Short feedback loops are great when the feedback is clear, but without a guiding framework it can be hard to know what to make of the feedback—especially in the short term. If things get uncomfortable, is that a bad thing or is that a counter-steering? If I’m getting results quickly, am I Goodharting to rapid yet unstable results? If I’m not getting results quickly is it because the problem is hard or just that I’m ineffective? Am I even wise enough to know what a good outcome looks like in this situation?
A physicist can take the mass of your car, power capability of the engine, and the height differential between the start and finish, and give you a lower bound for how quickly you can race up a mountain. This is useful because if the lower bound is 99% of your actual time, then you know that you have little to gain in “learning how to drive” and need to focus on improving the power to weight ratio of your car if you want better results. If the lower bound is 1%, then don’t bother upgrading the car because the answer is all in learning how to drive.
Even without the inputs to the equation the physicist can also judge your efficiency by watching your driving, and give criteria to optimize. If the accelerator isn’t pinned to the floor, first do that. If you’re revving past the power peak, upshift. If you’re under it, then downshift. Optimality, in this case, looks like keeping the engine as close as possible to the power peak at every moment when the tires aren’t at their limit of traction. If your engine is rarely near its power peak, you’re not going to approach this lower bound. If you’re up against the constraints, you’re doing all you can. What might this look like when applied to “psychological” problems? What are the constraints?
To explore this, let’s look at the maximally simple example of learning how to shoot a handgun effectively.
When learning to shoot a handgun there’s various knowledge and skills that required. Once you get past the basics of things like “how to hold the gun” and “how to align the sights”, there are still skills to be honed in order to do it well. You have to practice by feel so that you can keep the gun aligned and find your sights quickly. You have to practice to get a feel for what kind of sight picture is acceptable. You have to practice to get the feel of how to smoothly press the trigger so as to not pull the gun off target. And yet, even complete beginners can do a passable job at these things right at the start. The thing that produces really atrocious shooting is a failure of a different kind.
Shooting a handgun can be unpleasant. They are loud. Even with ear protection, they are still really loud. And startling. They kick back at you, and it can be uncomfortable. The hard part of shooting a handgun decently is in being okay with this and facing it without flinching. When you begin to press the trigger, your brain knows that this BANG is about to happen, and so if it interprets this “BANG” as a bad thing, it will change direction on you and reorient your focus towards damage mitigation. You push forward and down on the gun in order to brace for the recoil, and in doing so, you bring the gun far out of alignment. If the flinch is especially bad, you might even wince and close your eyes too, attempting to hide from the stimulus. This shift of focus from “smooth steady trigger pull while keeping the sights aligned and on target” to “Ahh!!! Protect myself!” at the last moment is the reason your shots fail to hit the target altogether—at the moment it counts, you stop trying to shoot straight and start trying to do things that preempt that. It’s a failure to even try.
So how do we fix a flinch? What is going on here, and how quickly can we expect to resolve this psychological issue?
Conventional wisdom is “not quickly”. For example, here’s Navy SEAL Chris Sajnog on the topic. He is pushing back against people saying that it can’t be done at all, and saying that it’s actually easy and should only take about a month. In the comments, two people report trying his thirty day challenge. One spent 5 weeks and was successful[1], and the other was not successful[2]. There was one more commenter describing their experience getting over a flinch (before watching the video), which took him 9 months of consistent effort[3]. Chris’s suggestion is quite “open loop” both in that he tells you to not test for live fire results for the full thirty days, and that while he tells you to “tell your body you’re a great shooter” he doesn’t mention listening for a response.
Let’s open the black box, and see if we can’t do better.
A fairly common way to address flinching in new shooters, and the way I was taught as a kid, is with use of “dummy” rounds randomly interspersed so that the shooter does not know whether or not the gun will go off.
