[DOTA] JacobianMatrix - MH - USWEST

[Gauss]

Replay Link:
(1) https://entgaming.net/findstats.php?id=6262093
(2) https://entgaming.net/findstats.php?id=6262562
(3) https://entgaming.net/findstats.php?id=6262837
(4) https://entgaming.net/findstats.php?id=6272798
(5) https://entgaming.net/findstats.php?id=6288993
(6) https://entgaming.net/findstats.php?id=6299783
Game Name:
(1) [ENT] DotA apem us/ca #25
(2) [ENT] DotA apem us/ca #70
(3) [ENT] DotA apem us/ca #86
(4) [ENT] DotA apem us/ca #64
(5) [ENT] DotA apem us/ca #51
(6) [ENT] DotA apem us/ca #80
Your Warcraft III Username: Noobsheet
Violator's Warcraft III Username: JacobianMatrix (aka 7708801314520 aka recovery aka middlebiddle)
Violated Rule(s): Maphack
Time of Violation (in-game or replay): 0-10 min of dota time (so I guess around 2-12min of replay time) (check explanation below)
Any further thoughts:
This guy knows not to fog click, so I didn't bother with it (proof : https://entgaming.net/forum/viewtopic.php?t=53893, second post)
However, I have found a pattern so obvious it cannot be just "random" or "map awareness" or whatever bullshit people might call it.
I decided to check the way he went for the runes the first 10 mins of each game.
Now, why 10 mins ? Well, the first 10 mins are extremely important, especially for meepo to get kills/lvl advantage, and in the first 10 mins nothing really happens, so people are not likely to detect MH usage. Once the first 10 mins are done, meepo starts ganking and one cannot really say whether chancing upon rune is pure luck or not (What I mean is that if Meepo ganks bot and the rune is bot, once the gank is finished, it isn't suspicious at all for meepo to find the bot rune once he comes back mid for example, so I did not take that into account. I only checked the way meepo finds rune when there is a strict 1/2 chance of finding it. That only happens at the start when Meepo have two choices : go bot or top when he is mid and there are no wards.

What I found out is that he SYSTEMATICALLY finds the rune, and never gets it wrong. He never checks for the rune, get it wrong, and go the other way. He just goes straight to it.

Bonus : in one game, before min 2, he goes traight to the secret shop to kill courrier. It might not be that suspicious in that an experienced player knows courriers are usually around there, especially at the start when most courriers are ground-based, but what is weird is that he doesn't do that often, but when he does, there just happen to be a courrier standing there, motionless, which doesn't happen in every game... So yeah, still somewhat suspicious. Here are my statistics :
Game number / times checking for rune / times finding rune straight away / interesting facts
(1) / 2 (3*) / 2 (3*) / *one of the three "check" isn't THAT suspicious, iirc he is coming back from a gank or something, and "chances" upon it. NOTICE HOW HE FINDS THE COURRIER STRAIGHT AWAY
(2) / 2 / 2
(3) / 2 / 2 / Notice how his team mate says the rune is bot and he goes top. of course, the rune might have changed spot since iirc it was after 2 mins of dota game, but remember how this guys always talks about being the best of the best player with awesome awareness, the probability of finding the rune was higher going bot and not top. The only flimsy justification would be him wanting to gank, and "stumbling" upon the rune... but I don't buy it
(4) / 2 / 2
(5) / 2 (3*) / 2(3*) / Okay, one of them isn't really suspicious. The other ones through ... 1/2 probability, still doesn't fail for both.
(6) / 2 / 2

In a nutshell : out of more than 10 random checks, he always goes straight to the rune. Keep in mind the probability of it happening is 1/2^10=1/1024, which is LESS THAN 0.1% (In other words, the probability of failing at least once is 99.9%, and he didn't. No amount of experience can affect this probability since it is totally random (or supposed to be))
Even if the "evidence" is empirical, it is extremely unlikely for it to be wrong, especially if we account for all the ban requests already against him
I only checked a few games (around 15-20% of them), there are probably a lot more

I hope I didn't "shuffle" game numbers and description, but actually description isn't even needed. one simply needs to watch the first 10 mins of geo play and rune seeking to notice how obvious his MH is. He usually finds 2-3 runes every game in the first 10 mins without ever being wrong. No amount of map awareness can foretell the spot of a rune randomly distributed between two sides without checking and possibily getting it wrong a few times.
So yeah, he either has a full-on maphack, or a softer hack which only indicates rune position (which i doubt), but still a third party hack program nonetheless.
You be the judge.

Memaru : if I'm wrong, care to explain how you systematically find the rune then ? Inb4 12 bs stories... happy writing

[dotaworx1]

@Gauss The information you presented might be a bit misleading. For example, if 1/2 were raised to 10, this could happen if you truly sampled randomly or the 10 checks were contiguous in one game, however, you found 2 instances in each of the 6 games you listed. I think the probability is more like: (1/2)^2 = 25% each game. Now, if you compare this with the number of games where this doesn't happen we can understand what this number means. Just tell us how many games total you have reviewed.

