Physical Review X is a new journal published by the American Physical Society (APS), the publisher of Physical Review Letters and Physical Review E, among others. PRX aims to bring valuable and innovative results to the broader physics readership. In scope it is similar to Physical Review Letters, but it has a few important differences, the most notable being open-access for all articles, Popular Summaries for every article, and no page limit. Since August, PRX has published three articles that caught my eye. Here I review the three articles and give my take on how well the journal is living up to its ideals. The verdict: mixed. The three papers vary widely in their editorial quality and accessibility.
The three articles of interest are Natural Human Mobility Patterns and Spatial Spread of Infectious Diseases by Vitaly Belik, Theo Geisel, and Dirk Brockmann, Anomalous Price Impact and the Critical Nature of Liquidity in Financial Markets by B. Tóth, Y. Lempérière, C. Deremble, J. de Lataillade, J. Kockelkoren, and J.-P. Bouchaud, and Predictability of Conversation Partners by Taro Takaguchi, Mitsuhiro Nakamura, Nobuo Sato, Kazuo Yano, and Naoki Masuda.
I most looked forward to reading the first of these three, by Belik, Geisel, and Brockmann. I know Brockmann and even contemplated joining his group after I received my Ph. D. This is the first article of the first issue, so I felt that PRX was off to a good start representing complex systems research. The paper discusses a model of disease spreading that takes community structure into account in new ways and makes some interesting predictions about the rate at which disease spreads through a system. Unfortunately, the article left me scratching my head in a number of places. The most notable frustration was the interchangeable use of a lower-case omega and a lower-case w in much of the paper. Such differences may be of no consequence to the untrained eye, but typically such differences denote related but distinctly different quantities. In my experience, the APS staff usually picks out these errors, and related articles by these three authors published elsewhere exhibit consistent notation. For this paper, the editorial staff gets a thumbs-down for missing these errors.
I know very little about Physics-inspired analyses of financial trading, but I did not think that would impare my initial reading of the paper by Tóth et al. Unfortunately, the authors make use of a number of terms that initially threw me off, incuding “limit orders” and “books.” I tried to infer their meaning from the context, but found that I was still confused. Fortunately, they define “limit orders” and “market orders” in their appendix and “order book” is described on this Wikipedia page. Armed with this knowledge, I understood more of the article on my second reading and found the paper to be more accessible, though still opaque in places. This is acceptable: at least the beginnings and endings of well written articles should be accessible to a broad scientific audience, and one can hardly avoid confusing non-experts with details in the middle. Although the article is not the first to provide a meaningful explanation for square-root price changes that occurr during large meta-orders, it makes fewer assumptions compared with previous work and links their analysis to the ever-intriguing notion of self-organized criticality. Overall, I give PRX one thumb-up for this article.
Unlike Tóth’s paper, Takaguchi et al. discuss research that comes very close to the interests of the Amaral lab, and in contrast to the first two papers, their work is experimental. I was trained as an experimentalist so perhaps my opinion is biased, but I found this article to be exceptionally well written. They adequately contextualize their work, they fully explain their experimental setup, they identify a number of shortcomings in their setup and analysis and address each of them, and they make excellent use of appendices for details that would otherwise obscure the discussion. They conclude that the temporal order of an individual’s conversations in an office-like setting is predictable. To make an example, if you track the people with whom I converse during the day, you will quickly see that I converse most with Julia, less with Peter, infrequently with Nico. You can calculate the probability that I speak with anyone in the lab with the simplest of statistics. Takaguchi showed that if you know my most recent conversation partner (Rufaro, for example), you can predict my next conversation partner with greater likelihood than if you started with static probabilities. Because of the quality of the writing and figures, I give PRX two thumbs up. I hope that this paper sees wide readership.
PRX is a new journal and it is still working to find its place among the journals of the American Physical Society. Hopefully in the coming months and years we will see more articles with the quality and clarity of Takaguchi et al. and less of the editorial snafus exhibited in Belik, Geisel, and Brockmann.
— David Mertens