8 ms·

Hi, game author here. First, this URL https://covidgame.info/ https://covidgame.info/ loads the game in English by default. There was a plan to popularize the

Hi, game author here. First, this URL https://covidgame.info/ https://covidgame.info/ loads the game in English by default. There was a plan to popularize the game outside of Czechia but, frankly, we missed the window when major news outlets were interested. To address few points mentioned in the comments: the game was tuned to the specific situation in Czech Republic around end of 2020 / beginning 2021. The goal was to build basic quantitative intuition about the behavior of the epidemiological model (SEIR model) that the general public and decision makers were lacking. The game doesn't allow for zero covid strategy - because that was never realistically considered. Likewise, we after some considerations, we didn't include cheap interventions like extensive testing, contact tracing and better government communication. This doesn't mean that we didn't support such interventions. However, Czech government proved unwilling or incapable to implement such measures and in the short/medium term broad restrictions were the only "levers" the decision makers could pull. There is an interview for a major Czech Internet server about some aspects of the game development: https://translate.google.com/translate?sl=cs&tl=en&u=https://www.seznamzpravy.cz/clanek/naucna-hra-o-pandemii-vznikla-z-frustrace-z-urovne-koronavirove-diskuse-138814 https://translate.google.com/translate?sl=cs&tl=en&u=https:/... We were working together with some real experts in Covid modelling and the model was in my opinion quantitatively sound (though, as we repeatedly said, the model was not optimized for scientific accuracy). At the end the was played by some members of the parliament, party experts and even ministers. We know that the game helped to facilitate consensus between the government and opposition in a very fragile political situation. I believe that, at the end, the game saved hundreds or thousands lives by helping to implement some necessary restrictions sooner than they'd be implemented otherwise.

