Train illusion

If you watch this for a bit, the train will appear to start going backwards. I’m sure there’s some fancy psychological name for this particular illusion, but I don’t know it. Thanks to several readers who sent this.

Oh, and today is the deadline for the cat beard contest. Get your cat beard photos in (if you’re brave enough to take one) by 5 p.m. Chicago time.



  1. Granny Sue
    Posted June 10, 2013 at 8:20 am | Permalink

    What does it mean if you can’t see the backwards part?

    • neil
      Posted June 10, 2013 at 8:23 am | Permalink

      You can make it go backwards or forwards at will by moving your eyes in the direction you want the train to go.

      • peltonrandy
        Posted June 10, 2013 at 8:26 am | Permalink

        I did this successfully. But I am at a loss as to which direction the train is actually going.

        • Notagod
          Posted June 10, 2013 at 12:13 pm | Permalink

          My guess is that the people are looking in the direction of the train coming toward them as opposed to in the direction of the train going away from them.

        • Jeremy Pereira
          Posted June 11, 2013 at 3:30 am | Permalink

          It’s coming towards the camera.

          1. the man at the back is looking towards the tunnel. He’s probably looking towards where the train is coming from.

          2. if the train was leaving the station, there would be nobody on the platform standing around – they’d be on the train.

          3. It’s Charing Cross station on the London Underground (you can see the station name reflected in the train’s windows). On the London tube, at the end of the platform where the front of the train is there’s always a TV monitor for the driver so he can see when the doors are clear. There is no monitor in this shot.

          4. If you’re on a tube train coming into Charing Cross (as I have been many times in my life), the platform is always on the right.

          • Jeremy Pereira
            Posted June 11, 2013 at 3:31 am | Permalink

            Forgot to click the subscribe button

      • Kevin Anthoney
        Posted June 10, 2013 at 9:31 am | Permalink

        If I look above the train it goes away from me, but if I look below the train it comes towards me. If I look at the train it goes in and out of the tunnel in a somewhat pornographic way.

        • Posted June 10, 2013 at 9:42 am | Permalink

          Stop watching porn and that effect may stop.

          • Notagod
            Posted June 10, 2013 at 12:06 pm | Permalink

            Maybe, but everyone knows that poor eyesight is the result of not watching enough porn.

    • John K.
      Posted June 10, 2013 at 9:48 am | Permalink

      I was able to make it “change” by changing the angle I was look at the screen. I.E. by leaning left or right.

  2. JBlilie
    Posted June 10, 2013 at 8:24 am | Permalink

    Photoshop …

  3. peltonrandy
    Posted June 10, 2013 at 8:24 am | Permalink

    I am able to see it going one direction and then with a slight change in the point of my focus I see it going in the other direction. But frankly, I can’t tell which is right. Is the train suppose to be coming toward me or away from me.

    • darrelle
      Posted June 10, 2013 at 9:26 am | Permalink

      Careful! The train will collapse as soon as you decide.

    • Gregory Kusnick
      Posted June 10, 2013 at 11:23 am | Permalink

      It’s coming out of the tunnel; see #18.

    • neil
      Posted June 10, 2013 at 12:25 pm | Permalink

      The fact that the people are standing facing the track suggests the train is coming in rather than leaving.

  4. Danny
    Posted June 10, 2013 at 8:28 am | Permalink

    It’s as your eyes get drawn further into the picture. further up the train the frame rate and the angle makes it look like the train is going back the other way. Look at the train in the foreground and it goes forwards again.

    • infiniteimprobabilit
      Posted June 10, 2013 at 11:48 pm | Permalink

      Well spotted. The effect is of course caused by the frame rate (or strobe effect), in the same way that wagon wheels in old Westerns used to rotate backwards.

      • infiniteimprobabilit
        Posted June 10, 2013 at 11:54 pm | Permalink

        … of course, the way to make it look as if one got the explanation in first is to post a ‘reply’ to an early message before all the other commenters have explained it in detail. As they did below. And which I didn’t read before I posted mine. D’oh…

  5. Dominic
    Posted June 10, 2013 at 8:30 am | Permalink

    It looks like it goes back in the front part but forward if i look at the nearest part. For me it flips like a Necker cube…

    • Mary Canada
      Posted June 10, 2013 at 10:44 am | Permalink

      Me too

  6. SMF
    Posted June 10, 2013 at 8:34 am | Permalink

    Is it a frame rate thing, like the “backwards wagon wheel” effect from old westerns?

    • Darth Dog
      Posted June 10, 2013 at 8:39 am | Permalink

      That’s what it looks like to me.

      • James M Peavler
        Posted June 10, 2013 at 3:50 pm | Permalink

        It is a timing thing. If an object in the picture has regularly timed segments (train, spokes on wheels, etc)and the length of time a frame takes in the recording device the to “float” in and out of synch — and hence appear to reverse directions from time to time. Happens in old movies.

