If you are bored

[Calisto]

This is probably possible to solve by using a couple of math equations. According to the page source, the "cars" move at rates of 9, -11, and 7 pixels per jump (setting left as the positive direction).

If each car is reduced to the size of one pixel, the position can be expressed by a sawtooth function.
p(t) = d ( (v / d)t - floor( (v / d) * t))
d = width of window (pixels)
v = car velocity (pixels/jump)
t = number of jumps (jumps)

I could be wrong though.

You're certainly on the right track. Keep going!

Awesome!

So I guess then that means that the individual position functions would be:
p1(t) = d ( 9t/d - floor(9t/d) )
p2(t) = d ( -11t/d - floor(-11t/d) )
p3(t) = d ( 7t/d - floor(7t/d) )

Measured d = 954 px and timed the period of each car to be 9, 8, and 12 seconds, meaning that the jump rate is about 10 jumps per second. Each holbs-mobile is roughly 50 pixels long.

Putting all of these numbers together in Matlab shows that (on my browser) the cars will align at around jump #478. This corresponds to roughly 47 seconds after the page loads:

This actually occurred 43 seconds after the page loaded, but the difference can be accounted for in the error in measuring the jump rate.

[Calisto]

Matlab script if you're as crazy as I am and want to calculate it for your own browser:

Code Select Expand

function cars(width, maxTime, v1, v2, v3, tolerance)

t = 0 : 1 : maxTime;
p1 = width * ( (v1 * t / width) - floor (v1 * t / width));
p2 = width * ( (v2 * t / width) - floor (v2 * t / width));
p3 = width * ( (v3 * t / width) - floor (v3 * t / width));

for i = 1 : maxTime;
    if (abs(p1(i) - p2(i)) < tolerance && abs(p2(i) - p3(i)) < tolerance && abs(p1(i) - p3(i)) < tolerance)
        disp (i);
    end
end

%plot(t, p1, t, p2, t, p3);

end

[Camaclue]

...
HOKAY THEN

[Shadow]

I'm extremely pleased at how this thread turned out

Bonus points - use error propagation formulas to work out if your estimate has a 4s tolerance built in.

[Camaclue]

nein