Я на днях просматривал исследование "DO TRENDS EXIST IN THE STOCK MARKETS: AN ANALYSIS BASED ON TAYLOR'S METHODOLOGY"

[www.researchgate.net]


Там по методике некого Тейлора проверяют есть ли трендовые участки или все укладывается в модtль случайного блуждания. Методика, как мне показалась похожа на вашу.

Так вот, выводы там не настолько радужные, у развивающихся рынков в целом есть тренды, а развитых уже нет. Но, например в России, средняя продолжительность такого не стохастического тренда - не более двух дней, что не дает возможности на нем заработать, т.к. это время нужно для его определения.

Цитата:

With respect to *U statistic, the results shown in Table 2 point out the following: Of 14 asset indexes of developed countries, the *U statistic accepts the null hypothesis of random walk (*1.65U <, in a one-tail N(0,1) test with 5% of confidence) in the first six asset indexes of Table 2, that is, Dow Jones, S&P, Nasdaq, FYSE100 UK, Nikkei and DAX of Germany. So, trends are detected in 8 out of 14 countries (*1.65U >: France, Italy, Spain, Holland, Australia, South Africa, Luxemburg and Israel. However, the mean trend duration is only higher than one day (m>1) in the cases of Italy, Holland, Australia and Luxemburg. On the other hand if *U statistic is applied to the index futures markets in the developed countries, the null of random walk is accepted in all cases. This fact reflects that for the asset indexes CAC40 of France, MIB30 of Italy, IBEX35 of Spain, AEX of Holland, ASX of Australia and JSE of South Africa, its index futures markets are more efficient that the proper spot market. In the case of Luxemburg and Israel the index futures series are not available.
For the BRIC countries the *U statistic accepts the existence of trends in the case of Russia, with a mean trend duration lower than 2 days, and China, with a mean trend duration of 9 days.

For the Asia-Pacific Securities the *U statistic accepts the existence of trends for 5 out of 6 asset indexes: Hang Seng of Hong Kong (m=5), TAIEX of Taiwan (m=10), SGX of Singapore (m<2) and VNI of Vietnam (m=1. Like before, no index futures markets available accept the existence of trends.

For the group of the rest of emergent markets, the *U statistic accepts the existence of trends in 7 out of 9 cases, specifically in the IBC of Venezuela (m=5), the BASE of Argentina (m<2), the SOFIX of Bulgaria (m=26), the IGPA of Chile (m=4), the IGBC of Colombia (m=3), the CASE30 of Egypt (m<2), and the NSE20 Kenya (m=74). Index futures data is not available in these markets. Finally, Table 2 shows that the *U statistic accepts the null hypothesis of random walk in the case of MXSE of Mexico and ISE of Turkey asset indexes, revealing a high degree of efficiency in those markets

..