    • Posted June 10, 2013 at 8:46 am | Permalink

      That’s also my guess: the apparent change in direction is caused by the train changing speed. I can’t tell whether it’s actually accelerating away from us, or decelerating towards us (I think the latter, ie. arriving in the station).

    • John K.
      Posted June 10, 2013 at 9:46 am | Permalink

      You got it. The Stroboscopic Effect.. Although, the TV thing had more to do with the strobe rate of of the TV display or TV camera. You got a similar thing with airplane or helicopter propellers.

      You could even recreate it with cake mixer blades in front of the screen, although I think modern day flat screens refresh too fast. You need an old school CRT.

      • Posted June 10, 2013 at 10:11 am | Permalink

        Oh, thank you. I’ve been wondering about this since 1967 and (as a kid) watched wagon wheels on those old cowboy westerns start to ‘rotate backwards.’

      • Posted June 10, 2013 at 10:14 am | Permalink

        lol. The wiki link even talks about wagon wheels in movies.

        I guess I should have read the link first. Now I look like I copied it…

      • Posted June 10, 2013 at 11:20 am | Permalink

        You can do it with guitar strings in front of a TV (pick your note appropriately!) so you can “see” the waveform of the string.

      • compuholio
        Posted June 10, 2013 at 11:20 am | Permalink

        Here is an even more impressive video where the blades of a helicopter are in almost perfect sync with the framerate.

    • Torbjörn Larsson, OM
      Posted June 10, 2013 at 12:13 pm | Permalink

      Yes. As the train gets up to speed, it somewhat but not quite matches the frame rate of the video. (The video can of course be manipulated to arrive at the effect, if it wasn’t a coincidence.)

  7. Nick
    Posted June 10, 2013 at 8:36 am | Permalink

    Hey Jerry, where do we submit the photo? Is there an email address or?

    • Kevin Anthoney
      Posted June 10, 2013 at 10:24 am | Permalink

      Follow the “Research Interests” link at the top for the email address.

    • Diane G.
      Posted June 10, 2013 at 10:25 am | Permalink

      Click on ‘Research Interests’ above to get his eddress.

    • Diane G.
      Posted June 10, 2013 at 10:27 am | Permalink

      Is anyone else getting incorrect time stamps on their posts? Don’t tell me I have to remember how to change my settings again…

      • Posted June 10, 2013 at 11:02 am | Permalink

        WP’s relationship with time of day is a mystery to me.

        Comment time stamps do not consistently relate to one certain time zone, and my WP app sometimes clicks over to a new day as early as five hours before midnight at my location, sometimes not early at all.

        IOW, no, I don’t think this is your problem.

        • Diane G.
          Posted June 10, 2013 at 11:28 am | Permalink

          Oh, good. It’s much better to know that I’m simply non-observant.

          • Dominic
            Posted June 11, 2013 at 2:38 am | Permalink

            Time is an illusion caused by a lack of alcohol.

            Oh dear, I don’t mean to encourage irresponsible drinking though!

            • Diane G.
              Posted June 11, 2013 at 9:30 am | Permalink

              No worries. I need no encouragement.

  8. Curt Cameron
    Posted June 10, 2013 at 8:41 am | Permalink

    If you look at the train on the right side of the image, it (to me) appears to be going away from me. But if I move my gaze to the part where it’s in the tunnel, then it suddenly appears to be coming towards me. So it depends on where I start looking, then the whole train goes that way.

    Did you notice the reflection in the train’s side, of the guy standing on the platform? His reflection looks like he’s dancing.

  9. lanceleuven
    Posted June 10, 2013 at 8:56 am | Permalink

    Whoah. That’s twisting my melon man. 🙂

    • SMF
      Posted June 10, 2013 at 11:07 am | Permalink

      Happy Monday BTW.

  10. Posted June 10, 2013 at 9:02 am | Permalink

    It is called the “waterfall illusion”, it is a motion aftereffect.

  11. Posted June 10, 2013 at 9:03 am | Permalink

    If you look straight at the train, it will probably appear to be coming out of the tunnel.

    If you look somewhere else, the train will probably appear the be going into the tunnel.

    I think that peripheral vision has a faster ‘frame rate’ than central vision*, and this difference affects how the movement is perceived.

    *This is probably a gross simplification.

  12. NewEnglandBob
    Posted June 10, 2013 at 9:21 am | Permalink

    I can will it to go forward or backwards.

    • Dave Hooke
      Posted June 10, 2013 at 10:39 am | Permalink

      The force is strong in this one.

  13. Kae Foggo
    Posted June 10, 2013 at 9:23 am | Permalink

    If you watch a tall waterfall for 10 seconds and look away to the side it to hill side or bank, the bank or hill appears to go up !!

    • Posted June 10, 2013 at 11:14 am | Permalink

      If you ride a motorcycle down a dirt track (with eyes focused much closer to the front wheel than they would be on a tarmac road) for 10 minutes, then stop, the entire world appears to be receding from you, for what I suppose must be exactly the same reason.

