I've been focusing studies on this "true roll" issue for about a decade now and would like to take a crack at this question.
Executive summary for Southern Ivy League grads : "Basically, worrying about true roll and practicing special techniques to get more forward roll than a plain-jane stroke does not benefit the putting result nearly as much (if any at all) as almost all people assume and is not worth the trouble. Simple enough, right? But perhaps you don't believe me without reading further. My motto: "Either get it right and make sense or shut up and stop claiming to know."
While Paul Hurrion and TaylorMade and Guerin Rife and Norman Lindsay and others have offered opinions and even science data about the rolling, NO ONE has ever actually studied whether quicker forward rolling matters. The "natural assumption" that it matters is just that, an assumption. But a close look at the issue throws that assumption into grave doubt.
What actually happens is that ALL putts transition from some skid-roll combination to no skidding and only rolling whenever the initial ball velocity on the ground reduces to 5/7th its starting velocity. It's physics and no one can do anything to change this.
Croquet science for skid-roll changeover to only-roll is the same as billiards and putting (note the 5/7 divider on the left):
This point in every putt happens roughly 15-20% along the total putt distance, the variation depending mostly on launching the ball and the green speed. (This has been noted since Cochran and Stobbs' book Search for the Perfect Swing in 1968.) But after this point, all putts are equal for quality of roll regardless of the skid-roll early phase.
So ... ALL efforts to generate so-called true roll are in reality efforts to end this transition phase a little sooner than otherwise. Paul Hurrion says the goal should be to end the phase between 10-20% of the total putt length. Uh, folks, that will happen no matter what the golfer does, so this suggestion is rather pointless, especially since there is no suggestion about HOW to get that done. What you want to know is whether it matters if you reduce a 20% skid-roll phase transition to a 15% skid-roll transition, not counting in that percentage any airborne travel, and to decide this issue requires focusing on the distance the poorer stroke takes to "catch up" to the first state of the better stroke once each ball has landed after launch. This "difference distance" is at most 1% of the total putt length.
There are two factors in promoting quicker forward roll at impact: putter design, and the golfer's stroke technique that puts that design to use in a certain way the putter head moves thru the ball at impact. Except for Harold Swash (pre-set the hands about two inches forward and then hit up with a "tangential strike" across the back of the ball), Norman Lindsay (hit the ball high on the face with a lower putter head COG and a recessed COG and the least loft required to generate a "vertical gear effect"), Michael Bonneau and Aserta Putters (hit the ball with the COG above the ball equator), Stan Utley (hit down on the ball with a 5-6 degree lofted putter using a hands-ahead forward press at setup that delofts the putter 3-4 degrees), and a few others, no one seems to have a suggestion about the golfer's stroke technique! (Just teasing -- .) Paul Hurrion doesn't mention stroke technique for true roll at all and discusses putter design only without reaching any conclusions on that issue either, but everything he teaches about stroke technique is taken directly from Harold Swash, who taught him.
Not withstanding all the confusion about what makes forward roll happen sooner off impact, EVEN IF one or more of the above techiques make the ball forward roll sooner than usual, what DIFFERENCE does that make? None of these guys have a clue about this, but they all have the same "natural assumption" that what they teach matters. Uh, you better check that before you spend a lot of time "grooving a move" at the expense of something else you should be focusing on.
All the science that I have seen about what happens in the skid-roll phase tells me that the effort chasing "true roll" is not necessary compared to a plain-jane sort of impact. Once any launching of the ball into the air is over and the ball is in steady contact with the green, any backspin is taken out of the ball in very short order. I mean VERY short order. Any backspin gets stripped out of the putt in about 2 inches of grass contact, but the ball is in transition and will be skidding a little as the rolling picks up steam and takes over from the skidding. The effort to start the ball off with forward roll and no backspin makes about a two-inch difference and that's it, and this two inches takes about 1/100th of a second to cover for normal putt distances and starting speeds, AND this two inches occurs when the ball has a full head of steam and lots of momentum early in the putt so that any little bumps in the way on the surface that might matter between a good roll and a poorer roll stroke get flattened! Most usual putt lengths start off with the ball traveling at 100 inches per second or greater and the ball mass is roughly 100s of times greater than a grass blade or the dirt of an ant hill or even a raised spike mark that has not yet dried and baked to a brick-like solidity. If the junk is big enough to knock this fast moving (even if largely skidding) ball off line, the quality of the roll won't matter a bit. Imagining the surface irregularity that matters for a mostly skidding / slightly rolling ball but not for one with less skidding and more forward rolling over these two inches where the poor roll is catching up to equal the better roll is pretty elusive. The actual chances that this combination of just this sort of surface irregularity getting in the way in just this particular two-inch span of the putt are very remote.
All the science REALLY shows that what causes line or distance variance is not a little more skidding than otherwise avoidable but launching the ball off the ground and side spin that produces a wobbling roll. Unless the so-called "true roll" stroke technique gets rid of or reduces excessive launching of the ball (or rebounding into the air after hitting down on the ball) or eliminates or reduces side spin on the ball at the start, it's not really helping to hold the line or make distance control more consistent. The ground pretty much washes out skidding and there are not really any "Skid Monsters" out in the dark ruining putts for line or distance. Golfers can be equally accurate and consistent for line and distance and even more accurate and consistent with a poor stroke technique for roll (more skid or less roll) than someone who has spent lots of time "grooving the move" of a so-called "true-roll" technique. What hurts is launching the ball or putting side spin on the ball.
