I'm a commercial roaster. I roast mostly the same coffees, from the same farms year after year. I sprinkle in a few surprises as I find them, but mostly, like a singer or band, I have to keep playing my hits. In other words, I know these coffees very well. Any changes to the green or the roast are very apparent.

Before I roasted commercially, I bought beans 1-5# at a time and was constantly experimenting. Even as my roasting style became more consistent, the coffee I was roasting was not. Roasting 5# of coffee, 1# at a time does not really allow for truly dialing in that coffee. And by the time I thought I had figured out one coffee, it was gone and I was on to the next. I was learning and I liked variety.

But when you roast for other people you lose that luxury. You need to be able to produce the same coffee, the same way, over and over again. This is where Scott's "steadily declining RoR" really shines. Aside from producing the best coffees I have ever tasted, once you get how to apply the process to your roaster, it does it over and over again.

My 5kg roaster (all roasters really) was really not designed to do this. I had to modify it heavily to get the data I needed and learn how to make the adjustments to achieve the RoR I wanted. It's is a very hands-on manual process, almost like playing a musical instrument. One mistake will affect the performance. And like a musical performance, the audience might not ever pick up on it, but I know I made a mistake.

Flicks are a mistake...always. They add an unnecessarily unpleasant roastiness to a coffee. Even if you like the taste of roasty coffee, there are better ways to get it than letting the RoR spike at the end of a roast.

Crashes are mistakes as well. They rob the coffee of sweetness.

Right now I'm playing with a different roast profile that extends the yellow phase of a roast. This means I have to turn the heat down prior to dry end and back up again mid-roast and it doesn't allow for a steadily declining RoR. So far the results have been meh, but I will try it a little more. Up to this point I have been trying to extend yellow without extending the overall roast time. I might have to change that and start doing some 13 minute roasts instead of my usual 10-11.

So to finally answer your question, yes, it really matters. Actually, everything really matters, but some things more than others.

The only people that have differing opinions on this are people that haven't tried it, or been able to achieve it the right way. Once you get your roaster to do a proper declining RoR roast, the flavor of the coffee just pops. It's an exponentially sharp spike in dynamic flavor and sweetness. If I make a "mistake" in a roast, and get a little plateau or a mini-flick, I notice the difference in taste immediately and it's not a good difference. The coffee is diminished. Not bad, and still very drinkable, but no longer exciting to me.

So far I haven't heard anyone that has achieved this process say they have found something they like better.

Crash and flick describes one particular kind of uncontrolled and unintended deviation from a roasting plan and avoiding that matters in the same way that one should always avoid uncontrolled and unintended deviations from the roasting plan. Stated that way it's almost tautologically uninteresting, but it's worth calling out for two reasons. The first is that if you're primarily looking at lighter roasts, getting this right might be the hardest part of really solidly learning to control your machine. The second is that there are places where you can no longer post a solid, well executed S curve without someone mis-reading that as a crash and flick and using the term to be smugly dismissive of the roast without adding anything of value to the conversation.

For the rest, I think it's important to put some context behind my position. I'm also a professional roaster, but my favorite part of the job is exploratory roasting, roasting plan design¸ and experiencing the tremendous diversity of flavors available in high quality coffees. Right now the lightest coffee I'm roasting measures 63.2 on the gourmet scale (I generally don't go out any lighter than 70) with an ending temperature of 401°F (below 390°F there's a particular raw coffee flavor that I personally always find objectionable in a supposedly roasted coffee which exists regardless of how you get to that end point) with time between start of yellow and the end of the roast at 5:57. The darkest coffee I'm roasting measures 32.3 on the gourmet scale (I'll go as dark as the mid-20s if I need to, but usually the darkest is close to 30) with an ending temperature of 461°F (it's extremely rare that I decide to end hotter than that and I can't get much hotter without resorting to unorthodox control techniques because my safety gas shut-off kicks in at 460) with 10:54 from start of yellow to the end of the roast.

I'm not the only roaster running that broad a range and there are places that go lighter and/or faster than I do and there are places that go darker and/or slower than I do, but I think it's fair to say that most roasters focus on a much narrower range of possibilities, to the point that one of the classes I've been teaching lately comes out of conversations I've had with those people where some are incredibly proud of their light roasts, but they see market demand for something darker and don't know how to make it taste good and other roasters who are incredibly proud of their dark roasts, but see market demand for something lighter and don't know how to make it taste good. Getting people to shed whatever artificial constraints that prevented them from getting good results outside of their comfort zone and replacing that with a more general framework for understanding the decisions that can be made in roasting plan design not only helps roasters expand their range, but also often gives them additional expressive capabilities within their old range. That's the major failing of so much roasting advice given today. It presents one of several potentially reasonable starting points as overly strict constraints and yes, if you're lousy at controlling your machine and haven't paid enough attention to how variations in roasting change the results in the cup, practicing within those constraints will, in the short term, make you a better roaster, but it doesn't allow for advancing to the next level in skills development and they don't help if what you're trying to achieve in your own roasts is incompatible with that guidance. Rao's rules in particular are where roasters go to stagnate.

Ultimately it comes down to this:

What do you want to achieve with a given coffee?
What did you do attempt to achieve that result?
What were the results obtained by that method?

If you didn't get the results you wanted, the follow up is:

What change is likely to correct an observed deficiency?
What are the results of trying that change?

In some cases, an always declining rate is going to be exactly what you want. In some cases, especially for darker roasts, an S curve is going to hit the hotter temperatures needed to develop the flavors you want without obliterating everything developed earlier in the roast. In some cases it might be something that doesn't conform to anybody's general guidance, but you don't start from the crazy looking plans. Those just fall out of specific changes introduced to achieve specific results in the cup.

The other thing that is important to keep in mind when considering the shape of the rate of change is that this is absolutely not universal across machines. If you ever find the need to match flavors across two different machines, this is generally achievable, but the only way you get the same shape of the rate of change is if your required calibration is a constant temperature offset. I won't say that never happens because I know someone who does a lot of that kind of work and I believe her when she says she's encountered that, but even between very similar machine designs the calibration is nearly always non-linear across the range of relevant temperatures. What looks constantly declining on my machine might involve a flat or increasing section on yours.