  14. Lyndon
    Posted June 10, 2013 at 9:39 am | Permalink

    I think we can call it a gestalt switch.

    It is essentially the same image that can be processed by the brain/mind in two different ways but not at the same time. Which shows that we are interpreting our environment (through sight, say) as much as we are directly reproducing it. Given our limited perspectives and visual capacities we use context and other tricks to build more complete representations of and “beliefs” about the world. But sometimes our representations and beliefs are unstable or illusory. They do not comport with reality or, say, are not a responsible representation of reality, for instance dangerous mirages.

    • Posted June 10, 2013 at 11:16 am | Permalink

      I was going to make a similar comment.

      Yes, our senses can deceive us, which is precisely why we need science: to mitigate our senses’ unreliability. We can interpret the gif in two different ways, because it’s really only two alternating still shots. But of course the train was in fact traveling in only one direction.

      Reality is really out there, despite pomo bluster.

      • Posted June 10, 2013 at 11:17 am | Permalink

        Or three still shots.

        • Gregory Kusnick
          Posted June 10, 2013 at 11:25 am | Permalink

          Four, in fact.

      • Andrew van der Merwe
        Posted June 10, 2013 at 11:44 am | Permalink

        I don’t go along with your reading. This is not a matter of being deceived by the senses but a matter of an ambiguous image. The fact that the train could only have been going in once direction, combined with the fact that the image is clearly seen to be ambiguous only serves to affirm the accuracy of the senses, not undermine them.

        As with all illusions of this sort, it trades on the selective removal of key visual cues which would otherwise enable a correct judgement about the scene it represents. This is easy to do with representations. The correct judgement about the image / representation of the scene is that it is ambiguous and anyone who calls it an illusion is correctly implying that it as ambiguous.

        What makes this illusion remarkable compared with others like the duck-rabbit illusion is that it involves movement. That is all. There is no deception of the senses only an ambiguous representation.

        • Lyndon
          Posted June 10, 2013 at 12:34 pm | Permalink

          Andrew, I agree with your assessment. If we study it closely and find some subtle clue that determines direction, we could then say that flipping to the wrong direction would be an illusion of the senses, but that probably gets dicey as well. What if the clue requires a more complex conceptual framework which then determines the direction, even though most immediate visual clues are ambiguous?

          But as it stands, for me, we take in all the relevant information of standard human senses and there is no determination of direction to be made.

          But Beef is of course right that science and scientific concepts play an important role in expanding our senses and the information that we gather from such.

        • Posted June 10, 2013 at 12:54 pm | Permalink

          Yes, this image is ambiguous, and yes, it is fairly easy in this case to determine that special care is needed to ascertain the reality of what’s going on.

          I can certainly imagine, however, simply glancing at this gif, not being careful, and concluding that the train was traveling into the tunnel rather than out of it.

          That it’s easy to catch this mistake and correct it in this case does not undo my point that this is a nice reminder that there are many, many more instances in which, despite careful observation, our senses will deceive us, and we will require the tools and methods of science to correct any mistakes we’ve made.

          • Gregory Kusnick
            Posted June 10, 2013 at 1:17 pm | Permalink

            Sorry, I’m with Andrew on this. Yes, our senses are fallible but that’s not what’s going on here.

            It’s like a bad digital recording of a pure tone from (say) a flute. If the sample rate on the recording is too low, it will introduce aliasing artifacts into the sound that will be audible on playback. Those audible artifacts are not the result of your senses deceiving you; they’re really there in the recording (as a Fourier transform will show).

            Similarly, what we’re seeing in this image is not real motion of a real train. It’s a digitally sampled recording of that motion, and the sample rate has introduced artifacts that cause the displayed image to exhibit a secondary mode of motion that was not present in the original. Such aliasing is a property of the recording, and varies with the parameters of the recording. Our senses are not to blame for this; they’re the victims of the illusion, not the cause of it.

            • Posted June 10, 2013 at 4:29 pm | Permalink

              I get your point, I really do, but I still think something like this can remind us that *because our senses are susceptible to illusions like this*, we need to enlist the aid of tools and methods that aren’t susceptible to them.

              • Posted June 10, 2013 at 4:31 pm | Permalink

                Oh, and for a moment I thought: “Cool! One of the commenters at WEIT I most respect agrees with me”.

                My name is also Andrew.


              • Andrew van der Merwe
                Posted June 10, 2013 at 5:25 pm | Permalink

                Gosh, I have to confess to feeling surprise more than flattery! I’m checking to see if there are any other Andrews here. But thanks, Andrew, the feeling is mutual.

                Re the deception of our senses, okay, I get what you’re trying to say, just that I’ve come to reject these so called “optical illusions” (they are not optical) as good examples of how our senses can be deceived and now see them as quite the contrary – evidence of how visually astute we can be. If we were not, we wouldn’t be calling them illusions, now would we?!

