2289: Scenario 4

Explain xkcd: It's 'cause you're dumb.
(Redirected from 2289)
Jump to: navigation, search
Scenario 4
Remember, models aren't for telling you facts, they're for exploring dynamics. This model apparently explores time travel.
Title text: Remember, models aren't for telling you facts, they're for exploring dynamics. This model apparently explores time travel.


Ambox notice.png This explanation may be incomplete or incorrect: Needs elaborating on the comic, and needs an explanation of the title text.
If you can address this issue, please edit the page! Thanks.

Although not directly mentioned, this comic is probably the 14th comic in a row (not counting the April Fools' comic) in a series of comics related to the 2020 pandemic of the coronavirus SARS-CoV-2, which causes COVID-19.

In 2278: Scientific Briefing, Megan and Cueball were briefing White Hat on things that were getting bad, hoping to convince him to do something about them. He chose to wait until things actually got bad. Evidently, that has happened, and now Megan and Cueball are delivering another briefing on just how much "Bad Stuff" there might be, according to their models.

In the context of the information (and misinformation) explosion associated with the COVID-19 pandemic (ongoing at the time that this comic was published), many graphs have been shown highlighting the prevalence of the disease - the number of cases at any one time and place, and the change in the number of cases over time. That being said, the graphs shown could easily apply to any number of scenarios where an upward trend is bad.

Several of these graphs have attempted to predict the future, using statistical tools ("models") to process existing data and generate a forecast. Inputs to the model(s) may include different assessments of, for example, the number of COVID-19 cases that have been recorded. Four scenarios are presented here, presumably showing what a particular model (probably only one despite the reference to "new modelS" in the comic) forecasts given different, unspecified, inputs.

Megan and Cueball present four scenarios, only three of which are possible.

  • The first, "best case" scenario recalls "flatten the curve" graphs that predict an occurrence will eventually cease to increase altogether. Using COVID-19 as an example, if strictest measures are put into place and adhered to, all those who have contracted COVID-19 will eventually be reported, and no further victims will contract it.
  • The second and third scenarios are increasingly worse cases, predicting that the occurrence will continue unceasingly. Again using COVID-19 as an example, the less measures are put into place or adhered to, the more COVID-19 cases that will occur. Scenario 3 appears to indicate an exponential increase best suited to a log scale; "pretty bad" is an understatement.
  • The fourth curve is not possible, as each point along the x-axis represents a specific time point. If the curve passes the same time point twice (as it does) then this means that on a given day there were two different number of cases. E.g. on the 1st of April there would have been both 100 and 1000 people infected, which makes no sense at all. The only way to make sense of it would be by using the common trope in science fiction of time traveling creating an alternate timeline in which events are different, thus the cases could be 100 in one timeline and 1000 in a different timeline. Hence the remark, "this model explores time travel", in the title text. This is a brain cramp to visualize, and the consequences of it actually happening would be calamitous on several levels. Real modelers might encounter such "graphing errors" while they are developing their models, entering data (especially if there are time-conversion errors), and testing their functions, but persons who went so far as to present such glitches in public, except for a laugh as here, would likely be asked to hand in their modeler's cards.
    • The 'time travel' remark is also suggestive of certain particle-physics phenomena captured in Feynman diagrams. Mathematically, an antiparticle moving forward in time looks like its equivalent particle moving backwards in time, so a particle-antiparticle annihilation or creation event could be interpreted as a single particle switching directions in time. In the context of this scenario, it is possible to read the fourth chart as predicting that the bad stuff will start traveling backwards in time as it increases, which we would see as a great quantity of "bad anti-stuff" appearing and decaying in number just as the "bad stuff" increases, until the two quantities meet at the halfway point and mutually annihilate. Even though there will be no more bad stuff after the annihilation (or time-reversal) event, particle-antiparticle annihilation releases enormous energies that might be even more catastrophic than whatever the bad stuff itself was.


