ok well i was talking with Sacrieur about this a few minutes ago. He suggested to post my idea here

For drophacking detection, Is it possible to detect what a person is typing in their msg box before they submit the text? If so, you could technically detect when someone is about to attempt a drophack and have them removed from the game while also providing a display msg to the rest of the players in the game letting them know he attempted a drophack and he was removed.

The second idea is for detecting Map Hack. If this works correctly, then hacks all over will need to be completly reworked to avoid detection. Apparently when a maphack is being toggled, it provides a msg on the left hand side of the screen with the other text that you are typing. Think of it that you can detect the exact text that would be there when you turn ur map hack on. Instead of just removing the player from the game and dropping them, you could just create a display msg to the rest of the players saying that this person has turned a maphack on. And when the player turns the map hack off, you can do the same thing to detect the text saying its off and provide another msg telling everyone that their map hack was turned off. Im guessing if this works correctly, you wouldnt need to check for certain key presses that the maphack is hotkeyed to?

What do you guys think about these ideas?

Yes. Kinda.
No.

This used to work, but as I recall, they found a way to display the text without it writing to the typical area where we read display text messages.

Also:
Again, not possible.

Edit: The address for what a person is typing starts at 0x0479008C. Never mind. It appears to change. I've found it on 0x0479008C, 0x041A008C, 0x06CF008C, 0x06D3008C, and 0x06A4008C.

Post has been edited 4 time(s), last time on Aug 28 2011, 5:05 am by Roy.

For the drophacking part. i was told people drophack by typing "/drop (player)". is that correct? would u just have to detect when someone types in /drop. alsoooo... this might be a stretch, but can u just auto erase what they are trying to type when they get to that point, so they cant even finish typing the drophacking trigger?

Roy, why do you say it's not possible to display messages to the rest of the players that someone was cheating?

This is entirely doable. If P1 is hacking, you Add 1 to a specific DC for All Players followed by Subtract 1 for Current Player. All players have a trigger that says "If DC exactly 1, display P1 is hacking message and subtract 1 DC".

If P3 is hacking, you add 3 to that specific DC for all players and have a trigger that says "If DC exactly 3, display P3 is hacking message and subtract 3 DC".

There's a slight chance you'll have multiple players hacking at the same instant, in which case you just need to use a DC per player or use binary arithmetic to isolate the player that the message is appropriate for.

But overall I don't see why this wouldn't work. I use a similar mechanism for various global messages in my DS map.

Edit: unless this is because the methods proposed rely on EUDs and therefore it's the whole local/global data problem.

No, it is not doable. If P1 is hacking, you add 1 to a specific DC for All Players, then that player is desynchronized.

Let's say Player 3 is hacking and there is a trigger that adds a death count for All Players. On Player 1 and 2's machines, all death counts will still be 0.
Player 3 drops from the game. No message appears.

If everyone has drophack protection on, then instant drop without warning from drophack protection means they were detected

Yes, this is usually the standard implementation. However, reading the address from where the player is typing is not as easy as one might think. The addresses it is found on are not base addresses, and it seems that the address being written to changes each game as I mentioned above. You could try to catch most of these addresses and just write EPDs for all of them, but that comes with some minor risks. I highly doubt the other addresses would have the int value for "/dro" exactly, but I can't say for certain.
Only with EUD Actions enabled.

And like Heinermann described, the EPD would be localized and the trigger would only run for the hacker. No other players' game data would change, so you'd have no information to know when to display the text message.

My experience with Anti-hack:

1. Put a Unit that'd crash people when they look at it at start location, without any way of the players seeing that unit unless they maphack
... this got patched in the next maphack ver
2. Put another Unit that'd crash people when they look at it.
... this got patched again
3. Use EUD's that detect something (not sure what, b/c someone did that for me.. this was in v1.5)
... this got patched as well

I thought disabling doodad state would be #1

I assumed that was one in the same as "unit that'd crash people."

Well #1 and 2 are saying the same thing, I think of that #2 would be unused crashing units/buildings and sprites

The ordering and details of UU's post are irrelevant; he was making the point that any anti-hack measures mapmakers take tend to get patched or worked around. Please try to stay on topic.

I think the only realistic anti-hack is to create units that can only be seen/selected by someone with map-hack on and use EUDs to detect if they've selected the invalid unit. If they have, they must be using maphack and can be dropped.

2 problems with this:
1. You have to be absolutely sure that regular players will not be able to select these units during regular play.
2. It relies on the map-hacking selecting the units. If they don't select the units, you can't crash them. Probably they will learn after the first time they do it and know not to do that any more. It may also be obvious from the unit placement or unit type that you've set this up as a specific map-hack detector and so canny players may not click on them anyway.

This method would have the advantage in that the hack would be map-specific and therefore not really fixable by the hack authors (unless they simply patch whatever method you are using to drop the hacking player).

Post has been edited 1 time(s), last time on Aug 29 2011, 11:29 pm by Lanthanide.

Hmmm. Is there a way to force the play to perform a left click during gameplay? If so, you could just have a crashing unit in the very beginning and have it click instantly on the crashing unit.

No. If EUD actions were enabled, you could force select a unit, but this would make every player select the unit regardless of vision, defeating the whole purpose. As far as I know, there isn't a way to create artificial user input through EUDs.

Unless hacks can be written to rewind time or predict the future, I don't think they can cure desynchronization. Once you detect a hacker, there's not much they can do. The name of the game is how to detect the hacker. The suggested methods in the OP do not seem to be reliable for accomplishing this task.

But what if you tried this at the very beginning of the game on an area where nobody can see. It would just be for the 1st second of the game and then have it removed after that test has passed. Though this would just test if the hacker had it on when it started i guess.

If you use EUD actions to select a unit, it will work for ALL PLAYERS, regardless of if they can see the unit or not. This means that you'd always detect everyone of "hacking" through this test.

I'm almost certain there's no EUD you can write for "Left-click mouse at this coordinate."

Ok thats understandable. Bummmmmer haha

Can't you detect what unit is selected and the coordination of the mouse?

Post has been edited 1 time(s), last time on Aug 29 2011, 10:47 pm by xboi209.

Yes, you can detect those things independently, but this doesn't progress IAGG's idea in any manner whatsoever. He was suggesting a forced click to fix the flaw in Lanthanide's suggestion. A forced click is not possible through triggers.