                But I’ve also taken a more subtle position regarding the use of scientific methods and instruments. These are only accurate where a numerical measurement is required and is otherwise often quite useless. For example, you could not easily tell me exactly, in miles per hour, for legal purposes, how fast that car was coming down the street, and one might conclude from that that the senses were not accurate. However, you could tell precisely enough to be able cross safely. At the same time, you could not cross safely based on a report of the car’s precise position and speed. Similarly, a good golfer makes no mathematical calculations but is able to strike with great precision, and I, when I make bread, do not need the use of a scale because I can tell if the water-to-flour ratio is right by the feel of the dough. The senses are enough.

                But here is the total best example (and I fear we will part ways on this one): the famous Checker Shadow Illusion ( I’m quite sure that you and 99% of the people here are sooner going to go by the colorimeter reading than literal common sense and agree that it is an illusion that the squares are different colours. It would be pretty unanimous here that the squares are actually the same colour because the science says so, but the science is quite obviously wrong.

                Or should I call the scientific measurement an incomplete observation? Have a look and tell me what you think.

              • Gregory Kusnick
                Posted June 10, 2013 at 6:13 pm | Permalink

                AvdM: I’m still with you on the Checker Shadow “illusion”. If you construct a real-world instance of it by setting a wine bottle on an actual checkerboard and carefully adjusting the lighting, then it becomes clear that it’s not an illusion at all. Your eye and brain are correctly discerning the actual colors of the squares, despite variations in lighting across the checkerboard.

                The only “illusion” in play here is the ability to interpret 2D patterns of pixels as 3D scenes with non-uniform lighting.

            • Launcher
              Posted June 10, 2013 at 4:52 pm | Permalink

              I agree. For a discrete-framed video like this one, aliasing would indeed be the main reason we perceive the train as going backwards. The sensory input, in this case, is “ambiguous” (as someone else put it) and is NOT a good test of the psychophysics of vision. **

              Nevertheless, there does exist an illusion for real-life (analog) viewing, in which a moving object appears to go backwards. In the article linked here, it’s called the “Continuous Wagon Wheel Illusion”:


              ** In human vision, discernible temporal frequencies go up to not much more than 60 frames/sec at reasonable contrasts. That’s appreciably higher than the train movie.

              • Andrew van der Merwe
                Posted June 10, 2013 at 5:51 pm | Permalink

                Yes, I’d call the Continuous Wagon Wheel Illusion a genuine proper, optical illusion because it has to do with the optics of the eye (frame rate and all that) but again, the fact that it is known to be an illusion is testimony to the astuteness of human observation. It is only when the wheel is abstracted from it’s context, when key visual cues are removed, that one can no longer tell which way it is actually rotating. And then usually you can tell that you can’t tell!

              • Gregory Kusnick
                Posted June 10, 2013 at 6:37 pm | Permalink

                “”…not much more than 60 frames/sec…”

                For me the limit seems to be around 72 Hz. In the bad old days of CRT monitors, it used to drive me nuts to use a screen with a refresh rate of 70 Hz or less. I always kept mine set to 75 Hz or higher, although my coworkers seemed unable to see a difference.

              • Launcher
                Posted June 12, 2013 at 3:47 pm | Permalink


                Actually, many visual illusions are attributed to processing at levels higher than the retina, and most of the others wouldn’t really be due to the “optics” of the eye (i.e. the lens,cornea, etc) but to the spatial and temporal resolution of the retinal cells that receive the optical signal. And many illusions happen during our everyday viewing experience without our even knowing (though they are most rigorously tested in controlled settings, mainly using calibrated monitors in a darkened laboratory; “out of context” as you put it). To a psychophysicist, illusions help reveal how our brains are wired to interpret visual (and other sensory) input; they truly push the limits of our perception, and not just for stimuli outside of real-world context.

                If you haven’t read it, I would encourage you to pick up Irvin Rock’s “The Perceptual World”. It’s a nifty little book about sensory processing and has many examples of visual illusions. The undergrad course I took that assigned that book (25 years ago!) was probably what influenced me most to pursue a research career in neuroscience.

              • Andrew van der Merwe
                Posted June 12, 2013 at 7:07 pm | Permalink

                Thanks, Launcher. I’ll look that one up. Interesting to think about the spatial and temporal resolution of the retinal cells too … And I’ll look that book up.

                I’d be interested to hear what you might have to say re some comments I’ve made over at “Uncle Eric once again goes after scientism and New Atheism, touting “other ways of knowing.” III. Scientism”. I used the checker shadow “illusion” as an example there too. It’s something I’m thinking a lot about but I didn’t really get any constructive replies, just a some raucous laughter which I had to put down. 🙂

              • Launcher
                Posted June 13, 2013 at 10:32 am | Permalink

                I’ll check out your comment there, Andrew. I think the checker board contrast illusion may very well be one of the examples in the Rock book. It’s also a good example of an “illusory” aspect of our vision that actually helps us perceive the world more astutely (to use your term from an above). 🙂

              • Launcher
                Posted June 13, 2013 at 11:24 am | Permalink

                OK, I looked over your comment (actually a whole wham bang discussion thread), and I do have a few comments of my own. First let me say that some of the posts there were a bit mean-spirited, but I admire your tenacity. I think a lot of the misunderstanding comes from what various people would call an “illusion”. Certainly the lame train video above is not an illusion by almost any standard, as the video frames were clearly altered and in any case are subject to aliasing effects. But the checker board IS an illusion in the narrow sense that the brain is not reporting the “true” pixel by pixel shades. The psychology world would, in any case, consider the checker board example an illusion.