daeglin 5y ago
BTW, source code is available here: https://github.com/swehq/corona-game https://github.com/swehq/corona-game
- uzakov 5y ago
  This is honestly great! Could you please add a license? I would love to translate this and deploy in other languages.
  - daeglin 5y ago
    Please contact us at info@koronahra.cz . The game is "open source" but we didn't explicitly specify what exactly that means. I consider the owners of the code to be the company swehq.com that did most of the coding.
noduerme 5y ago
Wonderful, sparse, essential simulation. The stark choices and the resulting reminders of what really happened (and how the player matches up) is a great eye-opener for the general public. The fact that it forces you to make decisions in realtime and feel the lag behind them is really brilliant as a thing to put in front of decision-makers in parliament. One of the most difficult things with covid was the inability to see a few weeks into the future, because our minds are not equipped to think in terms of exponentially multiplying consequences. This gives a really elegant, intuitive form to a very conceptually difficult set of variables. The beauty of the concept is its simplicity, as well. It's a bit like an IFR training panel, or early text-based video games. You can't see where the plane is or where it's going, you only see the instrument readouts after you adjust something. Since we have all been flying blind, this is a great experience to put in front of people, simply of what it's like to adjust and wait. Five stars, sending it to all the devs I know, also as a way to make a very engaging sim on a complex topic for people with no prior art =D
- daeglin 5y ago
  Moon Lander was fairly explicitly inspiration for this game :-)
Robotbeat 5y ago
This is really cool. I think it'd be interesting to model decarbonization in a similar manner, with social stability and debt and deaths.
- rob74 5y ago
  This Covid-19 simulation, where you can have a more or less reliable model because you are basically doing it after the fact, already has enough potential for controversy - e.g. the "social stability" metric, is it too sensitive? Not sensitive enough? For decarbonization, where you don't have a reliable model, you will probably be accused by both sides that you're underestimating or overestimating the various parameters involved. Do you want to open this can of worms?
  - zelphirkalt 5y ago
    Well, it would be a game or simulation after all. It could be improved over time with new knowledge. So it wouldn't hurt to try.
  - Robotbeat 5y ago
    Well, we’ve got to! and it would be nice to open up the game to different assumptions. Like, I tried Covid zero in this game, and although I never got to zero it ended up minimizingz the deaths. Games like this can be useful for at least explaining ones ideas even if the games aren’t “realistic”. I think for decarbonization, you wouldn’t need to take as drastic of steps is taken for Covid. But if your policies are unpopular, it will be a lot easier to lose political power as the timeline is ten times as long. It only takes slightly unpopular policies to be kicked out of power. So I think that aspect would help people who think we can do extreme measures for climate change (ie abandon most of the suburbs in a decade for denser urban cores more easily served by public transit) kind of temper their views a bit and think of more politically viable methods of fighting climate change (electric cars, etc). Of course it be nice to be able to modify the game to explain other arguments, like effects that redistribution might have making those changes more politically viable. But of course like most kinds of social science Model, it’s really more of an explanation/exposition than a prediction
RVuRnvbM2e 5y ago
> The game doesn't allow for zero covid strategy - because that was never realistically considered. The game should model this because it is the only strategy that has been shown to be actually effective, and might make people realise how badly governments handled this.
- edejong 5y ago
  I very strongly second this. The zero COVID strategy with localized lockdowns and a high emphasis on tracing would have prevented millions of deaths worldwide.
- noduerme 5y ago
  I think the point of the art is to show people how difficult it is to do the best you can with the tools available, and also to instruct them how to over-adjust with those tools and ease off them. To over-steer and then realign the direction of flight. What you're suggesting is: Give everyone an automatic transmission so no one ever shifts gears at the wrong time. What this game shows is: Here is how you shift gears before a curve, and what to expect if you shift too soon or too late. That is a fantastically more complicated and important thing to take away from a simulation than, "if we just turn everything on, problem solved."
  - littlestymaar 5y ago
    > What you're suggesting is: Give everyone an automatic transmission so no one ever shifts gears at the wrong time. […]"if we just turn everything on, problem solved." Not at all. A zero-covid strategy doesn't mean “Covid doesn't happen, nothing to do here”. It means locking down entire cities as soon as a few case of unknown origin are diagnosed, and doing so every time the situation occurs: it's not a one time medicine, you must be ready to to enact local lockdowns many times. Australia, Vietnam, China, South Korea, etc. saved hundred thousands of lives by taking extreme measures super early on. Locking down 7 million people when 9 cases have been found is an extremely tough political decision to make, but it has been immensely successful (both from the health and economy point of view). The western world has been completely oblivious to this strategy since the beginning of the pandemic, and not allowing this kind of strategy in a game is just going to reinforce this idea that there is no alternative to the deadly “flatten the curve” strategy.
pmontra 5y ago
Meta question to the author about the spread of information: you registered your account a few hours ago apparently to post answers here, how did you know that your game was on HN?
- mattowen_uk 5y ago
  I'm going to guess that the author checked where their site visitors were coming from.
- daeglin 5y ago
  Pointed out by a good friend :)
FabHK 5y ago
Awesome work. Astral Codex Ten (the reincarnated Slate Star Codex) has an informative article on the effectiveness of lockdowns that references and employs this simulation: https://astralcodexten.substack.com/p/lockdown-effectiveness-much-more https://astralcodexten.substack.com/p/lockdown-effectiveness... ETA: The post contains a graph of the frontier of pareto-efficient outcomes against which you might want to measure yourself when playing. See also this blogpost by one of the epidemiologists working on the game: https://translate.google.com/translate?sl=auto&tl=en&u=https://boundedlyrational.substack.com/p/o-dilematu-mezi-ivoty-a-ekonomikou https://translate.google.com/translate?sl=auto&tl=en&u=https...
pedrig 5y ago
Coolio! :) I built something similar [0] a while ago. My goal was to _visually_ demonstrate how fast a virus can spread. I wanted to make it is easy for everyone to get an understanding about exponential growth and how different measures can have a great impact on the spread of a virus. Disclaimer: This is an unpolished toy project. It was built in a weekend by simple dude who likes to code but has no idea about virology and epidemiology. [0]: https://patricsteiner.github.io/corona-simulator/ https://patricsteiner.github.io/corona-simulator/
daeglin 5y ago
[game author here] This forum is rate-limiting my posts so I am not able to respond to each particular thread. I am guessing that the renewed interest in our game is coming from the recent post at Astral Codex Ten: https://astralcodexten.substack.com/p/lockdown-effectiveness-much-more https://astralcodexten.substack.com/p/lockdown-effectiveness... It is well worth reading that post. Let me also link a post from one of our collaborators explaining how, for most countries, the dichotomy between saving lives and economy was somewhat false and relative to the real outcomes it was possible to improve both: https://translate.google.com/translate?sl=auto&tl=en&u=https://boundedlyrational.substack.com/p/o-dilematu-mezi-ivoty-a-ekonomikou https://translate.google.com/translate?sl=auto&tl=en&u=https... A lot of the discussion on this forum center around the question what was the optimal policy and how well the game simulated that policy. In particular, there is a lot of discussion around the two "extreme" approaches: "zero covid" and "no interventions". That is fine, we had these discussions many times and we can explain how these approaches are represented in the game (in my opinion, not too unrealistically) but, at the same time, it somewhat misses the point. The game was about building intuition about simple epidemiological SEIR model close to the regime where the real country was operating - and for political reasons it was very unlikely that the country will radically change the course. Lot of the game design decisions were some compromise between different goals of the game. Let me give an example: the team of authors of the game strongly believes that expanding contract tracing would a good thing. We were considering adding it to the game. However, what would be a point of adding that option to the game? It would further clutter UI (we had mobile users on mind). It would not add anything to the game because the decision to expand tracing would have no significant downside. And last it would hardly add anything to the public debate, it would just communicate that the authors of the game like contact tracing. There were already zillion of well argued opinion pieces on the topic. In fact, failures of contact tracing were more about failures of the specific agency than any conscious decision on the decision maker side. In short, tacking this issue in the game would quickly move from mathematical modeling into much more murky waters of public service efficiency. You should also keep in mind that the game was just a small piece in a bigger mosaic of truly incredible efforts of many other volunteers. E.g. "Interdisplinary group for epidemic situations" https://www.meses.cz/ https://www.meses.cz/. The game was never intended to provide comprehensive picture or recommendations.
daeglin 5y ago
[game author here] Note that I am Czech living in the US. Czech Republic and USA are the countries I follow most though I had discussions with other (mostly European expats). Note on the "social stability score": the game population is by design very "permissive" people will generally not revolt even if you implement very harsh lockdowns of if many people die. This is because we primarily wanted to allow the player to explore the epidemiological model even for fairly extreme strategies (and public opinion is hard to model anyways). In addition, we expect that most player will play the game only once and we didn't want to frustrate too many new players by "game over" screens. For context: here is how I see what was happening in Czechia. The spring 2020 was handled well and country avoided any significant surge in Covid cases. During summer 2020, people get used to the pandemic. New view got a significant traction: Covid is not that serious and the best course was to let it sweep through the population and gain natural herd immunity. Since then, there was a constant tug of war between side wishing to relax the rules and the side wishing for stricter rules. Every time the disease started to decline, the "relax rules" side gained more traction and every time the hospitals were close to the capacity (or arguably above the capacity), the government implemented fairly strict measures (there was a strong consensus that everybody must be able to get healthcare). This explains why the popular "zero covid" and "no intervention" strategies were never attempted. It is hard to tell how well our model works for these strategies. My guess is "not too poorly" but the we have much better confidence in the model in the regime much closer to real world behavior of the country.