It can sometimes be hard to tell if you’re flinching when the gun is going off every time. If the gun doesn’t go off when you expect it to, there’s nothing to mask your flinch. The only thing to look at is your flinch and you feel kinda silly for flinching for no reason. Busted. So if one out of every 5-8 rounds is a dummy round—and you don’t know which—then when you get to it you will still mostly expect the gun to go off, and if you’re flinching when you know the round is live, you’re almost certainly flinching here. This gives you a way to catch yourself failing, so you get to notice and have that “Ugh, I messed up” corrective nudge. It also keeps you on your toes knowing that you’re going to get caught out if you keep flinching, so that not-messing-up stays on your mind.
If the gun goes off when you don’t expect it to, you won’t flinch because you won’t have known to. If only one in every half dozen rounds is live, you’ll still know it might go off, but it’s much easier to refrain from flinching when you know it probably won’t go off anyway, and the last three times you flinched all that happened was that you felt “Doh! I flinched for no reason again”. By mixing in dummy rounds, you keep yourself in a state of uncertainty and unable to miss the results of your flinching, so it is a much more efficient method of learning to overcome one’s flinch than simply “shoot a bunch” and hope that you end up getting over it by intentionally putting yourself in a position where you might accidentally learn. When you mix in dummy rounds, you get a surprise every few rounds at most, and that surprise is where you learn something. Maybe it’s “I can fire a live round without flinching”, or maybe it’s “I’m flinching even on duds”, but you’re learning something. It keeps you from the repetitive monotony of failing over and over and over without surprise, without corrective feedback, and without change.
What if we were to try to maximize this learning rate? If only one in ten rounds is a surprise, you’re wasting most of your time. The first obvious thing to do maximize the expected information per round (“entropy”) by using an even mixture of live to dummy rounds so that every trigger pull brings a surprise—but even this is suboptimal. You aim, squeeze the trigger, and when the hammer drops you find out whether the round was live or not and whether you flinched or not. This means only part of the time are you updating on what you might be doing wrong, and the rest of the time you’re flying blind.
Consider this instead:
Instead of pulling the trigger only to find out later whether you flinched, pay attention now. Do everything slowly. Slowly enough that you notice the impulse to flinch before you succumb to it. Only increase pressure on the trigger when you know that you aren’t about to flinch. You won’t know when the gun will go off, but the point isn’t “Take so long that you won’t know when to flinch”. As you increase trigger pressure, you will go from very certain that that gun is not about to go off to less sure to heightened anticipation. You don’t flinch when you know it isn’t going to go off, you flinch when you think it will. Or might. By increasing trigger pressure only when you know you aren’t flinching, you quickly get to that point of heightened anticipation and you get to spend the vast majority of your time sitting there, knowing that the sear is near releasing, knowing that the BANG is near, and maintaining control of your focus on holding the sights on target, and only increasing trigger pressure when you can do so without losing sense of priorities.
When you’re on that edge, the gun is about to go off with an unknown-but-small increase in trigger pressure, and in that state of heightened anticipation, you have time to reflect. And instead of “Yes, actually, this is legitimately scary and deserves to be”, you tend to realize “Actually, it’s not that bad. I’ve been here before”. The nerves fade, you calm down a little, and as you do you increase trigger pressure to bring yourself right back to that same level of nerves while being closer to the point of firing.
What happens as you iterate this process is that you get to continually sit in the conflict between “It feels scary to shoot a gun” and “I’m okay with recoil. I want to hit the target”, and realign your focus as necessary, as your brain realizes which of these two is more important and more sensible to you. Eventually, you get to the point where the sear releases, the bullet flies out the end, and you repeat the the process. As you spend more and more time doing this, you accumulate more time realizing that the recoil and report ain’t that big a deal, and the nerves become less and less prominent. Each time it goes a bit quicker, and before long it’s not necessary to slow down at all in order to maintain focus where you intend to keep it.
This is the process of reprogramming your own mind at optimal speed and precision. It involves relentlessly directing your attention towards the experience of realizing your goals, and the moment this becomes difficult to achieve, stopping to notice what your revealed uncertainty is and what you want to do about it. Aside from a quick trigger reset, the entire time is spent in this “figuring out what the right move is” mode, and reconditioning your response as your best bet becomes apparent. Known-failure is never accepted, and you don’t flinch even on your first shot. This isn’t just “theoretical”. It actually works, both on those who have been struggling with a flinch and those who have never fired a gun before in their lives.