                It’s true that a checker board in the real world would be perceived CORRECTLY as having a regular pattern of black and white squares, regardless of an impinging shadow. But it should be noted that an artist’s illustration or photograph of a checker board is exploiting that “local contrast” phenomenon to CONVEY the perception of a black and white grid by using an array of grey shading. While that’s exactly how our brain is wired to work, it is a true illusion because in fact there is no checker board only a flat 2-D display of something that represents a checker board. (So although that real-world square really is white, the 2-D image’s square is not – it’s grey.)

                You could fairly claim that the image is SUPPOSED to be a checker board, and I’d agree it’s not a really good example of an illusion in the sense that the intended object is not being masked in any way. That’s why, in a psychophysics laboratory (like Irvin Rock’s), in order to understand a visual phenomenon like local contrast mechanisms one uses much better controlled non-real-world images. Trust me that there are contrast illusions out there that really ARE meant to trick the viewer into seeing something unintended. One other aspect of a proper experiment is that there will be some randomization of, in this case, the grey luminance levels, such that there is no CORRECT answer to the question “are these the same colors or not?”, because the answer will be subjective and thus vary from person to person (especially in those suffering from a particular type of vision loss). (Consider for a moment real checker board that by design is NOT made of black and white squares, but has some shade variance in it.)

                I should mention two other examples you made in the other post. Something about a moving car, I think, and an athlete striking a fast-moving ball. Because of the way our brains are wired, there are actually some people (usually with a head trauma, tumor, etc) who can see a car exist at point A and point B separated by time X, but nevertheless not perceive that the car is moving. Not an illusion, here, but just an interesting case study that reveals something about how our brain works. But the fast-moving ball example MAY be related to a common visual illusion. Check out this physiology paper by Dan and colleagues:


                The idea is that certain properties of cortical neurons may allow a viewer to PREDICT the motion of objects moving very fast (faster, lets say, then the synaptic latencies in our brain’s visual system would otherwise allow). That same response property may be related to motion-induced illusions that cause an otherwise static image to appear to move (the so-called “Waterfall Illusion”). So, like the checker board/local contrast phenomenon, this type of neural mechanism helps us perceive objects in our regular-world environment but nevertheless does NOT relay a 100% accurate report of the true physical properties of that environment.

              • Andrew van der Merwe
                Posted June 23, 2013 at 3:10 am | Permalink

                Launcher, I’m so sorry to only reply now. You talk of tenacity, but in fact I did run out of steam and had work pressure mounting. Really I appreciate your taking the time and your thoughtful response! This has been something that has exercised my philosophy muscle for some time and it’s difficult to find constructive engagement. I don’t mind a bit of sharp sparring but some people can be so dogmatic!

                My first thought re your first comment about what people call an illusion, is that illusion is not an objective thing. Take a mirage in the desert. That’s a genuine case of optics and light trickery and probably the oldest example of an optical illusion – or not. If you’re speeding up the pace of your crawling in the hope of a drink, then you have a case of an illusion, but if you see a mirage, then there is no illusion, because you are not deluded. Simple as that. You have to be deceived for it to be an illusion. Once you know it for what it really is then, as an illusion, it’s history. The very act of calling it an illusion makes it history. But it may remain an illusion for others.

                In the case of the checker board “illusion”, what you say is false, but it’s subjective. It might be true for you, but it certainly is not the case for me, and it can be the case for you. My brain correctly reports (using your reductionist terminology with a degree of irk 🙂 BOTH the the logical colour of B and the ““true” pixel by pixel shades.” The former is a common-sense, slam-dunk and latter is easily proved by the accuracy with which I could mix my paints to paint the scene.

                I probably wouldn’t be as good at it as some artists. Here is Robert Genn in a recent newsletter talking about Spanish painter Joaquin Sorolla (1863-1923): “Painters, according to Sorolla, need to think of themselves as truthful cameras. They need to develop the ability to see colours as they actually are, without the problems of previous understanding or careless rendition. He advocated sitting quietly out of doors while looking carefully at various elements in the surroundings–and mentally translating their colours into pigment. Sorolla, as well as Sargent, Monet and other great colourists, reported there to be nothing magic about it. Nailing the right pigment is an acquired skill.”

                If you have a look at his paintings ( you might notice how effectively they convey the Magic Hour light. What is remarkable is that he mixed those colours in-situ. Do you realise what that means?! Not only was he able to see the colour as it was, at the time, as distinct from midday 5000K colour, but he was able to mix his colour correctly on his palette while bathed in that same biased light. I’m distilling a theory about how one is able to do that, if you’re interested.