daeglin 5y ago

BTW, source code is available here: https://github.com/swehq/corona-game https://github.com/swehq/corona-game

uzakov 5y ago
This is honestly great! Could you please add a license? I would love to translate this and deploy in other languages.
- daeglin 5y ago
  Please contact us at info@koronahra.cz . The game is "open source" but we didn't explicitly specify what exactly that means. I consider the owners of the code to be the company swehq.com that did most of the coding.

noduerme 5y ago

Wonderful, sparse, essential simulation. The stark choices and the resulting reminders of what really happened (and how the player matches up) is a great eye-opener for the general public. The fact that it forces you to make decisions in realtime and feel the lag behind them is really brilliant as a thing to put in front of decision-makers in parliament. One of the most difficult things with covid was the inability to see a few weeks into the future, because our minds are not equipped to think in terms of exponentially multiplying consequences. This gives a really elegant, intuitive form to a very conceptually difficult set of variables. The beauty of the concept is its simplicity, as well. It's a bit like an IFR training panel, or early text-based video games. You can't see where the plane is or where it's going, you only see the instrument readouts after you adjust something. Since we have all been flying blind, this is a great experience to put in front of people, simply of what it's like to adjust and wait. Five stars, sending it to all the devs I know, also as a way to make a very engaging sim on a complex topic for people with no prior art =D

daeglin 5y ago
Moon Lander was fairly explicitly inspiration for this game :-)

