Automatic Motion Analysis versus Bots - an option?

TigerLotus

Thinking about the bot problem and a possible solution.

The bot's are run by scripts, they follow defined patterns if they
follow a patch, they warp, they teleport.

I have been traveling alot with my mobile GPS device in real
world and so I got the following idea:

The motion of all charakters are known to the server,
so the server can track the path of each player.
Like using a GPS device you can create graphs from
this data and analyse this data automatically.
With this data it would be possible to identify and ban many
types of bots.

Speed of Motion is the gradient/steepness of the motion graph,
if you know which speed of motion is possible with mounts and buffs
you can define a range of steepness that is allowed. All Characters
that are out of this value are hacking.

If a bot is teleporting, the steepness in the analysis would
go against infinite.

Additionally you could analyse the used commands and
the time between those commands. If the allowed
time between (legal) used commands is known,
you can sort out people who hack and use commands
more often than it should be possible.

To prevent people from moving "under" the world, you
could place unbeatable monsters there under the world, that speedport
to players in their line of sight (everyone who is under the ground).
Or analyse the height coordinate of the player on his location
in Relation to the groundlevel with a script.

It should also be possible to analyse the path, that a
character takes, and if this path (or a couple of switching paths)
is repeated. This should be possible with a pattern recognition.

For all those characters, that you "capture" by the mentioned
methods you could trace the routes of the cash flow, mail traffic
and direct trade, especially who is buying and who is selling large
amounts of items/gold. You could then mark this charakters as "suspicious".
If the same character becomes "suspicious" again, because
he is trading with botters again, he would be banned as well.

This way you could create a full automatic system agains bots.
Most of the functions described have been used in other
software, so there is not really something new to invent.

The recorded data could be used from GMs if there is someone
objecting his ban.

rvr1982

I might be wrong, and don't really know how bots work, but i guess that if it's recorded key and mouse movements, with the latency the path will be different each time. Slightly different, but still different.

TigerLotus

You are right with this, but those slight differences are more like
a "noise" on this path. So if someone runs a path 20 times
and you filter the noise out you should be able to clearly see,
if this is a bot.

mips_winnt

You do of course realize that the development of such a system would cost money, A LOT of it. Businesses operate on the basis of return on investment, if they can't expect a reasonable rate of return within an acceptable time horizon they don't spend the resources.

So first question is .... what's the rate of return on the system you're proposing?

TigerLotus

Hard to say, I am not an economist

But if there is a huge player-base shrinking fast because
of a bot problem, that money would not be wasted.
Especially because you could recycle this technology
for future projects or sell it to other developers of games

Censorious

Random is easy.

A child could program a bot to take a drunken walk from A to B that could never be distinguished from human behaviour.

Ever hear of Fractals?

mips_winnt

TigerLotus wrote: »

Hard to say, I am not an economist

It's not an economics question it's a question that anybody that proposes a project involving CapEx and OpEx is going to have to answer before their project gets approved.

But if there is a huge player-base shrinking fast because

We don't know if their player base is "shrinking fast", we don't know what their revenue and profitability targets are, so as far as "we" know ZOS is doing just fine without having to deploy additional capital and operating expenses.

of a bot problem, that money would not be wasted.
Especially because you could recycle this technology
for future projects or sell it to other developers of games

You have to understand the "bot problem" isn't a technology problem, it isn't created by this technology or that technology, it's created by the PLAYERS themselves. There's a certain percentage of the player base that demands the services that these bot runners provide and given that building botting software is fairly easy and low cost trying to solve the problem by throwing enormous amounts of CapEx and/or OpEx at it is not a great business strategy (unless of course your goal is to go OUT of business). This is especially true for a franchise like EOS which has not yet established itself as a viable long term business worth sinking large amounts of additional capital into.

So while your idea has merits from a purely functional standpoint, it's all pie in the sky from a business perspective because you don't have a case that supports a realistic expectation of acceptable ROI from it.