Thursday, June 25, 2009

Skills, spells, and classes

Today, a long-pending project finally got turned on for general use in the game: monster lore, and a few specialized versions for particular mob types. The current specialized skills are demon lore, elemental creature lore, undead lore, animal lore, and dragon lore.

In coming up with appropriate class distributions for these, I once again ran into the shortcomings of only having the four basic classes. There's really just a few things that don't fit nicely into combinations of those four buckets.

So I've been thinking really hard about just how much effort it would be to add other classes to the game. The first most obvious one would be a ranger class, which already has a handful of skills and could probably appropriate a bunch of stuff from cleric and mage.

I wouldn't just want to add one more class; if you're going to change the infrastructure enough to allow easy addition of classes, you might as well make a handful of them and get some really solid groupings going. I'd feel better about it if I had more than just ranger to work with; I think a lot of other skill sets can simply be specializations of existing classes, and don't justify a separate class.

Good examples of these would be necromancers (cleric), elemental casters (mage), and barbarians (warrior). I don't think any of these so much need a separate class as just appropriate spell and skill trees.

There's also a question of how many classes are ideal. Keeping with the exponential leveling scheme, a level 30 character should have levels of approximately 30, 25, 20, 15, 10, and 5, with levels in six classes. I'd actually expect there to be more people in the upper ranges similar to Boa, with something more akin to 33, 32, 31, 27, 22, 17, 12, 7, and 2, for a total of 9 classes. This indicates to me that even the most hardcore players would have difficulty bringing up to the top more than ten or so classes.

As far as code modification, there's a handful of specps and some other places where the code would definitely need work, but I'd like to think that once the fifth class is added, a sixth and more would be much, much simpler.

The final issue is: does this improve the game enough to justify the effort? It certainly adds some concept of end-game and more opportunities to advance high level characters, so perhaps the answer is yes. Then again, perhaps not. What do you think?

Sunday, June 14, 2009

Trading

I've been out for almost a week, so when I came back it was time to pull forward code from Locane and get some other minor things out of the way. As it turns out, he's been working on secure trading between players, and while I was away I read some articles on that same subject.

There seems to be quite a hubbub going on in the MMO world regarding trading, in particular gold farming. The general consensus seems to be that gold farming is something that should be utterly squashed wherever it crops up; further, anyone caught selling gold (or other in-game currencies) for real life cash, outside of the game environment, should be penalized as harshly as possible. Most games delete not only the accounts of the farmers, but the accounts of the purchasers as well, and occasionally the accounts of anyone they trade with.

This seems ridiculously harsh to me.

When I first started the game, I didn't really have an opinion on this topic. I originally didn't think trade would be big enough for it to really be a problem. Over the years, players began to donate to the game simply because they liked it; we (I and the other admin at the time) decided to give minor perks to those contributors as a way to say thanks. Many years later, we now offer a wide host of things for purchase from the game store.

While we were setting all this up, I got a lot of very good advice from some very bright people. The overwhelming majority of them were very clear on one point: thou shalt not sell things that unbalance the game. This includes characters, equipment, gold and a number of other nasties I have seen for sale on (less than successful) games in the past.

Players don't like the idea that someone without skill can buy their way to the top. If you take a look at our list of purchasables, you'll notice that everything on that list is either convenience, customization, or a minor boon. This policy has worked rather well for us over the years.

But back to the issue at hand, which is trading. I don't see any reasonable way that it can be completely prevented, and from my standpoint anything that's traded had to be obtained by someone somewhere. If they're going to be doing it anyway, I'd much rather put in some sort of secure mechanism for it, so people can't be ripped off.

Locane's code starts this off by handling in-game currency only. Once that's stable and working well, I'm going to add the ability for people to trade dollars, but only ones that have already been donated. Once dollars are donated to the game, you can buy donation things with them or trade them to other people. Eventually, I'll probably set up business accounts that let people withdraw those dollars from the game.

