
denote a fine tuning may be needed. But over all
call and put option , such position have positive
results are vary convincing that we can fine tune
gamma but negative theta value.
such system for Gamma scalping.

Volume 7, Issue 1, January-February-2021 | http://ijsrcseit.com

256
 Nishchal Sharma et al Int. J. Sci. Res. Comput. Sci. Eng. Inf. Technol, January-February-2021; 7 (1) : 255-259

V. REFERENCES
Gamma Scalping
1
[1] Burton G. Malkiel, E.F.F., 1970. Efficient
0.5 Capital Markets: A Review of Theory and
Empirical Work. J. Finance
Qnet

0
Probabil…

Column J [2] Chang, T.-S., 2011. A comparative study of

0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
-0.5 artificial neural networks, and decision trees
for digital game content stocks price
-1
Probability of Booking prediction. Expert Syst. Appl. 38, 14846–14851.
[3] Chen, A.-S., Leung, M.T., Daouk, H., 2003.
Figure 1. Results of simulation for a [4,3,3,1] neural Application of neural networks to an emerging
network. financial market: forecasting and trading the
Taiwan Stock Index. Comput. Oper. Res. 30,
901–923.
Standard Deviation 0.49
[4] Fernandez-Rodrıguez, F., Gonzalez-Martel, C.,
Average 0.07
Sosvilla-Rivero, S., 2000. On the profitability of
Average Portfolio Return below 30%
technical trading rules based on artificial
rate -0.12
neural networks:: Evidence from the Madrid
Average Portfolio Return over 30%rate 0.15
stock market. Econ. Lett. 69, 89–94.
[5] Guo, Z., Wang, H., Yang, J., Miller, D.J., 2015.
A. Structure of Neural Network A Stock Market Forecasting Model Combining
This paper doesn’t test the system with different Two-Directional Two-Dimensional Principal
structure of neural network, which may be Component Analysis and Radial Basis Function
significant for a fine tuned performance. This also Neural Network.
opens up the possibility of further research in above [6] Hassan, M.R., Nath, B., 2005. Stock market
method. forecasting using hidden Markov model: a new
approach, in: Intelligent Systems Design and
B. Title and Author Details
Applications, 2005. ISDA’05. Proceedings. 5th
Author: Assistant Professor Nishchal Sharma is a
International Conference on. IEEE, pp. 192–
faculty at Govt. College Kullu of Computer Science
196.
and Application.
[7] Huang, W., Nakamori, Y., Wang, S.-Y., 2005.
Forecasting stock market movement direction
IV. CONCLUSION
with support vector machine. Comput. Oper.
Res. 32, 2513–2522.
It is evident that a gamma scalping system with
[8] Kaboudan, M.A., 2000. Genetic programming
neural network has potential and a further study may
prediction of stock prices. Comput. Econ. 16,
be needed to fine tune such system to have an
207–236.
acceptable profitable range.
[9] Kara, Y., Acar Boyacioglu, M., Baykan, Ö.K.,
2011. Predicting direction of stock price index
movement using artificial neural networks and
support vector machines: The sample of the

