Tobold's Blog
Tuesday, April 19, 2011
 
World events

There is a rule of thumb for online games saying that the population of a server is between 5 and 10 times the maximum concurrent users. In other words, even during prime time at least 80% of the players on a server are offline. Another observation for MMORPGs is that if you gather X players on the same spot, the amount of data transfered goes up with X squared, because you need to send the information of every player's location to everybody else around him. Combine these two effects, and you realize that making a "world event" at a specific time and place generally a bad idea. More players log on than the servers can handle, and players at the spot can't all see each other, or there is a lot of lag.

Blizzard gave up on such world events after the lagfest that was the opening of the gates of Ahn'qiraj. After that world events were stretched out in time and in space. Trion just learned this lesson in Rift's first world event, the River of Souls. Some players got stuck in a 5-hour queue and missed phase 2 and 3 of the event, as those only lasted for half an hour. Trion having scheduled a free trial weekend at the same time didn't help.

The solution to running such world events is at the same time obvious and impossible: You need to have server hardware that can handle a far higher percentage of players being online at the same time, and in the same zone. That is a bit like building a supermarket with 100 checkout counters when on your busiest day only 10 to 20 are needed, just in case all your customers come in at the same time. I wonder if in future with the advances of cloud computing it would be possible to have huge extra server capacity for specific world event days. Until then, world events tend to disappoint a lot of players.
Comments:
I don't really understand why it always has the be a 'boss'. Why can't we be attacked by huge army of NPCs? Much more credible and easier to handle for the servers.
 
I think most established MMOs run on custom in-house cloud-like clusters already.

The real challenge is coming up with a one-size-fits-all model. For example, I'd imagine that Sony would be very interested in having all of their MMOs run on roughly the same software and hardware despite being developed by different teams. Blizzard has probably done some groundwork already with the massive WoW clusters and their plans to open up Battle.Net to third-party developers.
 
For multi-realm games, the cheapest method for reliable events would be to host events on a different realm each week. Then you add a load of hardware to one realm for the event and switch it to the next realm the following week.
 
@Nils

Well in Rift it was technically a "huge army of NPCs" . Unfortunately the objective was not to kill the army of NPCs, but rather kill their leaders. So Phase 2 in Rift was in fact the entire "world" overrun by multiple armies of NPCs, you would literally run into a group of NPCs making a beeline for the nearest town/quest hub everywhere.

The weird part about this World Event was that normal Rift Invasions [the zone invasions occuring all the time while you play] are in fact "prevent the army of NPCs from taking over your towns!" . That's the objective, you don't fight a boss, you fight groups of NPCs and protect various towns from being overrun. If you don't fight them off, they keep coming and eventually there can easily be a 30+ NPCs standing at your quest hub requiring a raid to take them out [especially if it's elite npcs] .

Now at the end of all that, THEN a boss appear, and that's usually the finale and a very quick end to the invasion [kinda like a cherry on top, but not considered the "meat" of the invasion] . I think that works out fine, since the invasion is alot more than just a boss fight.

So what did Trion do in the World Event? They simply allowed the bosses to spawn from the start and said "even though we are unleashing an army of NPCs on the entire zone[s] , ignore them, just kill the boss . I thought that was rather brainless of them. These NPCs did not even come to aid the boss even, they just ran to a quest hub where no one give a crap what they do.

I guess the good part in all this is, i think Rift can host a good world event, all the "tools" are there, they just need to figure out a better way of executing it.
 
The problem with world events (beyond the seasonal recurring ones, which I think barely qualify as 'events') is that they are guaranteed to generate huge amounts of QQ.

The majority (vocal minority?) of WoW players seem to prefer to a have an easy life, levelling undisrupted.

This has resulted in a move from innovative things like the WoW zombie outbreak to the elemental invasion of last summer, which I felt was a massive disappointment.

How do you reconcile 'opt-in' events with a sense of danger/excitement/immersion?
 
At some point, the guys at CCP are going to get smart and end up making more money on licensing their cluster technology to other MMOs than they do on EVE.
 
You really should give Rift (the gamer's game) a proper try rather than hoping that WoW will make some sort of miraculous turnaround and not be the brain-dead button mashing spamfest that it so obviously is.
 
@koneathris -
But there is a huge difference between 100's of animated, jumping, emoting toons on screen vs what CCP has to deal with.
 
You really should give Rift (the gamer's game) a proper try

I resent the assumption that I am not able to determine whether a game is fun for me after having played it for 10 hours. And I'm actually not interested in games which aren't fun to start with, and only keep you playing with a vague promise of more fun later.

Sorry, for me Rift is just a minor variation of WoW. And Rift having exactly the same problems with world events kind of proves that point.
 
