MArcomage

Marcomage is very good game but people don't know about it. Many people enter to site and exit at once, because they don't quickly find a game (or they are embarrassed to play with people, or tired of waiting of the opponent turn). Make simple AI with deck containing only attack cards. For example, let AI always select a card with maximum attack value, else discard most expensive card. And make large button "Play" in main page. After newcomers click to "Play", the game with AI start immediately without registration. This idea will greatly increase the conversion visitors to users and so increase number of players.

That kind of AI though would prove too weak as it would consume a lot of resources all the time. Also making a simple AI like that might cause bugs. In addition, an AI whether simple or complex might prove a bit heavy for the server, as probably everyone would start playing games, just to have something to do between turns.

If we want to keep visitors then we could do 2 simple things. Have the site send them a personal message along with their sign-up, which will welcome them and then warn them that players might take some time to finish a game so they should be patient, if they want a quick game they should look for the crossed swords and if they want to play a full game then to ask their opponent. Another solution would be to designate a couple of players each month to be challenged by newcomers and play them full games and explain to them the rules. I see some players who are online most of the time (myself included), so it won't be hard finding volunteers. We just include their names in the "Welcoming Message" and suggest that newcomers can challenge them in order to be tutored in the game easily.

Lord wrote:
That kind of AI though would prove too weak as it would consume a lot of resources all the time.

My own strategy is like this AI and very effective. Of course my strategy is not exactly the same. But I think such AI will be hard enough for newcomers. The deck should consist of cheap and effective attack cards. Also the deck should be balanced by costs of bricks, gems and recruits. Example of appropriate cards: Ghoul, Dire wolves, Wolf spirit, Stone gargoyle, Urk's tribe, Giant spider, Basilisk, Mercenaries, Army of volunteers, Obsidian gargoyle, etc...

Lord wrote:
In addition, an AI whether simple or complex might prove a bit heavy for the server, as probably everyone would start playing games, just to have something to do between turns.

MArcomage is interactive turn-based game. So MArcomage is not very demanding of server resources, because requests to server are submitted only when the player makes a turn or refreshes a page. Usually games like MArcomage used 1 server for 500-1000 online players.

Not all servers are created equal. That there exist servers that would easily handle 500-1000 simultaneous players does not preclude that this server is up to that task. The fact that it goes down sporadically every week or so would add weight to the argument that it is not.

I don't think we need to designate people in shifts to welcome newcomers. I always look at level and games played by new challengers and hosting players, and I offer advice and explanations throughout (even if it means giving myself a disadvantage). As long as there are people doing this anyway, and being communicative throughout, all is well.

The simple fact, however, is that some people are going to play a few games and then move on to something else, whether you're welcoming and helpful or not.

Writing a tutorial would be neat, leading a player through a preset game with explanations of what's going on and why numbers are changing the way that they do, all in the frame of a real game screen. Since it wouldn't need to calculate anything, there would be no need for an AI or potential system resource hogging.

Another thing worth noting is that the site Introduction includes a link to download single player Arcomage.

I feel like the number one way to make this site more welcoming and attractive to newcomers, would be to clean up the site's design. There has been some work done on this, but the site still looks like someone's GeoCities page from 1997. I love the site to death, but that's one of my few complaints.

I think we could make the site look a lot more modern if we went with a very simple colour palette, got rid of all the cheesy font effects (shadow, and highlight, and so on), and had, by default, a light, monochromatic background.

ETA: I think the argument that "It'll make the game too popular and crash the server!" is a very poor one. Don't we want the game to become more popular? Don't we want more people online to play with? I know I do.

DPsycho wrote:
Another thing worth noting is that the site Introduction includes a link to download single player Arcomage.

On the other hand there are many people that played and liked old Arcomage. Fast game parties are important reason they like and remember old Arcomage. But fast game parties without registration are possible only with AI.

Also I think AI will be interested for existing players too (for example, for testing some ideas). In future we can develop more complex and strong AI strategies.

In terms of a tutorial the simplest way I can think of to do this is just link to a youtube video of someone playing through a game or replay and narrating over what's going on. That doesn't involve any extra coding on the site, and theoretically anybody could do it if they had time. I wouldn't mind doing it, but I can't set my resolution high enough to be able to record the entire playing field.

While that could work, I think a series of screens (or a Flash animation) where the "player" clicks through at his or her own pace would eliminate a few problems.

I added download links to original Arcomage and Arcomage Tribute - a single player Arcomage open source project to main webpage in r1112.

Mojko wrote:
I added download links to original Arcomage and Arcomage Tribute - a single player Arcomage open source project to main webpage in r1112.

Better remove Arcomage Tribute. It is buggy unbalanced version. MArcomage is unbeatable. MArcomage is balanced and have 400 cards. I am a programmer and I think it is not hard to implement AI into MArcomage. You already have a ready-made interfaces and effective AI and deck idea. I will help if you accept the decision to make AI.

