Ranks and ranked games

Kevin · Post by **Kevin** » January 23rd, 2006, 1:42 am

i think it would be better if perhaps there was a way to rank players so you know if your playing a really good veteran or a begginner

and perhaps the only way to get a rank would be to play in a certain ranked area? But also keep around an unranked place for people who don't care about they're rank and just want to have a fun time? maybe?

feel free to tell me why or why not this would work...

scott · Post by **scott** » January 23rd, 2006, 2:22 am

This has come up a lot in the past. Some players like it and some don't. I think the biggest hurdle is the need to have player accounts, which apparently isn't trivial.

Kamamura · Post by **Kamamura** » January 23rd, 2006, 3:02 pm

Moreover, you would need a robust anti-cheating system that is even less trivial.

Artis · Post by **Artis** » January 23rd, 2006, 6:46 pm

Kamamura wrote:Moreover, you would need a robust anti-cheating system that is even less trivial.

It shouldn't be too diffucult to do all "sensitive" processing server side. The only problem I can think of is server load, or am I missing something?

Post by **Darth Fool** » January 23rd, 2006, 7:00 pm

Artis wrote:
Kamamura wrote:Moreover, you would need a robust anti-cheating system that is even less trivial.
It shouldn't be too diffucult to do all "sensitive" processing server side. The only problem I can think of is server load, or am I missing something?

The only thing you are missing is how difficult it is to overhaul the engine to do just that. Some steps that will need to be done before this is possible are:
1) enable encrypted WML to be sent.
2) enable a user/password system.
3) enable all random # generation and events to occur on the host machine.
4) make sure that the load doesn't kill the server.

So, while the developers support the idea, and in the long run it will undoubtedly happen, it will not likely happen anytime soon.

Artis · Post by **Artis** » January 23rd, 2006, 7:16 pm

Also visibility. It's certainly a time consuming problem, but not difficult in the "hard to solve" sense. Either way I'm not impatient--I have an excelent game to play.

Rune · Post by **Rune** » January 23rd, 2006, 10:16 pm

Darth Fool wrote: The only thing you are missing is how difficult it is to overhaul the engine to do just that. Some steps that will need to be done before this is possible are:
1) enable encrypted WML to be sent.
2) enable a user/password system.
3) enable all random # generation and events to occur on the host machine.
4) make sure that the load doesn't kill the server.

When you mean server, you mean the machine of the player who host the game, right? Unless I'm completely wrong, then that would mean that even with all those 4 points implemented the host can still cheat...

scott · Post by **scott** » January 24th, 2006, 5:16 am

There is a central server in the bowels of cybernoth that hosts the games. The computer that creates the game doesn't actually host it, as far as I know.

Kevin · Post by **Kevin** » January 24th, 2006, 8:43 pm

scott wrote:There is a central server in the bowels of cybernoth that hosts the games. The computer that creates the game doesn't actually host it, as far as I know.

Why is it that when the host leaves there is a connection error?

Is there some way to fix that?

scott · Post by **scott** » January 24th, 2006, 8:53 pm

Kevin wrote:
scott wrote:There is a central server in the bowels of cybernoth that hosts the games. The computer that creates the game doesn't actually host it, as far as I know.
Why is it that when the host leaves there is a connection error?

Is there some way to fix that?

There must be something special about the creating computer. The individual players have the calculations done on their own computers and send the results to the server, which updates the other players. I don't know for sure the role of the creating computer; this is just what I have picked up over time about how MP games work. There's an expert around here somewhere with the real answer.

Tux2B · Post by **Tux2B** » January 25th, 2006, 1:17 pm

When you play Wesnoth for a while, than you know who is good and who isn't.
In fact, I don't think it is even very useful. I've just come back on multiplayer after not having been on it since 6 months, and it was quite hard at the beggining. But since I discovered Wesnoth almost one year ago, I (could) be considered as being in an 'older' category than some others.

(Thinking twice about it, it's odd how it seems a short time ago we had version 0.8.1 with the old elvish graphics, no lvl3 orcish archer and those old gryphons...)

And, BTW, kill that animated avatar.
Or even change it, because there is already someone using a similar one.

Keelta MacRonan · Post by **Keelta MacRonan** » January 26th, 2006, 1:07 am

One second. Just because you have been playing for a while doesn't mean you are good. I have been playing for almost a year. But I'm still worse than most people on MP.

Wintersmith · Post by **Wintersmith** » January 26th, 2006, 2:47 am

It is possible to prevent cheating in the random number generation without making the central Wesnoth server do it all. There are a number of cryptographic protocols which enable two people communicating over a network to generate random bits in such a way that both parties can be satisfied that the bits are truly random and the other person could not have cheated to fix the outcome. If you have a copy of Schneier's _Applied Cryptography_ on hand, take a look at sections 4.10 and 23.7.

EDIT: The RNG issue became a new thread here

I haven't done any development work in Wesnoth before, but I would be willing to implement one of these cryptographic schemes into the game if the Dev team wants to see this happen.

What are the other possible methods for cheating besides the random number generation? The only one I can think of is to cheat by illicitly lifting the Fog of War / revealing hidden units (I don't have any suggestions for good ways to solve this). Are there other potential methods of cheating?

Post by **turin** » January 31st, 2006, 1:20 am

Cheating with fog/shroud would be nearly impossible to prevent unless you are going to disallow observers for all ranked games. Otherwise, one player could just create a dummy observer in another window and see what the opponent is doing behind the wall of fog.

Really, I don't think stopping cheating is an attainable goal, or particularly desirable; the losses of freedom would be less than the gains in security ( - a Schneier reference). What we could and should do is make cheating harder, and then we could allow ranked games, which would not be cheat-proof, but we wouldn't claim they were.

Duke Guillermo · Post by **Duke Guillermo** » January 31st, 2006, 4:29 am

I agree with Turin. (Holy Crap, what's happening?!?

) A completely cheat free system isn't necessary to have ranked games as no system is completely cheat free in any game. People have found ways to win at Halo, World of Warcraft, and the list continues ad infinitum. You shoudn't make rules based on the exceptions, those who will cheat. For those who don't cheat having a ranking system will still be advantageous.

And, maintaining a cheat-free Wesnoth would be easier to do than maintaining a cheat-free commercial game because the community is smaller. If necessary, an IP address can be tracked to find out activity etc. If players who cheated had their accounts destroyed and their IP addresses banned for a # of months then the # of cheaters wouldn't be very high. As for the observer issue, a game can be made to be unobservable. Or it could be made that no two accounts from the same IP address may be online at the same time. It's a start.

The Battle for Wesnoth Forums

Ranks and ranked games

Ranks and ranked games

Random Number Generation