Why your shoelaces come untied

Your editor was interested in this because virtually 100% of his untied shoelaces are on his left foot. Haven't quite figured the following out, but it may help explain it. 

Royal Society

Hypothesized knot failure mechanism

We developed a hypothesis for knot failure based on the aforementioned high-speed observations and additional normal-speed observations of knots in both normal operation (i.e. on a shoe, walking or running) and artificial operation (i.e. swinging a shoe at the edge of a table in order to isolate inertial effects, or stamping of a foot in place to isolate impulsive effects). Our hypothesis was further refined by data from the testing of knots mounted on the device described in §4. The hypothesized mechanism is as follows:

• (i) as the leg is swung forward, and then slightly backwards to impact the ground, the loops and free ends of the shoelaces (figure 3b) are all pulled forward (with respect to the knot centre) by their inertia. The relative motion causes an opening of the knot—that is, a widening of the centre space separating the two trefoils. The centre space is where the free ends were pulled back through the knot to create the loops (cf. figure 14h,i in appendix A);
• (ii) the impact force of the shoe during the striking of the heel is transmitted to the knot by the tongue of the shoe and the eyelets. As a result, the centre of the knot deforms;
• (iii) the opening of the knot, and the concomitant reduction in friction forces, facilitates relative motion of the free ends with respect to the knot centre. In other words, with the centre of the knot pulled apart, relative axial motion between the free end and the knot centre becomes possible, and it is easier for the free end to slide out of the knot centre. The tension force pulling a free end through the centre is due to the imbalance of the inertia of the free end (the same inertia that helps to pull the knot apart), the inertia of the corresponding loop (to which the free end attaches through the knot), and the frictional forces at the centre of the knot;
• (iv) the repeated impacts perturb the knot such that the free end is incrementally pulled through the knot. This happens slowly at first. But as the knot is repeatedly pulled apart and more length is fed through the knot centre to the free end, the inertial forces pulling the free end through the knot are increasingly magnified, while the competing inertia forces of the loop diminish as the loop size decreases (figure 3c); and
• (v) eventually, the free end is sufficiently long that in one or two strides (impact cycles), the inertial force associated with the free end completely overpowers the loop's inertial force, causing run-away knot failure as the free end pulls completely through the knot centre. This last event signifies total failure of the second trefoil (figure 3d).