On world events, I think they are great, and it would seem rift has some growing pains and lessons to learn as WOW did. WoW has changed, and found alternative vehicles to create "world" events... Fishing contest? holiday events? Argent tournament? Tol Barad/Wintergrasp?

Dobablo... I may be "less" expensive to do what you sugest, but I would imaging there is an oportunity cost and return that is lost. Cataclysm release was a world event... what were the numbers? What was the player turnout, what was the hype? what was the media coverage? now see if you have the same impact if you do that one server at a time. So, IMO, for business reasons.. it is not the cheapest, it could be the most costly in actuallity...
 
Sorry, for me Rift is just a minor variation of WoW

That's not really the correct attitude to have when trying something new now is it? But as you shy away from heroics and don't really understand the mechanics of raiding then maybe a game such as Rift is not really for you anyway. Saying that, I'm more than happy to guide you through some of the areas that you might find challenging, our guild welcomes players of all skills and abilities.
 
@Dave

After years of playing MMO's, I've come to learn that if you are even remotely happy with the current MMO you are playing, evyou aren't going to switch. The pros will seem much more evident and the cons much less so, and vice versa for the new game you are trying.

This has been true for me in at least four or five scenarios and have read basically as much on countless blogs. The only way you are going to truly be open to a new (in actuality or simply to you) MMO is if you are dissatisfied with your current MMO or aren't playing one. Otherwise you are just going right back to where you came from.

It's not the fault of the player, but in the case of blogs (not that there is anything saying they need to be objective) I wouldn't read too much into negative reviews if they are clearly biased towards a different game.

P.S. I tried Rift this weekend, and my wife and I ended up getting copies, so in our case, we were thankful for the free weekend :)
 
This comment has been removed by a blog administrator.
 
I think Trion would do well to have their shards be decomposed into a variable number of subshards. Each subshard would have a login limit, but you could transfer to another subshard quickly and without cost. All subshards would share chat common resources like chat channels and AH.

This would also solve the opening day spike/dead servers later problem. As average load declined subshards could be removed, transfering players to one of the remaining subshards. For world events, subshards would be added.
 
So what you're saying is a month after release, Rift still has so many people interested that they have 5 hour queues?
 
This comment has been removed by a blog administrator.
 
And if I wanted brainless challenge i would continue to play WoW.

Agreed on WoW being a brainless challenge, glad I unsubscribed when I did as it sounds like it's getting worse! I got to the stage you are now at with WoW - fed up of the idiots, the shallow gameplay, every zones being too safe even for casual players such as yourself - but the switch to Rift was definitely a good move as I find the world to be engaging, dynamic and challenging enough to keep me interested. Another thing - Rift really only starts to open up after level 20 so maybe you could set yourself a target of getting past beyond that and enaging in some open world dynamic gameplay?
 
Massive amount of players is the hard part , but wow can do 100-200 pretty good in same area. Planetside could do same amount AND twitch gameplay

I am sure we are within technical ability to have 500-1000 in same area (because planetside could do it and had twitch gameplay 5 years ago) , just there is no drive to do it.

In case of rift I am not surprised they cant do it - while Rift is polished they are not technical geniuses down there, rift client and network engines was made for small instanced content. It scales worse than wow and wow is over 10 years old
 
Throwing more hardware at a rare event is not cost effective until the cloud computing really is viable. The other solution is to make the "world" of the world event bigger. i.e., if the even happened across the world. there is no reason that the players in Goldshire are on the same computer server as the players in Icecrown or Ashenvale.

I like the EVE idea and single shard but it gets more credit from the non-computer scientists that it may deserve. Eve can have 50000 players active at one time in the only "realm", but even they can't really have much over 2000 of them "near" each other at any one time.

BTW, Dave certainly seems like EVE would have been a better fit than Rift. I am not sure if you are prone to nostalgia, but a few Daves can go a long way to curing that. Although, you are a subscribe/unsubscribe player while I tend to be slog on until you quit for good.

Not a complete digression: there are Erlang-C tables to allow you to calculate how many phone lines you need to be able to handle the load 90% or 95% or 99% that involve "Peak Busy Hour" I.e., the number of calls you get in a day/week is not relevant to capacity planning, just the number you get during your busiest time and your tolerance for being overcapacity.

I submit Cataclysm has increased the peak busy hour for WoW. I.e., the guild runs and rewards mean, e.g., you don't need to log on and do a heroic but you need to log on when you can find 2-4 guildies who will do it with you. So it encourages a higher % of players to log on in peak times.
 
My rule of thumb is, if it's not crashing the server - your not trying hard enougth.

Examples in Wow include the AQ and Sunwell Isle event. Both great events, both caused server issues. Enjoyable all the same.