[Megan and Cueball are standing in front of a large graph, with "Time" along the horizontal axis and "Bad Stuff" along the vertical axis. The curve on the graph shows a generally shallow upward slope.]
Megan: Our new models outline a few possible scenarios.
Cueball: #1 is the best scenario.
[The graph now shows a much steeper curve, before flattening out far in the future, similar to a logistic curve.]
Megan: Scenario 2 is not so great.
[The graph now climbs quite quickly, approximating an exponential curve.]
Cueball: Scenario 3 would be pretty bad.
[The graph starts curling up, like the exponential curve, but continues curving back, so that it no longer qualifies as a function, and may indicate time-travel to the past.]
Megan: Then there is scenario 4.
Megan: We think it's a graphing error.
Cueball: If not, we definitely want to avoid it.


  • This is the first Saturday comic since 2006.
    • It was released on Saturday since the previous comic was the April fools' comic for 2020.
      • The release of this comic, 2288: Collector's Edition, had been delayed two days because of technical difficulties with the complex interactive nature of the comic.
      • Thus the planned Wednesday release of this week was thus postponed to Friday.
        • To not cheat us from the planned Friday release, this comic was thus released the day after on Saturday.

comment.png add a comment! ⋅ comment.png add a topic (use sparingly)! ⋅ Icons-mini-action refresh blue.gif refresh comments!


Should definitely make a note re: this officially-Friday comic releasing late Saturday afternoon (EDT). TPS (talk) 22:06, 4 April 2020 (UTC)

   Or is this actually the april fool comic, except it fooled us by being on a Saturday? 22:12, 4 April 2020 (UTC) Sam
   Perhaps more likely because the actual April Fool's comic (due Wednesday) delayed 'til Friday. TPS (talk) 22:19, 4 April 2020 (UTC)

The title should probably be changed, the xkcd site uses the numeral "4" whereas we're using the word "four."--GoldNinja (talk) 22:50, 4 April 2020 (UTC)

   I accidentaly originally put it under Sequence Four. It shows in the image name.
       I hope I fixed it correctly Bugstomper (talk) 02:17, 5 April 2020 (UTC)

Another instance of a graph with poor labels ("bad stuff"), even without the time travel. 23:54, 4 April 2020 (UTC)

Scenario 1 is almost certainly intended to be a logistic curve. Scenario 2 starts off a bit slower, and it looks like one of the cases where you don't have enough data yet, but _hope_ it'll settle into a logistic. Scenario 3 is probably an exponential. These are the standard three scenarios: good, hopeful, and catastrophic. Then in his usual, now lets just get weird twist, comes the impossible one. I think that's all relevant, but I've been up too long to merge it into the text. MAP (talk) 23:30, 6 April 2020 (UTC)

The issue of time-traveling COVID-19 problems has already be considered in Onion Public Radio's The Topical. https://www.youtube.com/watch?v=E3GQwOcsChQ Apologies for any poor rule-following as this is my first edit. RandomEdditMemory (talk) 00:00, 5 April 2020 (UTC)RandomEditMemory

The time travel in this comic is probably a reference to the time offset resulting from the April Fool's comic, but possibly coincidentally the comic showed up here in New Zealand in the morning of the April 5 change to Standard Time when the clocks did turn back an hour. Bugstomper (talk) 01:55, 5 April 2020 (UTC)

The time travel is almost certainly not a deliberate reference to the April Fool's comic being late (or to any implementation of DST). Rather, there's 4 graphs, each with an increasingly higher curve. the first tapers off, appearing to be approaching an asymptote, with an ever-decreasing rate of increase -- or even heading to a decrease. The second has a steeper slope for a while, but then does start to taper off. whether it becomes linear, approaches an asymptote, or starts declining off the edge of the graph is not known. The third scenario appears to be an exponential curve. The 3.5 scenario (not shown) would be to have a vertical asymptote, where "bad stuff" shoots off toward infinity as time approaches T. Then the only other thing left to do with a curve is to have it continue back the way it came. Been too long since I was in that level of math, but I'm pretty sure it's problematic if you Y-axis has two values at a point on the X-axis. This isn't showing two different functions converging as time progresses, but rather that a high values of "bad stuff," time goes backward.