Does a rolling ball work better for line and distance than a skidding ball? That is not at all the issue, but that is what many people think is the issue. The issue is what reduction in the transition phase is realistic and does the span of the putt that differs between poor rolls and better rolls contain any monsters that will make the poor roll less consistent or accurate? Answer: two inches and no.
Companies that claim a huge reduction in the transition phase are getting funny with the accounting. You have to watch their definitions very closely. Yes! Putters claim the true roll starts in 3 inches versus other putters not starting true roll until 18 inches. But when you look at what they are actually reporting, they are saying that the Yes! putters get rid of backspin within 3 inches, while other putters take 18 inches before all skid goes away. That's an apples-oranges comparison.
The Hurrion study of just this comparison gives data measuring the state of the roll over the first 65 cm (25") of the putt, and this data shows the Yes! putters have more forward roll at the start than the Brand X putter. "On average the Brand X putters have almost 100% skid at the start of their putt and are still skidding by an average of 55% after 500mm. The C-Groove putt starts with an average of 80% skid which been reduced to 40% skid by 500mm. ... No evidence can be offered, at the moment, as to why extra rotation is obtained." The hard data shows the Yes! putter sends the ball airborne for the first frame (0.04 seconds) and then starts with 39 degrees of forward roll in the second frame (time 0.04 thru 0.08 seconds), while the Brand X putter sends the ball airborne the first frame (time 0 to 0.04 seconds) and then starts with 7 degrees of forward roll in frame 2 (time 0.04 to 0.08 seconds) and then MATCHES the Yes! putter's second frame by its third frame (time 0.08 thru 0.12 seconds) with 42 degrees forward roll. (P. Hurrion and R.D. Hurrion, An investigation into the effect of the roll of a golf ball using the C-Groove putter, World Scientific Congress of Golf IV, ch. 47, pp. 531-538 (St Andrews Scotland, 22-26 July 2002), PDF downloaded from Yes! Golf, accessed 6 Feb. 2009.) The DIFFERENCE is that the Brand X putter on a 20-foot putt takes about 2-3 inches after landing to match the rolling of the Yes! putter and after that the rolls are equal in quality. Hurrion doesn't notice this.
TaylorMade use a similar "funny definition" when claiming the TaylorMade ASGI makes a difference that matters. Here is the TaylorMade graphic illustrating their data:
What TaylorMade calls the "skid zone" actually shows balls in the air, so where is the "skid"? The first two putter photos show the ball bouncing in the "skid zone" with surface contact only very brief before the ball returns to airborne, and the TaylorMade putter's "skid zone" shows NO CONTACT with the surface at all, only airborne travel. Their definition of the forward rolling zone is really the point when backspin is gone and the ball is steadily on the ground and not in the air or bouncing. ALL putts get here sooner or later. HOW MUCH sooner or later is in the data: about two inches.
Guerin Rife at least correctly defines the "skid zone":
Notice how quickly the bad putter's backspin disappears once the ball lands (inch 9.5) and is in steady contact with the green (about 1/2 an inch right at inch 10 in the photo). The "bad" putter's ball rotation matches the GR "good" putter's starting contact by 11.5 inches at about 70 degrees forward rotation. The "good" Guerin Rife putter has the ball land with about 70 degrees of forward rotation (at inch 6 in the photo below). What matters is that the poorer roll takes only two inches to match the better roll.
Companies that report ONLY the state of backspin / forward spin right after impact are not even describing a difference between the putters, as this doesn't show what happens when the balls land or how long it takes a bad putter to match a better putter, let alone show whether any such difference matters at all to accuracy or consistency. Here is the Aserta data, showing immediate launch spin states:
Hurrion in his "independent" study for Gel putters (after he left Swash and went on the Gel payroll as their nominal "designer") does the same, reporting only the initial spin state. (See Quintic Biochemical [sic] Analysis, Gel Groove Equipment Co., PDF downloaded 7 Feb. 2009.) The DIFFERENCE over seven putters from best to worst is only 50 degrees of rotation, which is about 1/5th of a roll, and this DIFFERENCE evaporates in less than 3 inches once the balls land and engage the surface friction. Same deal, but the GEL report is basically worthless as science because it has no context or significance to a golfer and lacks the relevant data to allow any interpretation that matters. The truth is the spin state between poor and better putters doesn't matter a tinker's damn. Launching and side spin wobble matter. (So let's change golf instruction and putter design from HOW to get "true roll" to HOW to avoid excessive launching and side spin.)
I know this post appears overly complicated, but, heh, it ain't. All I'm saying is that REAL DATA and SOUND ANALYSIS rather convincingly proves that all the hoopla about "true roll" is based on unexamined assumptions that don't hold up -- the real difference you get for all the trouble and attention to generate a better roll than the usual stroke is not worth a hoot because the improvement only differs over 2-3 inches and 1 or 2 one-hundredths of a second when the ball is "high-balling it" down the freeway! There ain't no monsters in the dark after all.