                I’ll reply later to the other interesting thing you say (“it is a true illusion because in fact there is no checker board only a flat 2-D display of something that represents a checker board.”) I often get that reply and I don’t think it is adequately thought through.

  15. Posted June 10, 2013 at 9:42 am | Permalink

    I think this is a product of the train’s velocity and the frequency of the florescent light’s flickering.

    sean s.

    • Gordon Hill
      Posted June 10, 2013 at 5:59 pm | Permalink

      Precisely. You can see the same thing watching cars with mag wheels in the movies which run at either 24 or 30 FPS.

    • Jeremy Pereira
      Posted June 11, 2013 at 3:42 am | Permalink

      The lights flicker at 50Hz in the UK. The gif has only four frames and the frame rate is 10Hz. It’s the “strobing” of the gif, not the lights that is producing the effect.

  16. Posted June 10, 2013 at 10:09 am | Permalink

    Spoked wheels do the same thing when you watch them. First they seem to be rotating the correct way, then it stops, then they rotate backwards. All while the object is moving in the same direction.

    I noticed that when I was a kid and wondered why… AFter all these years, this train makes me wonder again…

    Sadly, I don’t even know the right question to ask beyond it’s got to be some sort of motion-sensitive optical illusion.

    • JohnnieCanuck
      Posted June 10, 2013 at 10:43 am | Permalink

      Look up stroboscopic effect.

      Briefly, if I show you a still image of a spoked wheel and subsequent ones taken after exactly one or more complete revolutions, it will appear to be motionless. Now if the relative speed of its revolution slows, you will see it earlier and earlier and it will appear to slowly move backwards.

  17. Andrew van der Merwe
    Posted June 10, 2013 at 10:09 am | Permalink

    I doesn’t quite do that for me. If I keep my focal point perfectly still it appears to jump in and out but if I allow my eyes to wander over the image it appears to recede or advance depending on which direction my eyes are travelling in. Sweep your eyes from left to right and the trian appears to come towards you, sweep your eyes from right to left and it appears to recede.

    As with all illusions of this sort – and I distinguish here between illusions and optical illusions – it trades in the the removal of certain visual cues which would normally be present to enable a conclusive judgement.

  18. krzysztof1
    Posted June 10, 2013 at 10:27 am | Permalink

    the backward effect is strongest near the mouth of the tunnel. For me, at least.

  19. Posted June 10, 2013 at 10:51 am | Permalink

    Blinking seems to change the train’s direction. There must be a fluctuation bias built into the photoreceptors depending on the excuse one gives for being late to work.

  20. Posted June 10, 2013 at 11:00 am | Permalink

    I got to thinking about “observed” effects such as this.

    Imagine a dark room with two doors. Person A enters door 1, and sees a “ring of light” in the middle of the room. Person B enters from the opposite side of the room, and not only sees Person A, observing, but also sees the same “ring of light”. Person A also sees Person B observing.

    What both observed was accurately reported, and the observation of each other gives veracity to their observations. However, the “ring of light” turns out to be a single pinpoint of light, rapidly describing a circle at the end of a slender (unseen) rotating arm.

    My point is, neither party saw the single point of light. They did not observe the true ingredients of the phenomena. Their mind working through their eyes, deceived them. All that the two can accurately report is a “ring of light”.

    All holy books are filled with stories with one observer reporting…not even two. Not one, not one “holy” story can be corroborated with real effects. Mere observation, as this thread so ably demonstrates, is fraught with uncontrolled variables.

    You’d think that with a crowd numbered as “multitudes”, at least ONE PERSON during the fabled “water into wine” episode would have thought, “This is amazing! I’m taking home some of this magic!!” Or even two. They could sell the stuff for a year’s worth of wine.

    But, fictional multitudes never produce anything tangible. But they typically fit right in with fictional events.

    • Andrew van der Merwe
      Posted June 10, 2013 at 11:24 am | Permalink

      I don’t follow your thought experiment about the ring of light. For A and B to have observed a ring of light is simply to have made an incomplete observation, not an incorrect one (closer inspection reveals how the ring of light is formed but it remains a ring of light). And since this is the case with all observations (one can add detail ad infinitum) it seems pointless.

      • Gregory Kusnick
        Posted June 10, 2013 at 11:30 am | Permalink

        In fact it’s easy to tell the difference between a stationary ring of light and a rotating point of light: just flick your gaze back and forth across it. The rotating point will describe a looping path on your retina (unless it’s spinning so fast you’d hear the whoosh of it).

        Hands-on science museums often have experiments like this with flickering or moving light sources.

        • Andrew van der Merwe
          Posted June 10, 2013 at 12:03 pm | Permalink

          But I’m sure it would be easy to set up a laser light that rotates silently and too fast for the eye to tell if it were a moving point or a stationery circle.