Efficiently climbing hills means choosing a gear that keeps the engine near the power peak and spending the whole time full throttle, with a few moments here and there spent shifting. Efficiently changing minds means choosing a rate and type of approach that keeps focus where it needs to be so that the both sides of the relevant question are at the forefront, and spending the whole time there weighing the options—with a few moments here and there spent resetting the trigger and deciding on new targets to address.
In practice, this often happens so rapidly that we can miss it. In the cases where I’ve explicitly taught someone how to not flinch while shooting firearms, there was no prolonged “I’m overcoming a flinch” phase, just a smooth steady trigger pull and no flinch—from shot one. Shooting a gun just isn’t that scary, and so it generally doesn’t take but a few moments of properly aimed attention to address. Anything lasting much longer than that is the result of failing to hold focus on the unresolved question, rather than the result of having a big hill to climb.
Not all conflicts are so simple, but the metric for optimality is the same. Is near 100% of the time spent with attention on that relevant uncertainty that separates one side of the disagreement from the other, with only a moment here and there when shifting between sub-disagreements? Or are words and actions thrown around without expectation of anything surprising happening, that someone can learn from?
“Tried it for five weeks. Every day. Maybe 10 maybe 15 times. Found more than that lessened my concentration. Went to the range yesterday. 45 acp 9 out of 10 in the black (3″ @ 30′). 9mm 10 for 10. Will continue dry firing every day. Bragging absolutely not more like pleasantly surprised. Thank You Chris”
“Excellent suggesrions Chris! I worked hard about 14 years ago to overcome the flinch. I did LOTS of dry fire practice with some .22 rimfire practice sprinkled in. It took me something like 9 months of consistenr effort to work through it.”
Recognizing Optimality
Short feedback loops are great when the feedback is clear, but without a guiding framework it can be hard to know what to make of the feedback—especially in the short term. If things get uncomfortable, is that a bad thing or is that a counter-steering? If I’m getting results quickly, am I Goodharting to rapid yet unstable results? If I’m not getting results quickly is it because the problem is hard or just that I’m ineffective? Am I even wise enough to know what a good outcome looks like in this situation?
A physicist can take the mass of your car, power capability of the engine, and the height differential between the start and finish, and give you a lower bound for how quickly you can race up a mountain. This is useful because if the lower bound is 99% of your actual time, then you know that you have little to gain in “learning how to drive” and need to focus on improving the power to weight ratio of your car if you want better results. If the lower bound is 1%, then don’t bother upgrading the car because the answer is all in learning how to drive.
Even without the inputs to the equation the physicist can also judge your efficiency by watching your driving, and give criteria to optimize. If the accelerator isn’t pinned to the floor, first do that. If you’re revving past the power peak, upshift. If you’re under it, then downshift. Optimality, in this case, looks like keeping the engine as close as possible to the power peak at every moment when the tires aren’t at their limit of traction. If your engine is rarely near its power peak, you’re not going to approach this lower bound. If you’re up against the constraints, you’re doing all you can. What might this look like when applied to “psychological” problems? What are the constraints?
To explore this, let’s look at the maximally simple example of learning how to shoot a handgun effectively.
When learning to shoot a handgun there’s various knowledge and skills that required. Once you get past the basics of things like “how to hold the gun” and “how to align the sights”, there are still skills to be honed in order to do it well. You have to practice by feel so that you can keep the gun aligned and find your sights quickly. You have to practice to get a feel for what kind of sight picture is acceptable. You have to practice to get the feel of how to smoothly press the trigger so as to not pull the gun off target. And yet, even complete beginners can do a passable job at these things right at the start. The thing that produces really atrocious shooting is a failure of a different kind.
