Start banning x-ray miners

Maybe some people think, ooo! I found diamond ore! There might be more around! ..? I know that's how I think when I go mining haha

 

I'm totally with you on banning x-ray-ers, however the critical point is: how do you tell an "x-ray tunnel" from a "non x-ray tunnel"? Investigator lets you know who destroyed which blocks (I have the feature), but how do you know that that particular path was "chosen" due to x-ray and not just by wandering randomly?

 

orthogonal (if you go the opposite way, you still go along it, though backwards. I know, I'm a math nazi, no need to remind me)

 

Technically it would just be perpendicular since you are probably just mining on the plane of y=11, orthogonal would be used if you were to go directly upwards, although many say the two are interchangeable.

 

I agree we need to catch hackers and xrayers, but the staff already handle this daily. They use plugins to detect hacking and suspicious activity. The problem with reporting someone over a sketchy tunnel is that people have very different mining strategies. I tend to mine around certain ore strains because I know that Minecraft often generates ores right next to each other (especially true at y=11). This could look like a "sketchy" tunnel but in reality I am just using methods I have found to work best. I think that the staff detect hacking well, even though you may not see all the work they do.

 

Turning after every diamond pocket.

 

Orthogonal doesn't mean to go directly upwards: that would be vertical. Orthogonal and Perpendicular are actually almost synonyms (a quick search on Wikipedia can confirm this); the slight difference is that orthogonality is a generalization of perpendicularity: in particular, perpendicularity applies only to "simpler" objects such as straight lines and planes, while orthogonality applies also, but not only, to curved lines and surfaces in a local sense.

The fact that a path appears to be random doesn't automatically mean that it was due to x-rays.
A random, strange turn would be a (very weak) evidence for x-rays only if you ascertained that it led to unequivocally unexposed ored.
And you would need to collect a lot of such weak evidence in order to be reasonably confident that those were truly due to x-rays and not just randomness.
Yes, "impossibly lucky" is suspicious indeed, but first of all you need to estimate how tiny the probability is: if it's not tiny enough, that might just be a false positives, and we don't want them. In order to proceed correctly, you need to:

• make a mathematical model of the ores distribution;
  • do a few simulations of "random miners" in order to have a statistical control sample;
  • in this way you can estimate how weak (or strong) a [strange turn towards unexposed ore] is as evidence;
• you would need to detect a ton of [strange turns towards unexposed ore]";
  • how many, depending on:
    • how weak an evidence you previously estimated a single one to be;
    • the prior probability (in the Bayesian meaning) you estimated of a random player being an x-ray-er (good luck with that);
    • the maximum rate of false positives you don't want to exceed (because we don't want to ban innocent players too easily);
  • all those [strange turns etc] must belong to the same player and, if I didn't stress this enough, must unequivocally lead to unexposed ore: otherwise they would be meaningless.

As you can see, even a professional statistician would need a lot of time to figure out a reliable data analysis algorithm, and then crunch the numbers.
So, again: I'm totally with you that x-ray shouldn't be tolerated and I strongly support the idea that proved x-ray-ers shouldn't be allowed back.
But I can't support your "I counted a bunch of suspiciously random turns, they can't all be that random" claim: that's just not how you proceed in Science, let alone how you want to prove someone guilty beyond reasonable doubt. If you want staff to prove someone guilty that way, they'd need to do a literal lot of work; but then it would require much less time to just follow suspects in /v and see them while they are mining, rather than reconstructing their behavior analyzing the tunnels they left.

TL;DR: it's unreasonable to prove someone guilty this way; the most you can do analyzing tunnels is to pinpoint suspects, and this alone would require quite some time.

 

It would be really amazing if someone coded an anti-hack plugin to fit what you just described. That would be really cool to take a scientific approach to detecting hacking.

 

Doing it at plugin level would be almost automated: the "only" time spent by human beings on it would be

• the developer's time to code it in the first place (a few months of work for one single person);
• mods' time to type the command "gimme the list of suspects detected" (almost negligible).

And the obvious pro would be that it would record suspect behavior real time, with no risk related to log cleanups.
But then: there's already various anti-cheat plugins to detect possible x-ray-ers; I don't know every single one, but I wouldn't be surprised to discover that at least some of them already implement something along the lines of what I described. Why are we assuming none does?

 

Where exactly did you get that "1 in a million" probability? One thing is to make rough estimates, and use them just for the sake of clarity (as I myself did here); another is to make up numbers totally out of thin air, just to please our own argument, and expect those numbers to be accurate estimates.
My very point was exactly not on the accuracy, but on the order of magnitude of such a probability. 1 in a million? If you truly believe it to be a reasonable, rough order-of-magnitude estimate, you clearly have no idea what the word "probability" means in the first place. And that makes me a saaad panda (cit.)

 

Guys, this is a suggestion. Stick to the topic at hand.
@Tandisong would you still like this suggestion locked?

 

As per request - Locked