          It all depends on the intention of or purpose made of the circle. If it were meant to appear stationery, then the report would be true and complete if it appeared that way. If it were a test of observation skills and if A said it was stationary and B correctly said it was a fast moving point producing the appearance of a static circle (thanks to your advice on how to look for it) then both reports would be true but A’s report would be incomplete. Only if A said it was not a fast-moving point would his report be false.

          • infiniteimprobabilit
            Posted June 11, 2013 at 12:01 am | Permalink

            If you go to a rock concert with a laser light show the light beam appears to form fans or cones or other shapes at will. I assume the laser light beam must be reflected off a controllable mirror that moves extremely fast in some programmed pattern so that it appears like a curtain rather than a traversing beam.

        • Posted June 10, 2013 at 12:23 pm | Permalink

          Well, hands-on science museums (such as where my wife works) don’t spend a lot of time fine-tuning simple illusions such as this. Your previous exposure to the phenomenon clues you to the deception. I picked this one, simply because of the stroboscopic aspect of it, compared to say, this one:

          As to Andrew’s remark, an “incomplete observation” is what the human mind is all about. All observations are incomplete. We are bombarded with information, and the rate of discard (rendering observations “incomplete”) is downright phenomenal. I wish I had the number right at hand, but this is too much fooling around as it were, to go find that rate of incoming visual data versus the discard rate. And humans do discard the important stuff in spite of their most earnest efforts (e.g. missing the tiger-strips among the leaf shadows).

          My point is simply that the stories of the Bible, such as the offspring of Adam and Even are buried in Hebron, never seem to be corroborated by anything material. Only by demonstrably fallible (and, fictional) observation.

          • Andrew van der Merwe
            Posted June 10, 2013 at 4:38 pm | Permalink

            Well, an interesting question for me would to which extent observations made in the Bible, and generally that far back in time, are incomplete rather than false.

      • Posted June 10, 2013 at 1:00 pm | Permalink

        It’s not that the observation was incorrect, it’s that the conclusion based on the observation was incorrect.

        • Andrew van der Merwe
          Posted June 10, 2013 at 4:31 pm | Permalink

          Yes, I have been careful to distinguish between the observation and the conclusion or report, as I called it. But even then, it’s not necessarily the case that the report is incorrect. It depends on whether the circle is static for all intents and purposes (visually static), and B’s report is only incorrect if he denies that it is a fast-moving point. It’s like if you ask me how much I’ve had to eat and I say “enough” but the scientist says “543 grams”. My answer can only be incomplete if a weight was called for and only false 543g is not enough.

  21. Gregory Kusnick
    Posted June 10, 2013 at 11:03 am | Permalink

    I see the illusion, and can make it switch directions at will, but it’s not obvious to me which direction is meant to be “forward” or “backward”. I suspect different commenters may have different ideas about that depending on which side of the road they drive on.

    • Gregory Kusnick
      Posted June 10, 2013 at 11:21 am | Permalink

      On inspection of the individual frames, it appears the train is “really” coming out of the tunnel.

  22. SMF
    Posted June 10, 2013 at 11:05 am | Permalink

    Bear in mind that it’s a looped GIF of a train slowing down (or speeding up – doesn’t matter.) Therefore if the strobe/wagon wheel effect is in play, at the start of the loop, it’s going one way. Then it switches (due to strobe effect.) Then the loop resets, so it switches again.

    It’s short enough to fool you into thinking that you concentrating on it is what makes it appear to change direction.

    Or: Magic Jesus did it!

    • Gregory Kusnick
      Posted June 10, 2013 at 11:18 am | Permalink

      The loop is only four frames long. I can sustain the illusion of motion in either direction for much longer than that.

      • Andrew van der Merwe
        Posted June 10, 2013 at 11:27 am | Permalink

        Yep, I noticed the same thing. Have you noticed the difference it makes moving your eye this way or that across the image?

      • Posted June 10, 2013 at 11:31 am | Permalink

        Yeah – you might have a point there. When I look at it again I can make it switch much quicker than I could when I’d had less wine.

        (TL;DR: wine is magic)

        • John Scanlon, FCD
          Posted June 11, 2013 at 5:49 am | Permalink

          Another example of chemical effects on perception that’s rather neat:

          Close one eye and hold one end of a drinking straw (a skinny one, not the kind designed for thick shakes) or similar narrow tube up to the other, against a background such as patterned wallpaper; you should be able to see right through, as well as the entire inside of the tube as a ring occulting a large circle of the background.

          But if the pupil happens to be sufficiently dilated, you can simultaneously see both the inside and the outside of the tube, and the entire background (except where your fingertips get in the way). Then, think about how that image is being projected on the retina…

  23. Gary Allan
    Posted June 10, 2013 at 11:30 am | Permalink

    This seems almost certainly to be due to a strobe effect and needs no fancy psychological term. The frame rate for the video combines with the speed of the train (which is changing throughout as it accelerates) to produce what seems to be reverse motion, but isn’t

  24. Posted June 10, 2013 at 11:40 am | Permalink

    Anyone else notice which station it is?