I requested a partner link on the Arcomage Tribute project page. For exchange I placed partner link on our main webpage. (btw. we have 598 cards not 400 :P).

Now regarding the AI. Implementing a dummy AI, that just plays turns on some very simple rules is indeed simple. But I don't see the point to implement such thing. Creating very smart AI, now that's complicated. Moreover I'm not convinced this is really necessary. If you want to experiment with decks you can play friendly games.

Mojko wrote:
But I don't see the point to implement such thing.

The point is implementing AI will greatly increase the conversion visitors to users. Do you want success to your project? Just see 3 and 4 pages of player rating. And there are 10-50 pages of users that are not shown in the rating because the rating doesn't show the users that visit site more than month ago.
Now there are 4 steps newcomers should pass to play the WHOLE game:
1. Register.
2. Go to "Games" (it is intuitively not obvious for newcomers).
3. Wait or find the game (it is intuitively not obvious too).
4. Start the game. Constantly wait for opponent turn. The game can last a week.
In each of 4 steps many newcomers exit the site because they are losing patience.
Implementing AI together with button "Play" in menu allow to skip these 4 steps and get MArcomage gaming experience.

Here is deck http://rus00.narod.ru/attack.csv. It is like survival mode for newcomers.

If I may play devil's advocate here, you're basing all of this on the assumption that we all want the site to expand significantly, and you're suggesting that we do this by appealing more to people who may be too impatient to play the game as it is now. Let's consider these things for a moment.

What might happen if the site were to become much more popular than it is now? There would be positive effects, but the site would require much more bandwidth for the increased number of games and forum traffic. The devs might have to migrate to a better and more costly server, and their workload would increase as there is higher potential for site issues and several more concept submissions (many of which would likely be redundant) to go over.

Additionally, this website is based off of an intellectual property that the devs don't own and uses hundreds of images that belong to several, some unknown, individuals who probably all have their own idea as to what constitutes fair use. More popularity means more exposure, and that means a greater possibility for copyright holders to take notice and decide whether to attempt to take action. While this scenario has been discussed at length on these boards already with differing opinions, in no case would a confrontation with copyright holders be a "good" thing.

You want to increase traffic by being more accessible to people who may not have the patience to take an entire week to play a game. While this would certainly broaden the numbers for the potential audience for this site, I can't help but wonder. Do we really want an influx of impatient players? Maybe we do, but maybe we don't. I personally don't mind a player base that's made up of a more mature and thoughtful group. It's quite possible that not appealing to people with short attention spans is actually part of this site's charm.

Now, I am playing devil's advocate here, meaning that aside from the lines where I used the word "I", I don't necessarily subscribe to these thoughts exactly, and I acknowledge that these are all things that could be specifically addressed and overcome. I just want these ideas out there -because- they would have to be addressed and overcome.

Hehehe, still it's a valid point DPsycho. One thing I'd like to add is that if we implement the AI then most players (especially newcomers) will just keep playing Single Player as it will be faster. Furthermore if someone doesn't wish to play a week-long game, then they can simply search for quickgames. For example Zika always asks for quick games and will always ask beforehand if we can keep playing.

Of course it would be nice to have a Single Player mode as well though. But it would require too much time and work. Maybe people with programming knowledge could help create this AI, but still I have doubts whether it will be useful for the site or not.

And btw, Arcomage Tribute is still a work in progress and it's been worked on pretty decently. Also the guy who made it is quite motivated and he just released a testing version of it. Maybe a cooperation with him could be useful in implementing harmonius's idea. Ah and you should also know that that Arcomage is only the original version that appeared in early Might & Magic games. The version we are using now is the new version that was implemented in M&M 8 I think.

It is now possible to play against a computer opponent. Since it is my very first AI I created, any feedback and improvements suggestions are welcome. Please report any situations, were you think the AI made a very bad choice.

I have several other improvements that can done, if necessary:

- AI difficulty setting
- AI strategy setting (specify the targeted victory condition) - can be useful when you give AI to play with your own deck
- AI challenge games (play vs overpowered AI opponent for gold rewards)

How come I can't chat with the AI? :(

This is pretty awesome. And looking at the replays so far, it looks like the AI is doing something right, as it has close to a 50% win/loss record.

Why would you want to chat with an AI? o_O

Currently he wouldn't be responding in any way. It would be possible to make a text parser like POD bots in CS had, but I don't think it's necessary.

Oh, when I suggested an AI in the thread about MArcomage's popularity, I didn't know it had already been talked about for like three months x)

The AI puts up a good fight. Kept lowering my resources. The fight dragged on but I really thought I was going to be put to shame by a computer... :p (admittedly I played fast and not very well because that wasn't a "serious" game, but still !)