3254: Detector

Detector

|< < Prev #3254 (June 3, 2026)

Title text: No other experiment has a lower false negative rate.

|< < Prev Random Explanation

Explanation

This is one of 44 incomplete explanations: This page was detected recently. Don't remove this notice too soon. If you can fix this issue, edit the page!

Ponytail and Hairy are showing Cueball a machine, claiming it is their most sensitive detector. Normally, detectors have a designated detecting job, such as smoke detectors which detect smoke. Being more sensitive means that it can detect (and perhaps quantify) far lower quantities/magnitudes of the target of its detections.

In this case, "more sensitive" seems to mean that it is sensitive to more different things. To quote Ponytail, it detects "gas, dust, particles, light, radio waves, gamma rays, protons, neutrons, electrons, fields, forces, events, potentials, or states", which runs almost the entire gamut of things that might be detectable, and leaves little room for there being any situation in which none of the aforementioned items are there to be detected. The constituent particles of the machine itself would be present for detection, and exist in "states" and have "potentials" relative to each other... assuming that the machine is sensitive enough to detect them.

Cueball points this out and Hairy says that it has been continuously lit this way since they turned it on. And Ponytail is left shocked when Cueball asks what would happen if the light labeled "Not detected" were to shine. Hairy claims that such a thing would be pretty bad, the presumption being that, if that were to happen, there would have to be no matter, light, forces, etc. within the detector's range of detection. (With almost no reason for the "Not detected" light to shine, they could have simply provided continuous power to the "Detected" light, but the reactions of the experts present show that neither of them think that this has been done.) Of course, a simpler explanation (and one far more likely to be correct) would be that the machine was just not working properly.

The title text states that the machine has the lowest false-negative rate out of any other machine as the "Detected" light will always continue shining. If it never ever states a negative, then it can never be wrong about it being negative, even though it is not clear what circumstances would result in a negative state being required, nor whether the detector will then (correctly) state that, rather than just continue to provide a (now) false-positive.

Randall had recently talked about detectors in 3249: Neutrino Project.

Transcript

[Hairy is standing to the right of a large machine labeled "Detector". The front of the machine has two lights, labeled "Detected" and "Not detected". The "detected" light is lit up in green. Ponytail and Cueball walk towards the machine from the left.]

Ponytail: Over there are our electron microscope, XRF scanner, and mass spectrometer.

Ponytail: And this is our most sensitive detector.

Cueball: What does it detect?

[The next panel zooms in on the detector. Ponytail's voice comes from the left of the panel.]

Ponytail (off-panel): Lots of stuff.

Ponytail (off-panel): Gas, dust, particles, light, radio waves, gamma rays, protons, neutrons, electrons, fields, forces, events, potentials, or states.

[The next panel zooms out. Cueball and Ponytail are standing to the left of the machine, and Hairy on the right.]

Cueball: I don't understand. Aren't most of those always present?

Hairy: Yeah, it's been saying "detected" continuously since we turned it on.

[Cueball and Ponytail still standing to the left of the machine, and Hairy on the right. Ponytail has her hand on her chin.]

Cueball: What happens if it says "not detected"?

Ponytail: Oh gosh.

Hairy: That would be pretty bad, I think.

  Add comment        Create topic (use sparingly)       Refresh 

Discussion

The deluxe edition of the machine probably has "Detected/Not detected" lights for each of those items. I would guess its cost would be significantly higher. 47.248.235.170 21:39, 3 June 2026 (UTC)Pat

To whoever wrote the initial transcript: the title text should not be included. Barmar (talk) 21:57, 3 June 2026 (UTC)

It doesn't even detect a vacuum? Useless thing, showing not even a vacuous truth. 2A02:590:1402:2E01:102F:245D:DD57:1937 22:15, 3 June 2026 (UTC)

It also has zero false positive Rtanenbaum (talk) 23:17, 3 June 2026 (UTC)

Not necessarily. We are told that it never says "No" when it should be saying "Yes", which suggests that there will be no false negatives. If were trust that assertion.