[AIDSHammer]

I heartily applaud the science and mathematics being applied to this matter!

[Gauss]

I have watched around 10 games, which means it does not happen in around 4 games indeed. But when I mean it does not happen, he either ganks or does not really "decide" to find the rune (e.g. he gets high lvl very fast and starts to gank. in this case the 1/2 probability does not apply obviously)
The fact that games are independent means that 25% each game is 0.25^n where n is the number of games.
Finding the rune once or twice the first mins is not suspicious. Doing it for 6 independent games is, without him failing or going the wrong way ONCE

I have never saw him failing a check in all the games I checked. So the 4 remaining games are 0 / 0

I know however I have no proof as to the fact that I did not cherrypick these 10 games out of the 60 (which is obviously possible for anyone with a malevolent intent to skew statistics).

Well the only way to prove the statistic right would be checking the 50 remaining games... but I don't hate the guy nearly enough to do that... if anyone else is willing though, he won't be disappointed. Actually I found this randomly, since I downloaded the 60 games thinking he was a good meepo, wanting to learn. He talked a lot about map awareness bullshit, so I decided to check... and he always finds the rune... wow, so now to be good at dota I need to find the rune straight away... but how harharhar... well. So yeah, the 10 games were chosen totally randomly, and Im a bit angry to have wasted my time over a bad player pretending to be good ("best player on ENT"). Him being known in dota 2 or whatever doesn't mean he's legit in here.

[dotaworx1]

@Gauss 0.25^n seems reasonable (if every game resulted in the same unlikely pattern), and according to your limited data, it is "possible" that he uses a third party tool. Another interesting metric would be the time interval between successive positive rune checks in the games you analyzed.

[Jiminy_Cricket]

@gauss

I watched a replay someone else posted on JacobianMatrix. I saw the same thing youre describing. Perfect backing and finding the right creep spot immediately and swinging over for the rune when it's one he really needs. All game, got around 10 time stamps of this. Like you said, the probability of being perfect all game is so low and when compounded with multiple games the probability nears impossible. Mh for sure. I sent Hazardous my thoughts on the replay I watched.

[Invoceur]

@dotaworx1 @Gauss 0.25^n is correct taking into account the first two rune checks. The time interval between succesful rune checks is irrelevant because runes do not disappear once they have spawned. Well played though, i dont think i have seen anybody get banned for succesful rune checking. But putting this into a probabilistic measure is definately a strong case. Also @hazardous @larz please get around to reviewing my initial request.

[Jiminy_Cricket]

@invoceur

It was your request I reviewed. You left many time stamps out that were significant. I reviewed the first replay you posted and thats all I needed to see. He's mh. I sent my review to Hazardous.

[dotaworx1]

@invoceur It may be interesting to note that if the time interval is constrained to a small range of values then it shows there was a methodical and planned/limited use of a map hack that resulted in statistically improbable successful rune checks over many games.

[Larz]

6262562.w3g

(286.21 KiB) Downloaded 22 times

I will review this again This thread will be locked until then. I will make one comment this player is on another level and I seriously doubt he hacks. I have reviewed over 15 of his games in the last two months and when everyone thinks he is hacking there is always something that shows he had vision. I have had the opportunity to play with and against him under various alias one of them being arty.
https://entgaming.net/findstats.php?id=6262093
in this game he goes right to fuzzybunnys pet and secret shop and kills it no way he knows the pet is there first suspicious movement at 4 min in there really isnt a reason to check secret shop.
At 2 min in goes to bottom river and get rune no vision 50/50 guess
at 4 min in sends meepo bot river again to get rune right to it no guess.
at 6 min in he separates and goes bot river again to get rune three perfect guesses
At 8 min in meepo tps one unit away and sends other to bottom river to again get rune with out vision and 4th perfect guess
Same thing at ten min in. after this its seems as he no longer needs the runes.
next game
https://entgaming.net/findstats.php?id=6262562
and here is the definitive proof
at -1.43 seconds replay time pink clicks roshan with out vision
-0:48 <JacobianMatrix> 0x19: Select subgroup: [Roshan], 0x00007B8200007A56
I apologize for any one i offended in this post
JacobianMatrix banned one year for map hack
most likely will dodge

[dotaworx1]

No need to apologize if you found a fog click that's some solid proof. I didn't know if this guy map hacked, but we'll have to wait for confirmation I guess. I hope @memaru appeals if this is just some big misunderstanding.

[Larz]

@dotaworx1 its already confirmed I already banned him and another mod has also confirmed

[dotaworx1]

@Larz surprising news, but a fogclick's a fogclick.

[Hatedmaru]

No dodging found, fogclick already confirmed previously.

Processing