Compare this to Blizzards run-up-to-cataclysm events. Dry as dishwasher, no chance of crashing anything.
 
@Dave: You fell for "Fake Tobold" baiting you with troll remarks.
 
@Everyone mentioning EVE

You know what CCP's big super secret is that lets them have a thousand people in one place and have it kind-of-sort-of work while these scrub fantasy MMOs can hardly manage a hundred?

A one second game engine tick.

Yep, EVE's engine updates every one second. It can get away with that because spaceships have inertia and because pushing a button in EVE doesn't immediately start an ability, it rather initiates a "cycle".

But if a fantasy MMO were to try to have a one second engine tick it would be non-responsive laggy mess. See, with fantasy games you have the ability to turn 180% at any time whatsoever with no inertia to stop you. You have the ability to activate any one of fourty or so ability and expect that ability to "fire" more or less immediately.

So if you want more people in one place in your MMO, you are going to have to expect less freedom of movement for your character and lower expectations for character responsiveness. Unfortunately, that's the opposite of where MMOs seem to be headed what with all this MMO-FPS hybrid talk.
 
I think the idea that "well, game X isn't all that fun from level 1 to whatever, but then gets more fun" is inherently unhealthy. If a game sucks below level 20, it will never get really good above level 20.

And yes, if you look close enough, then Rift is different from WoW, just like Cataclysm is different from WotLK, and WoW patch 4.1 is different from WoW patch 4.0. That changes nothing of the fact that the FUNDAMENTALS of Rift are exactly the same as those of WoW. There are NPCs with glowing symbols over their heads telling you to kill ten foozles, then you level up and move to the next quest hub, the next zone, and occasionally a dungeon. The addition of public quests, and a different talent tree system changes nothing fundamental.

Rift has the advantage of being newer, and being less ruined yet by the theorycrafters, but those are very temporary advantages. The players are fundamentally the same, and ultimately the same bad modes of behavior, the same elitist jerks will chide you for having a slightly suboptimal soul setup or gear combination.

As long as enemies are scripted and don't have any random abilities, there will always be a "best way" to beat them, and those who found that best way (usually by looking it up) feel an urgent need to express how superior they are to the "morons" who are doing things like playing around and trying out things. It takes more than a 1% change to the WoW recipe to overcome that.
 
I'm not going to atempt to tell you what to play or how to make your decisions, Tobold, because I really don't care whether you play Game X or Game Y.

I would say, however, that it is commonplace in MMOs for gameplay in the beginner levels to be unrepresentative of the general gameplay experienced in later parts of the game, and therefore ten hours is exceedingly unlikely to give you sufficient information on which to base an accurate judgment.

As a rule of thumb, I tend to allow two weeks of play before I make my mind up on any MMO. Around 40-50 hours. Even then I obviously can't know whether or not I might like the "end game", which will probably differ significantly.

And speaking of "rules of thumb", where do you get the "5 to 10" multiple from? I seem to remember the rule of thumb as "two subscribers offline for every one online".
 
how long till someone puts an mmo on the amazon cloud or google app engine. Those kinds of places would have scalability.
 
It's an unfortunate fact that the story of the game - the events - is hindered by the out-of-game technology concerns, but it seems to be a pretty established fact. Even when the world events aren't localized and server capacity isn't an issue, the events upset one half the player base or the other because highly diversified games like WoW or Rift attract people searching for extremely varied gameplay. The pre-naxx (or was it pre-wrath?) contagious disease World Event is one example, where so many people loved the event but an equal number of people hated it.

I was disappointed by the Trion world event, too, but mostly because I completely missed it. I assumed since the first phase was 2 weeks that the 2nd and 3rd wouldn't be a matter of hours. Ah well, though. Live and learn. WoW has (whether you agree with its lessons or not, it's certainly changed the type of game it is), and Rift will, too.
 
how long till someone puts an mmo on the amazon cloud or google app engine. Those kinds of places would have scalability.


Cloud is good for scaling non dependent traffic. like website one. In a game where one agent has to communicate with all others it doesnt work because of latency and bandwidth issues , unless you go eve style , which is boring because only allows for spreadsheet style gameplay.
 
Around 40-50 hours.

But what if you don't enjoy those 40-50 hours? Isn't life too short to play 40-50 hours of any sort of gameplay you don't enjoy? And even if the "end game" is "better", isn't it still somewhat based on the gameplay you hated for two weeks?
 
@The Real Tobold...

Thanks for the heads up on the fake, I'm guessing you have no control over who and what you can filter out.

