An example: if I want to model how the influence of legal practice has evolved during the course of the pandemic in the Netherlands I can use a timeline to select relevant institutions for my toy world and to locate the moments when they take action, perhaps even exhibit adaptive behavior. In terms of method, this can be used to build a toy world for the Netherlands, but we need to expand that method to determine which toy world variant is the best in terms of an @proxytruths approach.
Considering the position that the game industry has now acquired, it is obvious to call building toy worlds that aim to simulate the pandemic in a legal environment something like building COVID games instead. I cannot yet grasp the qualities of the terminology and will use the terms interchangeably.
I take the timeline for building COVID-game-NL from a graph. Actually two graphs, because COVID-game-NL looked at COVID-game-CN where it all started a few weeks earlier. Based on WHO data, I juxtapose two pandemic timelines, one for China and one for the Netherlands. Both cover the entire period starting in December 2019 and ending as I write this, in early March 2021. See Fig. 1.
The graphs show numbers per day in the form of running weekly averages. The scale is especially important for reading the graphs (the column on the far left). The numbers of new deaths per day are the raw numbers (the black line). The numbers of newly registered infections per day (the red line) are divided by 33 (to be able to show the red and black lines in a single graph). That skews a bit. The top of the first wave in China for example would be, in one and the same picture, a little lower than the Dutch one, for example. One period in the graphs of Fig. 1 equals almost 7 weeks. The title line contains the name of the relevant jurisdiction, the population size used for calculation, the number of deaths during the displayed period and how many deaths per million inhabitants his is. Overall, the Netherlands has 300 times more COVID deaths per million inhabitants than China. The pandemic has been dealt with differently in those two countries. I try to find out which mechanisms played which role in this.
Looking for Clues
To take my Durkheim-based approach seriously, I show Fig. 2 in reminder. I connect day 0 of Fig. 1 with the central starting point of the spiral and continue along that line in the attempt to give a location in Fig 2. to what I infer from Fig. 1. I further assume that knowledge of epidemic models such as the SIR and the SIRD models is known – in particular the exponential forms that the infection growth in the beginning of a pandemic as it is related to the ability of an infected person to infect several other persons before he has recovered (or died).
Looking at Fig. 1 it is clear in the third week (around January 16) that the numbers of infections and deaths are growing noticeably and exponentially. The numbers of infections peak in week 7 (around February 13) to decline exponentially in week 8 (from February 20). Assuming that it will take two weeks for the effects of measures to become visible in the statistics, measures were taken around January 30 (lock-downs + tests + quarantines) that were reinforced a week later (the construction of emergency hospitals in Wuhan). The graph does not clearly show whether and if so when the measures have been terminated. (The peak in week 15 is an administrative correction.) For the rest, the graph should be studied on a different scale to see what else has been discussed in terms of measures. There may have been 3 or 4 local lockdowns to control local outbreaks. I would expect them around June 20, November 27 and January 13. The big difference with the Netherlands is of course that in China the virus is under control from week 14 onwards, while in the Netherlands in week 60 this is by no means the case. In China, strict lockdowns have been implemented to curb incipient COVID outbreaks, supplemented by massive testing and controlled quarantines, including those who leave and enter the area. The lockdown in Wuhan was lifted on April 8 (week 14): the graph shows the virus in China was under control from that moment on.
It seems that a high @proxytruth value COVID-game-CN can be built with simple means: effective and covering bottom-up and top-down communication where normative top-down communication is maintained and followed by the entire population, supplemented with (1) high-quality healthcare and, when necessary, (2) social distancing, (3) mass testing, (4) (controlled) quarantines for who has symptoms or is tested positive, (5) lock-downs, (6) controls at the borders, (7) vigorous investments and (8) appropriate responses to random outbreaks of infections.
To be honest, I don’t expect such a game to be difficult to make, nor do I expect it to be exciting to play – unless the reality starts to diverge. This can be anticipated in the game with the introduction of a stochastic emergence of infections, followed by quarantine measures, test explosions, travel restrictions and, if necessary, a local lock-down. Predicting the latter is likely to be an important topic for (the design of) a @proxytruths-based evaluation method for COVID-games to be discussed in more detail. The placement of those features in Fig. 3 provides added value with the option of registering an intuition about what kind of @proxytruths play a major role in the realization of the measure. I will come back to that later.
Looking at Fig. 1, it becomes clear in the tenth week (around March 10) that the numbers of infections and deaths are growing noticeably and exponentially. The numbers of infections peak in week 14 (around April 8) and then decline more or less linearly. Assuming that it will take two weeks for the effects of measures to become visible in the statistics, measures were taken around 26 March (lock-downs + tests + quarantines). We cannot tell from the graph when measures have been lifted, but it is likely that this happened before another wave arrived, around July 17. Based on the peak in the second wave, a measure must have been taken around October 20 that only worked to a limited extent: around December 1, a third wave started before the second had worn off. The peak of this wave was reached around December 23, so that additional measures are expected to have been taken around December 10. Around 10 February, the downward trend in the daily numbers of new infections has come to an end. Today, March 2, 2021, the measures have nevertheless been partially lifted. It seems that large parts of the citizens can no longer stand it and / or no longer believe in them. There is one ray of hope: the number of deaths per day is still decreasing somewhat, presumably because there is now a vaccine and the vaccination of vulnerable elderly people has begun.
In a later post, I pay attention to whether I can judge the cursory analysis of the Dutch timeline on the basis of recorded observations. Now first another matter. It seems that it is much more difficult for me to design a COVID-game-NL with a high @proxytruth value than a good COVID-game-CN. In principle, the same measures are on the repertoire in the Netherlands as in China. However, we sometimes look at them differently in the Netherlands. This concerns measures 2 to 5 (distancing to lockdown). These are much less of a public law nature in the Netherlands than in China. Our (I am a Dutchman) “intelligent” lockdown is not understood in China and seen as a soft lockdown and considered almost immoral against a life-threatening, explosive pandemic. On the other hand, in the Netherlands the “hard” Chinese line is considered incompatible with Dutch culture and Dutch general understanding of human rights.
Fig. 2 and Fig. 3 provide a starting point to investigate whether we can make one model that can yield two games with high @proxytruths values. Perhaps that will provide clues for finding a method to help determine criteria to identify the better versions of games. For the moment it is my premise that analyzing timelines in terms of Durkheim’s heuristic is a good start, on my way to a more complete method.