Hacker News Reader

Comment by Enginerrrd | Hacker News Reader

Both, combined with quite a bit of luck.

>Like... Could you just fire the gun, then put the target where the bullet went an fire again and say its mission accomplished because the real goal is just setting up a gun that stable and precise and deterministic?

No... slight variations in wind and atmospherics across that distance change too rapidly and significantly to make that viable. (Also, you wouldn't need to, you'd just adjust the windage/elevation on the optics and/or fancy rifle mount.)

Also the when the bullet drops from supersonic to subsonic speed, there is a non-deterministic kick that it receives which dramatically reduces accuracy beyond that distance.

View on HN

>Also the when the bullet drops from supersonic to subsonic speed, there is a non-deterministic kick that it receives which dramatically reduces accuracy beyond that distance.

Fun fact, while video games typically depict suppressors as slowing bullets down, and lowering range (presumably with subsonic ammo). A lot of rifles are actually have slightly longer range with a suppressor and standard ammo because it allows the powder to burn more completely with a slightly higher muzzle velocity and therefore it takes longer to hit that subsonic transition

> non-deterministic kick

Actually non-deterministic? Or just difficult-to-deterministic? Genuinely curious.

Very likely deterministic if you knew the _exact_ state of both the bullet and the air, but effectively impossible to know and likely not that easy to model correctly even if you did.

So nondeterministic in the same way an electron passing through two slits is.

I think slightly less non-deterministic than that, because there's at least some way you could know, if you were god.

As far as we know, quantum nondeterminism is real nondeterminism, not just nondeterministic because you don't know the exact state of both the electron and the slits.

Presumably the Navier–Stokes chaos of turbulent flow will eventually amplify quantum fluctuations (where two molecules do or don't interact) into macroscopically divergent fluid dynamic behavior, but I don't know if the relevant Lyapunov time is short enough to make a bullet's flight quantum nondeterministic.

I doubt it. The bullet movement may be chaotic, which means that tiny changes in initial conditions will drastically affect the trajectory, but it is theoretically possible to predict the trajectory. However, it is likely that quantum effects such as the movement of an electron through slits is actually non-deterministic, so we can only predict it in a statistical way even in principle.

Non-deterministic in the extreme literal sense. Our (turbulent) atmosphere is chaotic, and the transsonic shock is a fluid dynamics problem happening in that chaotic environment.

Even with a perfectly still atmosphere it'd probably be fairly difficult to predict.

> No... slight variations in wind and atmospherics across that distance change too rapidly and significantly to make that viable.

This doesn't sound right. If it's true, then the same thing -- slight variations over time in wind and atmospherics across a seven kilometer span -- would make the aimed shot nonviable too. The shooter isn't adjusting for air movement between himself and the target -- because that can't be done -- so if he's hitting the target reliably, the air movement can't really matter.

He's not hitting the target reliably, it took more than 60 tries, and would take just as many to do again

What's supposed to be impressive about "I hit a bullseye by coincidence"?

Impressive that they were able to control things well enough that it only took 69 tries. (Vs. 69,000 or so.)

(Yes, statistical "proof" that it wasn't 10,000-to-1 luck would require far more experimental data. With their setup, that wouldn't be quick or cheap to get.)

> Yes, statistical "proof" that it wasn't 10,000-to-1 luck would require far more experimental data. With their setup, that wouldn't be quick or cheap to get.

Saying that this wouldn't be quick or cheap to get is just an admission that you think it was 10,000-to-1 luck. Why wouldn't it be quick?

Read the article more carefully. Their hugely customized "gun" was only capable of 1 shot every 2 or 3 minutes (to limit heating and thermal expansion ). They had 5 spotters near the target, watching where every individual bullet landed. (Then reporting back to the shooter, so he(?) could make little adjustments.) And it sure looks like an amateur group of "extreme" performance enthusiasts - so who would pay for the collection of a large (say, n=25,000) data set? From a quick web search, just the ammunition (.416 Barrett) would cost ~$400,000. The barrel of their extreme-performance "gun" might need replacing a dozen times or more, and...

The probability of a 1 in 10000 event occurring in the first 69 attempts by chance is about 0.7%.

The 1 in 10000 probability seems arbitrary and underestimated in my opinion though. What would that even mean? The probability of hitting the target with random aiming or purposeful aiming from a fresh start? The probability of hitting the target with a perfected aim and relying on random variance? The probability of doing it without the engineered gun?

But given the precision required, wouldn't it make more sense to aim the gun where it would hit the target given median atmospheric conditions and then just fire with that exact aim until it does? Like you couldn't possibly dynamically aim to accommodate that.

>Like you couldn't possibly dynamically aim to accommodate that.

Yes, you can.

Not doing so would result in dramatically reduced accuracy.

That's just windage 101.

The horizontal variation due to even slight wind is massive at this distance and would dwarf all other factors. Failing to adjust for this on a per shot basis would result in dramatically reduced accuracy.

Seems to me you would not be able to adapt your aim to the wind to the level of precision required to hit a small target. You would be much better off trying to identify a working shot for an average wind condition and then trying to time your shot to match that condition.

That's not true at all.

Even at 500-700 yards, wind holds can be significant (as in off target). That changes second by second. I can tell you that even as a decent shooter with a better spotter, when the spotter makes a wind call, the slight 1-2 second delay before I can actually press off the shot can be enough where I need to change the hold slightly from what he said to make a good hit.

Over this distance, there will undoubtedly be some fishtail winds and dead spots that are changing on a per second basis. In this case, I'd expect you have a team of spotters reading the mirage at several distances to target, and one lead spotter that is integrating those into the wind call for the shooter.

There's also some new tech that's aimed at this problem actually where instead of getting a wind read at one distance, it measures the wind over the entire distance and gives you a hold based on that. Very fancy stuff, they may have used that in this case.

ok wind aside, if that's true, how is it possible to aim it then? It can't possibly be true that you can precisely adjust your aim to hit a specific target if you can't even hit the same target by not changing your aim?

That's why this is such an impressive achievement!

It’s also why they succeeded on the 69th try instead of the 2nd.

But that means the approach described above works perfectly. Fire the gun. Place a bullseye over the bullet hole. Then keep firing the gun until you hit the bullseye "again".

Well, if you consider that "perfectly," yes, but it could take all day. The ancestor post seemed to be suggesting that if it was so "stable, precise and deterministic" that you could do this trivially in a couple shots.

Considering the guy who made the shot commented "we were amazed we got it in so few shots", taking all day is expected anyway.

Reply on news.ycombinator.com