著者

石鎚 英也

出版者

専修大学情報科学研究所

雑誌

情報科学研究 (ISSN:02866048)

巻号頁・発行日

vol.30, pp.1-34, 2010-03-01

---

It is widely known that various anomalies and stylized facts go with stock prices. Volatility clustering and calendar effects in price time series are well-known examples. Some traders take advantage of such market distortions to increase their profits or reduce the risks of trading. Pairs trading is one of their practical and typical ways. In this paper, time-series data of stock prices in the first section of the Tokyo Stock Exchange are investigated, and risks in and rewards for pairs trading is evaluated through computer simulation.

著者

石鎚 英也

出版者

専修大学ネットワーク情報学会

雑誌

専修ネットワーク&インフォメーション = Network and Information (ISSN:13471449)

巻号頁・発行日

vol.28, pp.1-16, 2020-03-10

---

In the world of games and sports, various rating methods have been developed and used in order to rank players or teams according to their real abilities - Elo rating in chess and Massey rating in college football, for instance. Players and teams are re-rated according to the results of matches, and rating scores and rankings can be updated instantly. However, the followings may be pointed out as drawbacks of prevailing rating methods: It possibly takes time to get stable rating values. It is difficult to properly set parameters included in the updating formulae. And rating values tend to be inflationary.In this study, traditional and standard rating methods are extended by Bayesian statistical modeling so as to mitigate such drawbacks. Then, they are applied to real records of games and sports.

著者

石鎚 英也

出版者

専修大学ネットワーク情報学会

雑誌

専修ネットワーク&インフォメーション (ISSN:13471449)

巻号頁・発行日

no.9, pp.3-9, 2006-03