Shooting a handgun can be unpleasant. They are loud. Even with ear protection, they are still really loud. And startling. They kick back at you, and it can be uncomfortable. The hard part of shooting a handgun decently is in being okay with this and facing it without flinching. When you begin to press the trigger, your brain knows that this BANG is about to happen, and so if it interprets this “BANG” as a bad thing, it will change direction on you and reorient your focus towards damage mitigation. You push forward and down on the gun in order to brace for the recoil, and in doing so, you bring the gun far out of alignment. If the flinch is especially bad, you might even wince and close your eyes too, attempting to hide from the stimulus. This shift of focus from “smooth steady trigger pull while keeping the sights aligned and on target” to “Ahh!!! Protect myself!” at the last moment is the reason your shots fail to hit the target altogether—at the moment it counts, you stop trying to shoot straight and start trying to do things that preempt that. It’s a failure to even try.
So how do we fix a flinch? What is going on here, and how quickly can we expect to resolve this psychological issue?
Conventional wisdom is “not quickly”. For example, here’s Navy SEAL Chris Sajnog on the topic. He is pushing back against people saying that it can’t be done at all, and saying that it’s actually easy and should only take about a month. In the comments, two people report trying his thirty day challenge. One spent 5 weeks and was successful[1], and the other was not successful[2]. There was one more commenter describing their experience getting over a flinch (before watching the video), which took him 9 months of consistent effort[3]. Chris’s suggestion is quite “open loop” both in that he tells you to not test for live fire results for the full thirty days, and that while he tells you to “tell your body you’re a great shooter” he doesn’t mention listening for a response.
Let’s open the black box, and see if we can’t do better.
A fairly common way to address flinching in new shooters, and the way I was taught as a kid, is with use of “dummy” rounds randomly interspersed so that the shooter does not know whether or not the gun will go off.
It can sometimes be hard to tell if you’re flinching when the gun is going off every time. If the gun doesn’t go off when you expect it to, there’s nothing to mask your flinch. The only thing to look at is your flinch and you feel kinda silly for flinching for no reason. Busted. So if one out of every 5-8 rounds is a dummy round—and you don’t know which—then when you get to it you will still mostly expect the gun to go off, and if you’re flinching when you know the round is live, you’re almost certainly flinching here. This gives you a way to catch yourself failing, so you get to notice and have that “Ugh, I messed up” corrective nudge. It also keeps you on your toes knowing that you’re going to get caught out if you keep flinching, so that not-messing-up stays on your mind.
If the gun goes off when you don’t expect it to, you won’t flinch because you won’t have known to. If only one in every half dozen rounds is live, you’ll still know it might go off, but it’s much easier to refrain from flinching when you know it probably won’t go off anyway, and the last three times you flinched all that happened was that you felt “Doh! I flinched for no reason again”. By mixing in dummy rounds, you keep yourself in a state of uncertainty and unable to miss the results of your flinching, so it is a much more efficient method of learning to overcome one’s flinch than simply “shoot a bunch” and hope that you end up getting over it by intentionally putting yourself in a position where you might accidentally learn. When you mix in dummy rounds, you get a surprise every few rounds at most, and that surprise is where you learn something. Maybe it’s “I can fire a live round without flinching”, or maybe it’s “I’m flinching even on duds”, but you’re learning something. It keeps you from the repetitive monotony of failing over and over and over without surprise, without corrective feedback, and without change.
What if we were to try to maximize this learning rate? If only one in ten rounds is a surprise, you’re wasting most of your time. The first obvious thing to do maximize the expected information per round (“entropy”) by using an even mixture of live to dummy rounds so that every trigger pull brings a surprise—but even this is suboptimal. You aim, squeeze the trigger, and when the hammer drops you find out whether the round was live or not and whether you flinched or not. This means only part of the time are you updating on what you might be doing wrong, and the rest of the time you’re flying blind.