From the explanation: "This is another comic in the coronavirus series." But ... is it? It certainly isn't explicitly so. The implicit argument is easy to make, but the fact that it is just "bad stuff" as a function of "time," it could easily be relevant to any number of bad scenarios: velociraptor attacks, Macarena flash mobs, mobile game IAP monetization, nationalistic views in politics, cat-based cheeseburger memes, or so on. It's not much of a stretch to say that the comic is topical to current events (especially given that there are many others in a sequence of implicitly or explicitly CoViD-related comics), but it still is a stretch to _definitely_ say so absent Mr. Munroe actually acknowledging so elsewhere, and then a citation would be needed, right?

It definitely is. All those graphs (except the fourth, obviously) can be found in real countries' data. South Korea would be an example of scenario 1. The United States would be an example of scenario 3. The virus is on everybody's mind, so there's no way it's a coincidence. (I think labeling 2283: Exa-Exabyte as a coronavirus comic is way more of a stretch.)

I suggest the mathematical background of the graph 'bending over backwards' should be explained in more detail because the contradiction between what is 'natural' tendency of the graph and what is possible mathematically is what makes for the core of the joke. I mean, let's imagine that the graph is a picture of some tangible object, as a non-mathematically inclined person might do. Let's say, it's a rope. Then after observing it 'bend upwards' more and more with each scenario that gets progressively 'worse', it would be reasonable to conclude that continuing to bend this object even more and 'overbending it' would naturally mean some kind of a catastrophe. In reality, of course, it is impossible just because of the way the graph is being plotted. Each next segment is added as time goes by and placed more to the right because the time is shown to flow right on the horizontal axis. Thus the only way this graph could bend like this is for the next added segment to be in time 'before' the last one. And since it is impossible to travel back in time (citation needed), such a graph is unlikely to be predicting a real scenario. --SomethingLike (talk) 06:30, 5 April 2020 (UTC)

There are graphs, however, that have multiple Y-values for single X-values (graph of a square root function, at least for positive values of X, graphs of circles, or the batman equation. Might need an ELI5 why those are okay but the line curving back in time isn't.

You laugh here, but I have in fact seen graphs in corporate presentations which folded back. The presenter (a) didn't understand data analysis, (b) thought Excel was the right tool, and finally (c) decided the graph looked "better" by using the (incompetent) Excel pseudo-curve-smoothing graphics tool.Cellocgw (talk) 18:33, 6 April 2020 (UTC)

To me this comic seems a commentary on alocalypse. Some see COVID-19 as the start of a coming apocalypse, and some worst apocalypse scenarios involve either an explosive AI researching things like time travel or ending our timeline as physics knows it, or all of us going back to survival mode on a landscape without any modern infrastructure. 22:03, 5 April 2020 (UTC)

I think that "The only way to make sense of it would be by using the common trope in science fiction of time traveling creating an alternate timeline in which events are different, thus the cases could be 100 in one timeline and 1000 in a different timeline."..etc is utterly wrong creating a new timeline would have two 'forward' lines over a stretch of chart but would not have a single inflection joining a forward over ibto a backwards one. Maybe a (reverse) Z-bend if you include the retrograde (tachyonic?) leg, but then the true alternate timeline (also as per a single line splitting into two forward-going streams at a given t, whether or not that was invoked by time-travellers arriving at that point or 'mundane' quantum superpositioning if alternate outcomes) would not be backwards. (Alternative time-arrow, maybe, but that's more like a continuation of the existence of the usual one, which has no existence beyond the rotation of time into a backwards framing... However that happens - and this graph seems to indicate gradually, like the rate of time goes for +ve to -ve by having less seconds/'second' and passing zero, perhaps by somehow rotating in the imaginary time plane (similar, then, to a spacial one?) in which case there's probably more to worry about than the (presumably unrelated) Bad Stuff. (Darnit, forgot to sign...). 15:49, 6 April 2020 (UTC)

Consider for a moment that each graph is not one line segment, but two. That makes scenario 4 the best possible scenario. 10:51, 7 April 2020 (UTC)

Wouldn't an asymptote best represent a 'non time related' apocalypse? Vee00101010 (talk) 20:47, 8 April 2020 (UTC)

I have trouble seeing how it's useful to include the explanation of antiparticles as backward-in-time-traveling particles. It strikes me as only remotely related, and does not really explain anything about the comic. I think a simple explanation in terms of graphs and functions is appropriate here. -- Redbelly98 (talk) 02:07, 14 April 2020 (UTC)