Robotbeat 5y ago

This is really cool. I think it'd be interesting to model decarbonization in a similar manner, with social stability and debt and deaths.

rob74 5y ago
This Covid-19 simulation, where you can have a more or less reliable model because you are basically doing it after the fact, already has enough potential for controversy - e.g. the "social stability" metric, is it too sensitive? Not sensitive enough? For decarbonization, where you don't have a reliable model, you will probably be accused by both sides that you're underestimating or overestimating the various parameters involved. Do you want to open this can of worms?
- zelphirkalt 5y ago
  Well, it would be a game or simulation after all. It could be improved over time with new knowledge. So it wouldn't hurt to try.
- Robotbeat 5y ago
  Well, we’ve got to! and it would be nice to open up the game to different assumptions. Like, I tried Covid zero in this game, and although I never got to zero it ended up minimizingz the deaths. Games like this can be useful for at least explaining ones ideas even if the games aren’t “realistic”. I think for decarbonization, you wouldn’t need to take as drastic of steps is taken for Covid. But if your policies are unpopular, it will be a lot easier to lose political power as the timeline is ten times as long. It only takes slightly unpopular policies to be kicked out of power. So I think that aspect would help people who think we can do extreme measures for climate change (ie abandon most of the suburbs in a decade for denser urban cores more easily served by public transit) kind of temper their views a bit and think of more politically viable methods of fighting climate change (electric cars, etc). Of course it be nice to be able to modify the game to explain other arguments, like effects that redistribution might have making those changes more politically viable. But of course like most kinds of social science Model, it’s really more of an explanation/exposition than a prediction

RVuRnvbM2e 5y ago

> The game doesn't allow for zero covid strategy - because that was never realistically considered. The game should model this because it is the only strategy that has been shown to be actually effective, and might make people realise how badly governments handled this.

edejong 5y ago
I very strongly second this. The zero COVID strategy with localized lockdowns and a high emphasis on tracing would have prevented millions of deaths worldwide.
noduerme 5y ago
I think the point of the art is to show people how difficult it is to do the best you can with the tools available, and also to instruct them how to over-adjust with those tools and ease off them. To over-steer and then realign the direction of flight. What you're suggesting is: Give everyone an automatic transmission so no one ever shifts gears at the wrong time. What this game shows is: Here is how you shift gears before a curve, and what to expect if you shift too soon or too late. That is a fantastically more complicated and important thing to take away from a simulation than, "if we just turn everything on, problem solved."
- littlestymaar 5y ago
  > What you're suggesting is: Give everyone an automatic transmission so no one ever shifts gears at the wrong time. […]"if we just turn everything on, problem solved." Not at all. A zero-covid strategy doesn't mean “Covid doesn't happen, nothing to do here”. It means locking down entire cities as soon as a few case of unknown origin are diagnosed, and doing so every time the situation occurs: it's not a one time medicine, you must be ready to to enact local lockdowns many times. Australia, Vietnam, China, South Korea, etc. saved hundred thousands of lives by taking extreme measures super early on. Locking down 7 million people when 9 cases have been found is an extremely tough political decision to make, but it has been immensely successful (both from the health and economy point of view). The western world has been completely oblivious to this strategy since the beginning of the pandemic, and not allowing this kind of strategy in a game is just going to reinforce this idea that there is no alternative to the deadly “flatten the curve” strategy.

pmontra 5y ago

Meta question to the author about the spread of information: you registered your account a few hours ago apparently to post answers here, how did you know that your game was on HN?

mattowen_uk 5y ago
I'm going to guess that the author checked where their site visitors were coming from.
daeglin 5y ago
Pointed out by a good friend :)

FabHK 5y ago

Awesome work. Astral Codex Ten (the reincarnated Slate Star Codex) has an informative article on the effectiveness of lockdowns that references and employs this simulation: https://astralcodexten.substack.com/p/lockdown-effectiveness-much-more https://astralcodexten.substack.com/p/lockdown-effectiveness... ETA: The post contains a graph of the frontier of pareto-efficient outcomes against which you might want to measure yourself when playing. See also this blogpost by one of the epidemiologists working on the game: https://translate.google.com/translate?sl=auto&tl=en&u=https://boundedlyrational.substack.com/p/o-dilematu-mezi-ivoty-a-ekonomikou https://translate.google.com/translate?sl=auto&tl=en&u=https...