It will be interesting to see if this model works for any other MMOs. So far, I'm not aware of any that have tried it.

One major concern I have is regarding duping and generation bugs. If someone finds a way to exploit gold or certain items into existence, it may quickly become very difficult to clean up.

Sunday, June 7, 2009

Customization

After a week of forcing myself slowly, piece by piece through this code, I finally managed to boot in player customization on character create. I don't know if it's just me being in a coding funk for a week or not, but it seemed abnormally hard to get this done and installed.

The upshot is that we now have basic character customization for new players. They get to pick a handful of obvious traits, from which a generic long description is created. This will hopefully solve two problems:

1) The long description problem, which is that most players seem utterly incapable of generating proper long descriptions. Those who are capable don't want to take the time. Everyone now has a sane long by default.

2) Far more important is the issue of player investment. Specifically, new players are insufficiently invested in their character to return after their initial playing session. The hope is that customization will pay off in the long run, as the people who do take the time to customize will be more likely to return. I should be able to verify this over the next few months by tracking stats.

There's still a lot of work that needs to be done here, including ways (other than long descs) to see your features, and ways to change them. Changing them is actually going to be more interesting than I expected - as was pointed out to me, age is technically quested for via the fountain of youth, why not quest for most of the other attributes as well?

Things like changing hair length or style are clearly easy to do (got some scissors?), but other things might be interesting for players. Changing skin color for example, could be a pretty involved process.

Now that this is done, there's a whole raft of low level changes on my todo list. Lack of player investment is only one part of the problem; the remainder is that the game just isn't as fun as it should be for low levels. I have several more quests to add, as well as a couple more shopkeepers and more utility items. There are also some structural additions to be made, but those can wait for a while.

Monday, May 25, 2009

Newbie Statistics

I had a brainstorm the other day for how to pull a certain class of statistics from the various logfiles and other data we track in the game. The idea is that we grab all the new characters created in a one month time period, then see how many of them continued to log in for successive months. I decided to try this on four month time periods to filter out storage characters and other infrequently used but legitimate pfiles.

Here's a chunk of sample data, looking at how many of the players we got in January continued to log in for February, March, and April:

Total created in 2009-01: 230
Total last saved in 2009-04 or earlier: 198
Total last saved in 2009-01: 146 Average Total Levels: 8
Total last saved in 2009-02: 21 Average Total Levels: 22
Total last saved in 2009-03: 16 Average Total Levels: 28
Total last saved in 2009-04: 15 Average Total Levels: 30

Some quick observations to distill it down:

- Of the 230 characters created in January, 198 have not yet logged in for May. Only 32 of 230 (14%) have been active this month.

- Nearly all of the dropouts are in the first month, with 146/230 logins (63%) never bothering to log in again.

- Once you get past that critical first month drop out period, we lose people at a relatively steady rate.

- The first month players actually get more than one level. They get 8 on average. Even having leveled multiple times, they still don't return.

When I looked more closely at the data, I found some obvious patterns for nearly all of the people that drop out in the first month. In particular, they play the game for a reasonable period of time (15 minutes to two hours), get a handful of levels, learn a few things, then log off. About half of these never sign in again. The remainder sign in a small number of times, generally for only a couple of minutes, then sign back out.

There's a really obvious reason: The game isn't fun enough to hook newbies at low levels. They're bored, and don't know what to do when they return.

Ways to improve this:

- Add more really low level quests in the encampment. (I already have a couple ideas for this.)

- Add more quest guides for the Vemarken and other Sloe quests.

- DClient changes to make quests and other things more obvious and interesting (in progress.)

- Build the low level generic skills (cooking, skinning, etc) and make them useful. This may involve some form of low level crafting.

- Possibly give return bonuses. I believe I saw a game at one point that gave increased experience if you returned after a long absence.

This is more than enough stuff in the short term to keep me and probably several other people busy. Unfortunately, the only things I can offload onto others are the first two, and I'll probably still end up being involved at least incidentally to do cleanup and monitoring. I generally have a better idea of where the dropout points are for really low level players, because I spend more time watching them.

