Search
× Search
Saturday, February 8, 2025

Archived Discussions

Recent member discussions

The Algorithmic Traders' Association prides itself on providing a forum for the publication and dissemination of its members' white papers, research, reflections, works in progress, and other contributions. Please Note that archive searches and some of our members' publications are reserved for members only, so please log in or sign up to gain the most from our members' contributions.

Using ARIMA - GARCH Trading Strategy on the S&P 500

photo

 Jacques Joubert, Quantitative Analyst at NMRQL

 Monday, August 1, 2016

I think that a lot of the people from this group would largely benefit from reading some of the articles on the quant blog aggregator www.quantocracy.com. The following article was sourced from there. https://www.quantstart.com/articles/ARIMA-GARCH-Trading-Strategy-on-the-SP500-Stock-Market-Index-Using-R


Print

9 comments on article "Using ARIMA - GARCH Trading Strategy on the S&P 500"

photo

 private private,

 Monday, August 1, 2016



Quite skeptic on the replicability of this strategy. On a daily frequency asset prices are almost always random walks; this backtesting should be extended a higher number of assets to confirm this result


photo

 Roberto Ordoyo Romero, Systems Solution Architect & Technical leadership

 Tuesday, August 2, 2016



interesting...


photo

 Paulo Ribeiro, CFA, CMT, HEAD OF PRODUCT DEVELOPMENT at CAIXAGEST

 Tuesday, August 2, 2016



fully agree with mattia, anyone can do a very successful trading system with just an asset and a good dose of over fitting


photo

 Robert Curcio, Account Manager/Technical Sales Weatherford International

 Wednesday, August 3, 2016



Hmmmm.....Long Term Capital. Quants and markets = lethal juxtaposition.


photo

 Joaquin Poudereux, System Administrator at Grupo Anaya

 Thursday, August 4, 2016



I can tell you that I have backtested this same strategy for Forex, not randomly subsetting the assets as the article above suggests, just iterating through all the symbols in my Forex basket. It steadily wins about 50% in a year and a half. And it can be improved if lower timeframes are used, or if you invest in the best 5 or 6 assets in stead of just the best asset (the one with highest |p-0.5|). I have already talked to the author of the article to let him know and to thank him for his idea, I will probably be using a similar approach in my real trading.


photo

 Joaquin Poudereux, System Administrator at Grupo Anaya

 Thursday, August 4, 2016



Sorry, my comment was mistaken. I was talking of another article previously aggregated by quantocracy: http://www.tulipquant.com/2016/06/18/montecarlo-and-arima-for-stock-selection/


photo

 Guy R. Fleury, Independent Computer Software Professional

 Sunday, August 7, 2016



Sorry to put my 2 cents here. The article mentioned by @Joaquin is missing in some regards. First the stock selection process. Picking 100 at random is not that good an idea. No one in his right mind should accept this. When you look at the test results, you find penny stocks in which you would have put your total equity as well as stocks trading below 10k shares a day, which would make those trades difficult to execute. Therefore, the selection process needs to be worked on. You can certainly do better with very little effort.

Second, the trading principle itself appears sound. You look at the best performers of recent past (2 years), order them according to relative ARMA slopes, pick the highest of the group and set your day trade for the next day. It should not be at the close. The weakness is in the day trade. One should extend the trading interval somewhat to capture more of this trend following methodology. Catching more of the drift if you catch my drift.


photo

 Guy R. Fleury, Independent Computer Software Professional

 Sunday, August 7, 2016



Another point I did not cover concerning the article mentioned by @Joaquin is that most of the profit generated by that strategy were from random events, outliers, some of which from penny stocks (ex.: COSI listed at $0.23). In fact, only 13 trades (see returns2.csv) out of 177 (7%) generated the bulk of profits (about 80%).

I think that by doing a more reasonable stock selection, and extending the average holding period on more than one stock could greatly help that kind of trading strategy. Because for sure, predicting that you will hit an outlier on your next trade might be pushing it a bit.


photo

 Justin Duval, Consistently Profitable Day Trader

 Wednesday, August 10, 2016



Yeah, not to mention how much you'd push some of the really thin stocks trying to enter and exit. Good points Guy.

Please login or register to post comments.

TRADING FUTURES AND OPTIONS INVOLVES SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS
Terms Of UsePrivacy StatementCopyright 2018 Algorithmic Traders Association