They found a general law (a square root relation) and tested that law on two types of materials. Now of course they can’t go around saying that material X will

They found a general law (a square root relation) and tested that law on two types of materials. Now of course they can’t go around saying that material X will have n cracks at speed v because material Y had m cracks at this speed. However, given a glass which produced n cracks at an unknown speed, they can take an – ideally – identical glass, hit it at a known speed, count the number of cracks and infer the previously-unknown speed. If the square root law is indeed as simple as I imagine (I didn’t read the paper), they will only need one data point to estimate the unknown speed, though more are obviously helpful. Over time, one could easily build up a database of materials, since there are only so many different cars, that shouldn’t be too difficult.