Volume 7, Issue 1, January-February-2021 | http://ijsrcseit.com

257
 Nishchal Sharma et al Int. J. Sci. Res. Comput. Sci. Eng. Inf. Technol, January-February-2021; 7 (1) : 255-259

Istanbul Stock Exchange. Expert Syst. Appl. 38, modified RBF neural network, in: 2010 IEEE
5311–5319. 17Th International Conference on Industrial
[10] Khan, Z.H., Alin, T.S., Hussain, M.A., 2011. Engineering and Engineering Management (IE
Price prediction of share market using artificial EM). Presented at the 2010 IEEE 17Th
neural network (ANN). Int. J. Comput. Appl. International Conference on Industrial
22, 42–47. Engineering and Engineering Management (IE
[11] Kim, K., Han, I., 2000. Genetic algorithms EM), pp. 424–427.
approach to feature discretization in artificial [19] Vanstone, B., Finnie, G., 2009. An empirical
neural networks for the prediction of stock methodology for developing stockmarket
price index. Expert Syst. Appl. 19, 125–132. trading systems using artificial neural
[12] Kumar Chandar, S., Sumathi, M., Sivanandam, networks. Expert Syst. Appl. 36, 6668–6680.
S.N., 2016. Prediction of Stock Market Price [20] Xiao, Y., Xiao, J., Lu, F., Wang, S., 2013.
using Hybrid of Wavelet Transform and Ensemble ANNs-PSO-GA Approach for Day-
Artificial Neural Network. Indian J. Sci. ahead Stock E-exchange Prices Forecasting. Int.
Technol. 9. J. Comput. Intell. Syst. 6, 96–114.
[13] Kuo, R.J., Chen, C.H., Hwang, Y.C., 2001. An [21] Zekic, M., 1998. Neural network applications
intelligent stock trading decision support in stock market predictions-a methodology
system through integration of genetic analysis, in: Proceedings of the 9th
algorithm based fuzzy neural network and International Conference on Information and
artificial neural network. Fuzzy Sets Syst. 118, Intelligent Systems. Citeseer, pp. 255–263.
21–45. [22] Zhang, Y., Wu, L., 2009. Stock market
[14] Martinez, L.C., da Hora, D.N., de M Palotti, prediction of S&P 500 via combination of
J.R., Meira Jr, W., Pappa, G.L., 2009. From an improved BCO approach and BP neural
artificial neural network to a stock market day- network. Expert Syst. Appl. 36, 8849–8854.
trading system: A case study on the BM&F [23] “Stock Market Volatility and Learning - ADAM
BOVESPA, in: Neural Networks, 2009. IJCNN - 2016 - The Journal of Finance - Wiley Online
2009. International Joint Conference on. IEEE, Library.”
pp. 2006–2013. [24] “Data | Free Full-Text | Current Core
[15] McCulloch, W.S., Pitts, W., 1943. A logical Competencies Trend of Small to Medium
calculus of the ideas immanent in nervous Enterprises (SMEs) in China—A Concurrent
activity. Bull. Math. Biophys. 5, 115–133. Comprehensive Evaluation and Active
[16] Mizuno, H., Kosaka, M., Yajima, H., Komoda, Learning Study of Newly Listed Chinese Stocks
N., 1998. Application of neural network to from 2015 through 2017.”
technical analysis of stock market prediction. [25] “The contribution of technology selection to
Stud. Inform. Control 7, 111–120. core competencies - ScienceDirect.”
[17] Perwej, Y., 2012. Prediction of the Bombay [26] “Using artificial neural network models in
Stock Exchange (BSE) Market Returns Using stock market index prediction - ScienceDirect.”
Artificial Neural Network and Genetic [27] “Application of artificial neural network for the
Algorithm. J. Intell. Learn. Syst. Appl. 4, 108– prediction of stock market returns: The case of
119. the Japanese stock market - ScienceDirect.”
[18] Sun, B., Li, T. k, 2010. Forecasting and [28] “Predicting stock and stock price index
identification of stock market based on movement using Trend Deterministic Data

Volume 7, Issue 1, January-February-2021 | http://ijsrcseit.com

258
 Nishchal Sharma et al Int. J. Sci. Res. Comput. Sci. Eng. Inf. Technol, January-February-2021; 7 (1) : 255-259

Preparation and machine learning techniques -

ScienceDirect.”
[29] “Predicting the Direction of Stock Market Index
Movement Using an Optimized Artificial
Neural Network Model.”
[30] Sharma, Nishchal. (2020). Index Option Greek
Analysis with Heikin-Ashi Transformed Data
and Its prediction with Artificial Neural
Network. International Journal of Scientific
Research in Computer Science, Engineering
and Information Technology. 166-169.
10.32628/CSEIT206136.
[31] “A hybrid stock trading framework integrating
technical analysis with machine learning
techniques - ScienceDirect.”
[32] “A Deep Neural-Network Based Stock Trading
System Based on Evolutionary Optimized
Technical Analysis Parameters - ScienceDirect.”
[33] Sharma, Nishchal & Chauhan, Chaman. (2019).
Heikin-Ashi Transformation and Vix Index
data for Stock Market Index Prediction and It's
Effects. International Journal of Scientific
Research in Computer Science, Engineering
and Information Technology. 363-365.
10.32628/CSEIT195195.

Cite this article as :

Nishchal Sharma, "Gamma Scalping using Neural

Network with Heikinashi Transformed Data and its
Performance", International Journal of Scientific
Research in Computer Science, Engineering and
Information Technology (IJSRCSEIT), ISSN : 2456-
3307, Volume 7 Issue 1, pp. 255-259, January-
February 2021. Available at
doi : https://doi.org/10.32628/CSEIT217152
Journal URL : https://ijsrcseit.com/CSEIT217152

Volume 7, Issue 1, January-February-2021 | http://ijsrcseit.com

259