If I get all of this done before the end of May, I will be very surprised. It's a good short term goal though. (The return bonuses are excluded for now, as I don't know if we want to go that route. I'll have to think about that one a bit more.)

Monday, May 18, 2009

Advertising/Public Relations

Last night on a whim, I searched for the mud Dragon Realms just before I went to bed. DR is arguably one of the biggest and most successful muds of all time; they are fully commercial, with accounts going for as much as WOW accounts. They regularly top 700 concurrent users, and that's just for the one server. They have at least one additional.

An hour later, it was way past my bedtime and I must have had over 50 browser tabs open. This sort of thing is an amazing way to get ideas of where to post and link to Alter Aeon; they've already done most of the hard work, and Google has made it accessible.

Today, I began trimming tabs. I'm down to probably only 15 or so, most of which are waiting for some event to occur. I know I can repeat this process for a handful of other large games easily enough. I'm about 90% certain that the main AA pages can breach Google Page Rank 5 by the end of the year, possibly sooner.

While copying isn't a great way to be a leader, it can at least get you into second or third position. Given that we're in about 30th place right now, I'm totally up for doing some copying.

Thursday, May 14, 2009

MishMash

A quick note - I built up a Muds and Mudding lens over at Squidoo. Squidoo is one of those new things which I don't entirely understand, like I don't understand Facebook or Myspace. But other people do, so perhaps it's not such a bad idea. The url is:

http://www.squidoo.com/mudding

I've been thinking about and doing a bit of work on the client recently. One of the newer changes was the addition of an enemy hitpoint bar; the current release client has this bar between the automap and the buttons. I moved it to be in the regular prompt bar, and this serves to make the screen a lot less 'busy'.

All this has got me thinking about what to do next. I really should allow for an unhooked/floating automap, so users can position it wherever they want. I've also been thinking about having a sort of 'list window' where people can perform actions on list items, such as looking in containers, getting things from containers, etc.

Until recently, the mechanics of how to do this in the code have been somewhat questionable. As usual, talking to my girlfriend about it generated a solution: uuid strings for objects and characters, that you can always use to get an exact command reference to that object or character.

I started on the first stage of this process last night, with infrastructure. The game has had a really, really bad PRNG for years; I threw something together with basically no experience about ten years ago and we've been using it ever since. Looking at the code, I'm actually somewhat surprised it didn't cause a lot of visible trouble over the years.

We now have a brand new, 10-register lagged fibbonacci PRNG driving the game. Not only is it faster, but it actually produces reasonably high quality random numbers to drive the game mechanics. Using this, im now generating proper UUIDs that I can have some amount of confidence in.

The list window however I'm less sure about. I suspect this needs a proper design.

Tuesday, May 12, 2009

Articles and Propaganda

There's a new article out for thieves. It's a pretty comprehensive overview of the various stealth skills, such as sneak, hide, stealthy movement, searching, and even how these affect other skills such as backstab. It's a real treasure trove of information:

The Thieves Creed - Stealth and Hiding

With this addition, we've now cleared 30 official articles, not counting things like FAQs, the quest guides, or other related things. The web site is actually becoming useful to players.

The articles are continuing to show their value, both in greater and lesser extent than I had expected. I regularly refer people to the articles when discussions and Q/A happens on public channels. I really had not expected this, and I'm constantly surprised at the things people ask. A good example is the article on Time, Age, and Aging, which I refer people to regularly. Other articles, such as the article on attack strategies, are much less active, though much more comprehensive.

One other area that did not work out as I had expected has to do with the web traffic to the articles. I was expecting substantially higher numbers in regards to web traffic than what we've been seeing. I think this might be a critical mass problem however, in that traffic has been picking up recently.

In other news, the maps and quest guides have been very, very popular on the web pages. They are now the second and third heaviest hitters on the site.