pedrig 5y ago

Coolio! :) I built something similar [0] a while ago. My goal was to _visually_ demonstrate how fast a virus can spread. I wanted to make it is easy for everyone to get an understanding about exponential growth and how different measures can have a great impact on the spread of a virus. Disclaimer: This is an unpolished toy project. It was built in a weekend by simple dude who likes to code but has no idea about virology and epidemiology. [0]: https://patricsteiner.github.io/corona-simulator/ https://patricsteiner.github.io/corona-simulator/

daeglin 5y ago

[game author here] This forum is rate-limiting my posts so I am not able to respond to each particular thread. I am guessing that the renewed interest in our game is coming from the recent post at Astral Codex Ten: https://astralcodexten.substack.com/p/lockdown-effectiveness-much-more https://astralcodexten.substack.com/p/lockdown-effectiveness... It is well worth reading that post. Let me also link a post from one of our collaborators explaining how, for most countries, the dichotomy between saving lives and economy was somewhat false and relative to the real outcomes it was possible to improve both: https://translate.google.com/translate?sl=auto&tl=en&u=https://boundedlyrational.substack.com/p/o-dilematu-mezi-ivoty-a-ekonomikou https://translate.google.com/translate?sl=auto&tl=en&u=https... A lot of the discussion on this forum center around the question what was the optimal policy and how well the game simulated that policy. In particular, there is a lot of discussion around the two "extreme" approaches: "zero covid" and "no interventions". That is fine, we had these discussions many times and we can explain how these approaches are represented in the game (in my opinion, not too unrealistically) but, at the same time, it somewhat misses the point. The game was about building intuition about simple epidemiological SEIR model close to the regime where the real country was operating - and for political reasons it was very unlikely that the country will radically change the course. Lot of the game design decisions were some compromise between different goals of the game. Let me give an example: the team of authors of the game strongly believes that expanding contract tracing would a good thing. We were considering adding it to the game. However, what would be a point of adding that option to the game? It would further clutter UI (we had mobile users on mind). It would not add anything to the game because the decision to expand tracing would have no significant downside. And last it would hardly add anything to the public debate, it would just communicate that the authors of the game like contact tracing. There were already zillion of well argued opinion pieces on the topic. In fact, failures of contact tracing were more about failures of the specific agency than any conscious decision on the decision maker side. In short, tacking this issue in the game would quickly move from mathematical modeling into much more murky waters of public service efficiency. You should also keep in mind that the game was just a small piece in a bigger mosaic of truly incredible efforts of many other volunteers. E.g. "Interdisplinary group for epidemic situations" https://www.meses.cz/ https://www.meses.cz/. The game was never intended to provide comprehensive picture or recommendations.

daeglin 5y ago

[game author here] Note that I am Czech living in the US. Czech Republic and USA are the countries I follow most though I had discussions with other (mostly European expats). Note on the "social stability score": the game population is by design very "permissive" people will generally not revolt even if you implement very harsh lockdowns of if many people die. This is because we primarily wanted to allow the player to explore the epidemiological model even for fairly extreme strategies (and public opinion is hard to model anyways). In addition, we expect that most player will play the game only once and we didn't want to frustrate too many new players by "game over" screens. For context: here is how I see what was happening in Czechia. The spring 2020 was handled well and country avoided any significant surge in Covid cases. During summer 2020, people get used to the pandemic. New view got a significant traction: Covid is not that serious and the best course was to let it sweep through the population and gain natural herd immunity. Since then, there was a constant tug of war between side wishing to relax the rules and the side wishing for stricter rules. Every time the disease started to decline, the "relax rules" side gained more traction and every time the hospitals were close to the capacity (or arguably above the capacity), the government implemented fairly strict measures (there was a strong consensus that everybody must be able to get healthcare). This explains why the popular "zero covid" and "no intervention" strategies were never attempted. It is hard to tell how well our model works for these strategies. My guess is "not too poorly" but the we have much better confidence in the model in the regime much closer to real world behavior of the country.