Consider this instead:
Instead of pulling the trigger only to find out later whether you flinched, pay attention now. Do everything slowly. Slowly enough that you notice the impulse to flinch before you succumb to it. Only increase pressure on the trigger when you know that you aren’t about to flinch. You won’t know when the gun will go off, but the point isn’t “Take so long that you won’t know when to flinch”. As you increase trigger pressure, you will go from very certain that that gun is not about to go off to less sure to heightened anticipation. You don’t flinch when you know it isn’t going to go off, you flinch when you think it will. Or might. By increasing trigger pressure only when you know you aren’t flinching, you quickly get to that point of heightened anticipation and you get to spend the vast majority of your time sitting there, knowing that the sear is near releasing, knowing that the BANG is near, and maintaining control of your focus on holding the sights on target, and only increasing trigger pressure when you can do so without losing sense of priorities.
When you’re on that edge, the gun is about to go off with an unknown-but-small increase in trigger pressure, and in that state of heightened anticipation, you have time to reflect. And instead of “Yes, actually, this is legitimately scary and deserves to be”, you tend to realize “Actually, it’s not that bad. I’ve been here before”. The nerves fade, you calm down a little, and as you do you increase trigger pressure to bring yourself right back to that same level of nerves while being closer to the point of firing.
What happens as you iterate this process is that you get to continually sit in the conflict between “It feels scary to shoot a gun” and “I’m okay with recoil. I want to hit the target”, and realign your focus as necessary, as your brain realizes which of these two is more important and more sensible to you. Eventually, you get to the point where the sear releases, the bullet flies out the end, and you repeat the the process. As you spend more and more time doing this, you accumulate more time realizing that the recoil and report ain’t that big a deal, and the nerves become less and less prominent. Each time it goes a bit quicker, and before long it’s not necessary to slow down at all in order to maintain focus where you intend to keep it.
This is the process of reprogramming your own mind at optimal speed and precision. It involves relentlessly directing your attention towards the experience of realizing your goals, and the moment this becomes difficult to achieve, stopping to notice what your revealed uncertainty is and what you want to do about it. Aside from a quick trigger reset, the entire time is spent in this “figuring out what the right move is” mode, and reconditioning your response as your best bet becomes apparent. Known-failure is never accepted, and you don’t flinch even on your first shot. This isn’t just “theoretical”. It actually works, both on those who have been struggling with a flinch and those who have never fired a gun before in their lives.
Efficiently climbing hills means choosing a gear that keeps the engine near the power peak and spending the whole time full throttle, with a few moments here and there spent shifting. Efficiently changing minds means choosing a rate and type of approach that keeps focus where it needs to be so that the both sides of the relevant question are at the forefront, and spending the whole time there weighing the options—with a few moments here and there spent resetting the trigger and deciding on new targets to address.
In practice, this often happens so rapidly that we can miss it. In the cases where I’ve explicitly taught someone how to not flinch while shooting firearms, there was no prolonged “I’m overcoming a flinch” phase, just a smooth steady trigger pull and no flinch—from shot one. Shooting a gun just isn’t that scary, and so it generally doesn’t take but a few moments of properly aimed attention to address. Anything lasting much longer than that is the result of failing to hold focus on the unresolved question, rather than the result of having a big hill to climb.
Not all conflicts are so simple, but the metric for optimality is the same. Is near 100% of the time spent with attention on that relevant uncertainty that separates one side of the disagreement from the other, with only a moment here and there when shifting between sub-disagreements? Or are words and actions thrown around without expectation of anything surprising happening, that someone can learn from?
“Tried it for five weeks. Every day. Maybe 10 maybe 15 times. Found more than that lessened my concentration. Went to the range yesterday. 45 acp 9 out of 10 in the black (3″ @ 30′). 9mm 10 for 10. Will continue dry firing every day. Bragging absolutely not more like pleasantly surprised. Thank You Chris”
“Have a bad flinch. I tried this method and it didn’t work for me. I’m not complaining. It might work for some people. This is just feedback.”
“Excellent suggesrions Chris! I worked hard about 14 years ago to overcome the flinch. I did LOTS of dry fire practice with some .22 rimfire practice sprinkled in. It took me something like 9 months of consistenr effort to work through it.”