But we aren't given any assurance at all (even a simple nodding statement with no provability behind it) that the machine will not continue to say "Detected" even if it somehow really shouldn't. Should it do so, it would be be a false positive. Even though various other existential problems might be more important to anyone around who might care about it, at that point. 81.179.199.253 23:45, 3 June 2026 (UTC)

My point is that there is no circumstance that the machine would exist where all of the items it detects are no longer present i.e. the case of a false positive just cannot happen. Rtanenbaum (talk) 03:03, 4 June 2026 (UTC)

Failing to detect states would be a particularly worrying prospect for people in the USA. --2A10:D586:3E93:0:CC7B:253E:A0AA:2FA9 08:00, 4 June 2026 (UTC)

I would claim the assertion about false negatives is incorrect. It's impossible for a true negative to be displayed by the machine, so it's enough for a single malfunction that ever makes the machine display a negative to get a 100% false negative rate. 2A00:79E0:2820:8:9966:10CC:BAF3:62E 09:13, 4 June 2026 (UTC)

Not necessarily true. If the machine is set up to scan a volume of space that includes its own green lamp, then it should be recording the matter, light, etc of that green lamp, so no question that it would (self-reinforcingly) be detecting the thing that says it is detecting things, and never .

But if it's monitoring a distinct space a few feet away, and (because of handwavy occurances that I'll leave to your imagination) that whole area of space gets changed/replaced/displaced by some utter shielded void of some kind, that could result in there now being a true negative result to report, whether or not the machine can or will do.

Which is moot if there's no way for the Not Detected lamp to light (there's no way for the circuitry to light it... there is no circuitry attached to it... it doesn't even have a bulb... there's an ED209 standing nearby to aggressively shoot anyone who ever looks like they're trying to install the necessary bits to overcome any or all the deliberately omitted ways for the lamp to even apparently be lit up...) the instances of it being negative/"Not" might be ruled out by even the weirdest quantum effects (there just is no solution to the universe's wavefunction that results in the other light being lit), even in the apparent philosolhy of "if anything is possible, it must happen in an infinite universe".

What it comes down to is whether the possibility of the light ever lighting is (unlike the unspoken possibility of a bubble of non-existence spontaneously forming or passing over the right spot) sufficiently unlikely within the xkcd-verse to make it less unreliable (for negative reporting) than all the other detectors. Like how many times have neutrinos passed through the neutrino detector without being detected (very high false-negative; false-positives may also happen when someone cannonballs into the pool during the pool-party), but less so for more reliable experimental detectors. 82.132.239.222 12:26, 4 June 2026 (UTC)

If the Not Detected light were to turn on, it would be detected, and would thus turn off, providing opportunity for oscillation. But at what frequency? 173.188.194.118 12:56, 4 June 2026 (UTC)

Perhals (Planck time)^-1 if it was immediate (and no silly 'speed of light' delay within itself, or other processing delays like collating/multiplexing the various inputs for all the different sub-detectors prior to choosing which light to light) plus could detect an intensity of light as low as the NDL emits. (Imagine that it only triggers when a supernova-level of light is detected? It may be a very insensitive detector, by what degree of detectable-stuff it can measure, even though it is also sensitive to a wide variety' of different detectable-stuffs, at least once they get beyond some arbitrary individual thresholds for their respective quality of detectability.)

Though the Detected-light would already be detected (as light and mass and electrical fields and everything else), by the same assumption of its light-sensitivity (assuming the NDL isn't purposefully dimmer than the DL, exactly enough to prevent this NDL-detection), so if the lighting of the light is relevent to a Detection event, then Detection events are assured regardless of whether one starts from a Non-Detection Event baseline (immediately flips to being a Detection, then stays there). 82.132.239.222 15:15, 4 June 2026 (UTC)
        Add comment