    • Posted June 10, 2013 at 12:06 pm | Permalink

      Zoo Station?

    • Simon Hayward
      Posted June 10, 2013 at 12:21 pm | Permalink

      It says Charing Cross on the reflected sign, looks like the Bakerloo line, hence my comment below.

    • Jeremy Pereira
      Posted June 11, 2013 at 3:44 am | Permalink

      Charing Cross

  25. Simon Hayward
    Posted June 10, 2013 at 11:43 am | Permalink

    I watched that long enough that the front end of the train must have arrived at Waterloo by now

    • Posted June 10, 2013 at 11:51 am | Permalink

      by sunset?

      • Simon Hayward
        Posted June 10, 2013 at 12:24 pm | Permalink

        It’s June – sunset is pretty late 🙂

      • infiniteimprobabilit
        Posted June 11, 2013 at 12:03 am | Permalink


  26. JohnnieCanuck
    Posted June 10, 2013 at 11:48 am | Permalink

    Very nice. Thanks for pointing it out.

    I’d have expected it to be busier, for being so famous.

  27. JBlilie
    Posted June 10, 2013 at 11:50 am | Permalink

    What struck me was the completely FROZEN people on the platform. Looks faked-up to me.

    • Posted June 10, 2013 at 11:53 am | Permalink

      Not totally frozen. As someone pointed out above, that dude down the end must be a vampire (or something) – his reflection on the train is dancing a jig.

      • Gregory Kusnick
        Posted June 10, 2013 at 12:03 pm | Permalink

        Actually I wonder if that tells us something interesting (or maybe not that interesting) about how the clip was made.

        Suppose the jig-dancing guy was caught in mid-step in this clip. But having him twitch back and forth on the platform distracted from the train illusion. So somebody edited that out by copying and pasting the platform portion of the image from one frame to the other three frames, to keep the people motionless. But a trace of his actual motion remains in his reflection, which couldn’t easily be edited out.

    • Jeremy Pereira
      Posted June 11, 2013 at 3:47 am | Permalink

      There’s only four frames with a delay between each one of 0.1 seconds, so it’s only 0.3 seconds between the first and last frame. I would suggest anybody standing still would appear frozen.

  28. misstexaskitty
    Posted June 10, 2013 at 1:32 pm | Permalink


  29. Posted June 10, 2013 at 3:26 pm | Permalink

    For me it depends which eye I use, it changes direction depending on it.

  30. tombesson
    Posted June 10, 2013 at 4:12 pm | Permalink

    This reminds me of the old “flipping the box” trick where you drew the outline of a three dimensional box on the board and stared at it for awhile. When one eye got tired of staying focused on the box, it shifted the load to the other eye and the box appeared to ‘flip’. When I taught this optical illusion to my students, I always said it was magic and that the box would flip if they kept repeating “Flip, flip, flip”. Funny stuff, but totally explainable using what we know about how the brain works.

  31. Launcher
    Posted June 10, 2013 at 4:53 pm | Permalink

    As I commented about elsewhere above, check out this link on the “real” Continuous Wagon Wheel Effect:‎

  32. MrHolbyta
    Posted June 11, 2013 at 2:27 am | Permalink

    If you enjoyed this, check out Brain Games on the Nat’l Geo channel (and occasionally Fox). They deconstruct lots of quirks of the brain including optical illusions, peripheral vision, focus & distractions to generally leave me feeling befuddled & intrigued by the ways the brain has evolved to help us model the world through which we move.

  33. Posted June 11, 2013 at 3:19 am | Permalink

    I think the train was going into the tunnel because there’s a reflection of a red signal in the windows which it would have triggered as it passed.

    Also, in looking this up I found it’s a good idea to double check facts, because, not being able to see the reflected light at first, I couldn’t see any signals, and assumed they were at the other end of the platform. Which would mean the train was going the other way…

    After a search on Google, it looks like the Northern Line at Charing Cross. Note red signal!

    • Jeremy Pereira
      Posted June 11, 2013 at 3:54 am | Permalink

      It’s coming towards us. See my post above for four reasons why. The signal is reason number 5.

      If the signal was to stop the driver from going into the tunnel we can see, it would have to be visible from the platform. As it stands, the driver would not see it until he had gone past.

  34. Richard Wein
    Posted June 11, 2013 at 4:09 am | Permalink

    The clip contains just 4 frames, repeated. If you assume that only a single carriage is passing during that time (and the clip isn’t sampling multiple carriages) then it shows the carriage moving towards us. That is, each successive frame after the first shows the carriage closer than the one before.

    Our brains try to interpret this as a normal picture of a train moving at a fairly constant speed, with multiple carriages passing. But, if we interpret each loop as another carriage passing, the jump from frame 4 to frame 1 is much bigger than the jumps between the other successive frames. So the repeated sequence isn’t a very good fit to the the interpretation of a train coming towards us at a fairly constant speed. If it was, I suspect we would be much more strongly inclined to prefer this interpretation over the “going away” interpretation.

%d bloggers like this: