Question about picking velocity & orientation

Hello dear friends,

I have been practicing usx for a while…Now i have 2 questions about right hand picking technique…

  1. Do you prefer to slide across the strings with small movements or play forcefully (possibly like gypsy jazz picking letting the adjacent string stop the pick) and destroy the strings?

  2. I realized that when practicing the famous chromatic picking exercise of petrucci usx picking feels easier when ascending and dsx when descending ( and i can pick faster) Is it a mistake to carry on picking like this (switching from usx to dsx when descending) and practice usx both ascending and descending? What do you think?

Thnaks in advance.

Others with far more articulation will hopefully chime in.

As controversial as some may find it…That chromatic exercise is a colossal waste of time. Don’t try to get good at it. For all intents and purposes…its pointless. And I did that exercise a LOT.

@Tom_Gilroy as done a massive amount of research and work on this. Before I’d even taken my first lesson with him, he saw a video of me playing that exercise and others like it. His comment was “At best they’re a waste of your time, most often they’ll reinforce fretting hand postures which are suboptimal for actual playing”

This post is full of great insight and worth the time:


Hi @pickfly . Sorry I couldn’t respond to this question sooner, I’ve been very ill this last week and I’m just starting to improve. I hope you can forgive any typos, etc.

I very strongly recommend that you aim to make large, powerful movements which change direction rapidly and which feel effortless.

The usual idea of economy of motion is that smaller movement are faster or more efficient is incredibly naive. It is totally incompatible with the physical and motor control realities of the situation.

In a physical sense, our picking hand can be (roughly) modelled as a forced oscillator. The equation which describes the amplitude of a forced oscillator is given below:


With terms
A = amplitude of oscillator
F_max = amplitude of driving force
w_d = frequency of the driving force
w = natural frequency of the oscillator
m = mass
b = damping

It’s important to understand that the frequency of the forced oscillator is equal to the frequency of the driving force w_d. In movement, this frequency is neurological in origin; it is the result of our ability to rapidly fire the neurons which trigger the contractions of our antagonistic muscle pairs. The other factor we have control over in this equation is the amplitude of the driving force F_max. The mass m and the natural frequency w are not adjustable parameters in our context.

Now, here’s the critical thing. The amplitude A is linearly proportional to the amplitude of the driving force F_max, but almost inversely proportional to the square of the frequency of the driving force w_d. What this means, is that with the same driving force, the amplitude A (movement size) will inevitably get smaller as the frequency increases.

If we begin with a movement which is small and lacking in force at a lower frequency (that is, tempo), the movement will simply be too small or too weak to be functional at a higher frequency (tempo).

The only way which we can deliberately make a movement smaller at a given frequency is to increase our damping coefficient b. This can only be done through muscular exertion and absolutely does not result in greater efficiency or frequency. Damping is to be avoided.

Small movements aren’t fast. Fast movements are small. The direction of implication that everybody teaches is completely backwards. It’s an overly naive reduction which doesn’t hold up to rigorous examination. It’s like saying that rich people buy Lamborghinis, so if I buy a Lamborghini I’ll be rich.

So, our goal is to make an oscillator which remains functional at our highest potential driving frequency w_d (which again, is neurological in origin). This means that we need to ensure that we have sufficient force production to drive the oscillator and that we have minimal damping. Essentially, if we’re practicing at a lower tempo, our movements should be large and powerful, but they should feel almost effortless.

This is the physical reality of periodic movement and it’s absolutely inarguable. I have heard some people try to find an “out” for the conventional ideas of economy of motion. Their argument being that focusing on making smaller movements is still (somehow) the best way to learn to play faster, despite the physics being backwards.

That’s an absolutely ridiculous claim based purely on the underlying physics. Beyond that, it’s completely incompatible with the realities of motor control. Specifically, the idea violates a principle in motor control called Fitts’ Law, which is the most well established and tested law in the field.

The idea is actually quite brilliant. Fitts used ideas from information theory to describe movement difficulty. His metric is ID, the index of difficulty, given by


Here, D is the distance to the target and W is the width of the target. Think of W as a tolerence which determines a “hit” or a “miss”.

Fitts also created a metric IP called the index of performance, which relates ID to the movement time MT and which is relatively constant in any task:


Then, we find a linear regression model:


So what does this all mean?

Well, it’s important to understand that whether we play faster or slower, our accuracy is subject to a target width W which is essentially constant.

Suppose we practice slowly with small movements. Now, D is small, so ID is small. Infact, it’s too small, because ID (index of difficulty) for this type of practice is far below a realistic ID for fast playing. Since MT is also large, the index of performance IP is also very small. As such practicing slowly with small movements does not in any way test that our index of performance is sufficient for fast playing.

With the understanding that IP is relatively constant, the only way we can make slower playing representative of faster playing is if the index of difficulty ID is greater for the slow playing than the fast playing. Thus, as we play faster and MT decreases, ID decreases proportionally and IP remains constant. Now, slow practice is representative of the actual difficulty of fast playing.

But how can we do this? Well, there’s really only one answer. Since the target width W cannot be adjusted, we can only increase ID by increasing the distance D to target.

So, when practicing slowly, we again want our movements to be large, powerful and easy.

The final element is that guitar playing requires not only spatial accuracy, but temporal accuracy also. For this, it has been demonstrated that faster movement speeds result in greater temporal accuracy. So, when practicing slowly (lower tempo), our hands should still be moving quickly. We should focus on making fast movements which accelerate and change direction rapidly.

Listen to this podcast for more information.

@Ruefus has already summarised my position on this exercise. Unless you love chromatic lines and want to optimize your playing around them, don’t bother.



I don’t want to hijack the post but I found the modelling of the picking hand as a forced oscillator quite interesting. @Tom_Gilroy When you say “It’s important to understand that the frequency of the forced oscillator is equal to the frequency of the driving force w_d.”, are you implying that the picking hand is always operating at the resonance peak? If that is the case, is the neurological system somehow finding what that natural frequency is and adapting the driving force frequency to that value?

No. Every forced oscillator oscillates at the frequency w_d of the driving force, not the natural frequency w of the oscillator. That’s universal, there’s no requirement to “tune” the natural resonant frequency w of the oscillator to to driving frequency w_d. We just won’t get peak resonance.

The natural frequency is not affected by the frequency of the driving force or by damping.

I think there may be some component of this, but I think it’s very unlikely to be the primary means by which picking mechanics are driven. The natural frequency is


Where k is the spring constant and m is the mass. We cannot practically alter the mass (unless you want to start cutting off your fingers), so tuning the natural frequency would come down to tuning the spring constant k. This could be feasibly done through increasing cocontraction of antagonist muscles, which might explain why some people have a tendency to tense up as they try to get faster.

The problem is that this would be quickly fatiguing, the muscles would have no recuperation time. There is always some slight amount of coocontraction required to remove muscular slack and allow rapid onset of movement, but in my opinion based on my own experience and my work with students, this degree of necessary cocontraction is so small that it’s better not to even think of it.

Another problem would be that we would be building a perceptual-motor control law which would be coupling time information (driving frequency w_d) to tension information (tuning cocontraction). As we fatigue, our perception of tension changes and our time would inevitably drift.

This is speculative, because I can’t cover an elite picker in electrodes and stick them in an MRI to measure their brain activity (@Troy if you want to chime in and tell me what’s up, you know I’d welcome it), but I would imagine primary means by which the picking mechanics are driven is through the simple push/pull cycle the antagonist muscles which drive the picking movement. The frequency of this cycle is the driving frequency w_d. It’s literally a direct coupling, and it would directly connect to our natural perception of rhythm also.

In any case, whether there is a component of resonance or not doesn’t change the big picture. If slow practice is to represent fast playing, it needs to be based on large, powerful movements which acclerate and change direction rapidly, but which feel easy.

Ok, I got you now. You just meant w_d is the frequency of the input force.

Yes, that’s it exactly.

In this case, the natural frequency is essentially zero, if the concept applies at all. A picking motion is not a harmonic oscillator because the neutral position for picking does not correspond to a force equilibrium. There is no spring constant to work with, and gravity acts simply as a DC force offset. Without a driving force, the pick would never return from a downstroke. On the upstroke, if the force is discontinued instantaneously after the pick passes the string, the rebound would be more of a ballistic problem then a harmonic oscillator, but setting that aside, the ‘resonant frequency’ of this system would be highly dependent on the amount of upstroke force, and in most cases be small enough compared to the driven frequency of high-speed picking to be neglected entirely.

This is not at all a criticisms of Tom’s forced oscillator analysis, which I liked very much.

1 Like

@pickfly just to give you a different perspective:

I did the Petrucci chromatic thing a good amount when I was “getting serious” about practicing, and at the time it was a good exercise for me because it got me used to playing with a metronome for multiple bars, across all the strings, and with all fingers. As the speed increased, I had to start implementing “chunking”, ensuring I was accenting the quarter note to not get lost. I don’t ever practice it anymore, but at the time it helped.

Tying this back to your question (#2), specifically USX for ascending and DSX for descending:

There’s no wrong way to do it, just keep in mind that the pattern is even notes per string (4), so it’ll feel different if you add in odd number groupings. I’d guess if you’re descending with DSX, you’re either swiping the last note into the first note of the subsequent string, or you’re developing some sort of helper motion. If it works for that particular chromatic exercise and you like the results, then no worries. I would try to play a few different lines, and see if the technique is carrying over to those as well. An easy modification / test is to do the same Petrucci chromatic pattern but in 3s.

For question #1, I prefer small movements.

Thank you Tom Gilroy. Actually i tried to apply this approach today in my practice session and it already feels like i am on the right track

1 Like