GC Directional Forecast Report XU030 GC Financial Research
GC Directional Forecast Report: XU030 GC Financial Research Published 02-Jun-2012 Copyright 2012 Milenyum Research & Development, www.gunduzcapital.com The information contained in this report may not be reproduced or redisseminated in whole or in part without prior written permission from Milenyum Research & Development.
Introduction... 1 Legal Notice... 1 Prediction results... 2 Overview... 2 Success rate... 2 Details... 3 Method... 3 Positive Return Probability... 3 iv
Introduction GC Directional Forecast Models. Direction-of-change forecasts are indispensable for succesfull trading strategies and efficient asset allocation. Gunduz Capital (GC) Financial Research has developed thoroughly tested quantitative models for directional forecasts of Istanbul Stock Exchange listed companies and indices. Objective forecasts. Quantitative models are rigorous mathematical models. They are objective and allow rational, unemotional decision making. Our models are constantly updated in order to quickly respond to rapidly changing market conditions. Evidence of past performance. Besides a prediction, also the out-of-sample performance of our method is given. A model may have a good in-sample performance, but may perform poorly out of sample. In order to assess the predictive power of our models, also backtesting results are presented. Legal Notice Uyari Notu. Burada yer alan bilgiler Milenyum Teknoloji ARGE LTD tarafindan genel bilgilendirme amaci ile hazirlanmistir. Yatirim Danismanligi hizmeti; araci kurumlar, portföy yönetim sirketleri, mevduat kabul etmeyen bankalar ile müsteri arasinda imzalanacak Yatirim Danismanligi sözlesmesi çerçevesinde sunulmaktadir. Burada yer alan yorum ve tavsiyeler, yorum ve tavsiyede bulunanlarin kisisel görüslerine dayanmaktadir. Herhangi bir yatirim aracinin alim-satim önerisi ya da getiri vaadi olarak yorumlanmamalidir. Bu görüsler mali durumunuz ile risk ve gitiri tercihlerinize uygun olmayabilir. Bu nedenle, sadece burada yer alan bilgilere dayanarak yatirim karari verilmesi beklentilerinize uygun sonuçlar dogurmayabilir Burada yer alan fiyatlar, veriler ve bilgilerin tam ve dogru oldugu garanti edilemez; içerik, haber verilmeksizin degistirilebilir. Tüm veriler, Milenyum Teknoloji ARGE LTD tarafindan güvenilir olduguna inanilan kaynaklardan alinmistir. Bu kaynaklarin kullanilmasi nedeni ile ortaya çikabilecek hatalardan Milenyum Teknoloji ARGE LTD sorumlu degildir. IMKB isim ve logosu 'Koruma Marka Belgesi' altinda korunmakta olup izinsiz kullanilamaz, iktibas edilemez, degistirilemez. IMKB ismi altinda açiklanan tüm bilgilerin telif haklari tamamen IMKB'ye ait olup, tekrar yayimlanamaz. VOB isim ve logosu VOB adina tescilli olup, izinsiz kullanilamaz, iktibas edilemez, degistirilemez ve/ veya çogaltilamaz. VOB ismi altinda açiklanan tüm bilgilerin telif haklari tamamen VOB'a ait olup, tekrar yayimlanamaz. Bu kapsamda, hiçbir açik veya örtülü lisans verilmemektedir. VOB ismi ve logosu veya VOB ismi altinda açiklanan bilgiler kullanilarak olusturulan uygulamalardaki hatalardan, eksikliklerden ya da bu bilgilere dayanilarak yapilan islemlerden dogacak her türlü maddi/ manevi zarardan ve her ne sekilde olursa olsun üçüncü kisilerin ugrayabilecegi her türlü zararlardan dolayi VOB sorumlu tutulamaz. Kullanim izni ve sartlari. Kullanicilar bilgi ve dokümanlari, sadece bilgi edinme amaciyla kullanabilir. Burada mevcut bulunan bilgiler, Milenyum Teknoloji ARGE LTD nin önceden yazili olarak izni alinmadan kismen veya tamamen kopya edilemez, uygulamaya konulamaz, dagitilamaz, kiralanamaz, çogaltilamaz, alt lisansla kullandirilamaz, degistirilemez, ileride kullanilmak üzere saklanamaz. Bilgi sayfalarindaki tüm görüntüler Milenyum Teknoloji ARGE LTD tescilli olup, Milenyum Teknoloji ARGE LTDin önceden yazili izni alinmadan ticari amaçli olarak kullanilamaz. Bu web sitesinde yer alan bütün ürün ve hizmetler önceden bildirimde bulunmaksizin degistirilebilir. 1
Prediction results Overview Instrument Report date Time horizon Our expectation Confidence XU030 31-May-2012 1 Month (21 days) Negative Return Moderate For personal assistance with your investment strategies and risk management, please contact us at info@gunduzcapital.com. Success rate During the past year, we made predictions for XU030 every day. In total there were 253 predictions. We made 154 negative return predictions, out of which 112 were actually negative returns. A graphical illustration of this success rate is shown below. For more detailed statistical measures we refer to the next chapter. 2
Details Method Our quantitative model forecasts an estimate for the sign of the return. Several custom indicators which posses predictive power are used during this process. Positive Return Probability Estimate. On 31-May-2012, the estimated probability for a positive return of the instrument XU030 in 21 business days is 46.1094 percent. The 95% confidence interval is also given below. Current date 31-May-2012 Prediction date 29-Jun-2012 Positive return probability (%) 95% Confidence interval 46.1094 [ 37.3435, 56.933 ] Backtesting. For the indicated time period, direction of change predictions for XU030 were made every day. Backtesting period 03-May-2011::02-May-2012 ( 253 business days) Optimal threshold. The above value quantifies the positive return probability. The optimal cutoff value (threshold) obtained from the receiver operating characteristic (ROC) curve after backtesting is given below. Probability cutoff value (%) 55.7103 Statistical measures. The out-of-sample performance is summerized by the confusion matrix: True Positives: 61 False Positives: 38 False Negatives: 42 True Negatives: 112 The True positive (negative) rate quantifies the ability to correctly identify positive (negative) returns. True positive rate (%) True negative rate (%) 59.2233 74.6667 The Positive (Negative) predictive value, or precision rate, is the proportion of cases where returns were positive (negative), when predictions were positive (negative). 3
Details Positive predictive value (%) Negative predictive value (%) 61.6162 72.7273 Confidence guide. Our recommendations are not solely based on the above positive return probability, but also on the position of the cutoff value with respect to the 95% confidence interval: Confidence Moderate Fair Rule if cutoff value in 95% confidence interval if cutoff value not in 95% confidence interval 4