Back to your comments, I can't help feeling that unsubscribing for a while isn't going to cut it with you, there's taking a break and there's feeling so frustrated that you simply can't continue any longer. I think you fall into the latter. I guess the question now is, what and where from here? Anything on the horizon that might reinvigorate your interest? SW:TOR? GW2? Or do you think they will be more of the same? I think it's unfortunate that you haven't found more in Rift, a lot of the gameplay, particularly after I broke out of the early (<=20) levels, I found to be very satisfying as it offers experiencing that can no longer be found in WoW. I'm currently at 45 and in no rush to get to the end game, and I'm also finding many players also not rushing to 50 and enjoying the experiences that Rift has to offer.
 
@Dave

I’m about the same place in Rift as you. I honestly don’t think the game is much different from level 15 to level 45. You do quests like we’ve been doing quests for years. WoW has much more fun mini game style quests than Rift has. The only thing that changed for me at around level 25 was my class, cleric, finally felt like something that wasn’t bland.

Gordon at We Fly Spitfires commented on this a lot early on. Rift takes awhile to develop the classes, however once it does the classes become fun. The rest of the game feels exactly the same at level 15 as it does at level 45. Rifts, quests, PvP are all very similar across the entire leveling experience so far… to me at least.

Rift is fun because the community has had time to evolve into what WoW community is, but it will become that. As some people have said it’s like playing Vanilla WoW with a different skin. It feels good, it feels different, but it doesn’t feel new.

As far as how long does it take to decide if you like a game, I say within 5 minutes. I’ve been quoting Brenda Brathwaite a lot lately but she said something that really stood out in a blog post awhile back.

“Focus on second-to-second play first. Nail it. Move on to minute-to-minute, then session-to-session, then day-to-day, then month-to-month (and so on). If your second-to-second play doesn’t work, nothing else matters. Along these lines, if your day-to-day fails, no one will care about month-to-month, either”

If the first few seconds of a MMO aren’t fun how can we expect the rest of the game to be fun? How far do people on trial accounts really get? The game needs to be great right out of the gate to convert those trials into paying subs.
 
@Tobold

No, life's not in the least too short to take a couple of weeks to find out whether something is to your taste. Life is a marathon, not a sprint, as they say. Took me twenty years to learn to love olives!

Obviously, if I strongly dislike something I wouldn't persist with it for any significant amount of time, but for me at least, strong negative reactions like that are unusual. If I get as far as trying something I would expect to find some merit in it and usually that's what happens, so it takes a good while to evaluate which ones are worth pursuing further.

Of the many MMOs I've tried over the last dozen years, there aren't many that I didn't find worth giving at least a couple of weeks of my time. It has to be a pretty bad MMO not to be fun for at least a fortnight, I'd say.
 
@Epiny

I agree that if you break a game down into its constituent parts that it's very easy to come to conclusion that its 'more of the same' However, what I think sets Rift apart is the experience when those parts are brought together - Rift offers a far more rich and rewarding experience than any other MMO out there right now, especially WoW with its very shallow and repetitive gameplay (not to mention the awful community). As for giving a game 5 minutes to decide its fate, I can't see how anyone can possibly judge anything objectively within such an small timeframe.
 
This comment has been removed by the author.
 
If someone were playing a game to review it for say... massively I would expect them to put in a good 2-4 weeks.

However asking a game to "hook" you in the first 5 minutes isn't to much. MMOs hand out free trials that last anywhere from a weekend to a week. They are saying that it will take that long to get you interested in the game enough to buy it.

Therefore I think it is completely reasonable to consider 5 minutes a long enough time to decide if you want to continue playing a game. There have been a few games I've bought off Steam I've only played for 5-10 minutes and decided they weren't for me. I got 3 games with my PS3 when I bought it, 2 of which I've played for less than 30 minutes combined. If a game isn't fun right away why should you continue playing?
 
Note that I *did* play Rift for over 10 hours before deciding it wasn't for me. I think a 50-hour requirement is a bit over the top. Many single-player games these days only HAVE 10 hours of content. Would you say you need to play those through 5 times completely before you can even say whether you like them?

I think the willingness to MMORPG players to put up with many hours of crap for the promise of some purple pixels or other sort of "fun" later is one of the underlying illnesses of the genre.
 
You know what CCP's big super secret is that lets them have a thousand people in one place and have it kind-of-sort-of work while these scrub fantasy MMOs can hardly manage a hundred?

A one second game engine tick.

It certainly helps, but having hundreds of players in an FPS is not unheard of. Joint Ops had 160 players in 2004 and MAG scales up to 256, and there's an attempt to go to 1000. Does anyone have any numbers for Planetside?
 
This comment has been removed by a blog administrator.
 
Post a Comment

Links to this post:

Create a Link



<< Home
Newer›  ‹Older

  Powered by Blogger   Free Page Rank Tool