Introversion

[Chris]

Six weeks on and we’re still feeling the shockwaves from our Usability testing in Birmingham. Hilariously our project plan included a “generous” one week to incorporate any required changes based on feedback, and six weeks later I’m still busy. It’s not that the usability tests went particularly badly – in fact they ran very smoothly, and we were very pleased with the results, but they certainly identified massive problems in the understanding gamers have of Multiwinia, and that really did get me thinking.

For a long time I’ve considered usability and complexity to be opposite ends of the same scale, and this has always been a problem – you make something more “usable”, easier to understand and control, but you take away depth and complexity that make the task interesting or challenging. I’ve given this view a lot of thought since the usability testing, and I’ve come to the conclusion that it’s a load of bollocks.

The current game industry masters of usability are undoubtedly Valve software (who might very well have rekindled my love of PC gaming single-handedly, such is the quality of the Orange Box). All three games are connected in one way that is relevant to this discussion – they are all extremely easy to get into, and none of them are dumbed down. Valve have totally nailed the sweet spot of providing an intuitive control mechanism, gradual training and feedback, and fast visceral rewards for success in all three orange box games. In particular, Team Fortress 2 is an extremely complex game on paper (as far as First Person Shooters go) – nine radically different classes with 3 different weapons each, with dynamic objectives moving extremely fast – yet somehow the game is trivially easy to get into and have fun. I’ve never once wondered what sort of enemy I was fighting, or what his strengths were. Valve are actually teaching you the mechanics of the game in the very animations that move the characters around. It’s all too tempting as a player to simply think “this game is simple, there was no need for a complex control mechanism” and not appreciate what they’ve achieved, but we know different because we’ve experienced those playtests in Birmingham on early versions of Multiwinia. You’d be amazed how basic left/right mouse control can confuse the player if you don’t make it intuitive and well explained, and it’s a big step to accept that this is your fault as a game designer, not their fault as a player. It’s one of those areas of game design that is totally invisible when it’s done extremely well, and fatal when done badly.

We should be starting our beta test process in January, and this will give us yet more feedback on the controls and from that we’ll make yet more refinements. It’s telling that Multiwinia is the first game we’re planning to release that does not include a tutorial (apart from the first Darwinia demo, but we shall strike that shocking mistake from the record in the interests of our personal sanity). Our aim is to reach those levels of usability that Valve have mastered – no tutorial is required, and most of the controls are obvious to the player, even though hours of work goes into making that the case.

Fundamentally I now believe that complexity and usability are orthogonal properties, rather than ends of the same scale. We shall see how effective this new viewpoint is once people start to play the game. Soon!

Last edited by Chris on Wed Apr 09, 2008 2:47 pm; edited 1 time in total

[Feud]

Ha! I make fun of IV's Dev Journal schedule in the reverse question thread and two hours later there is a new one up!

So, if I make a joke about the beta schedule could we be playing by Thanksgiving (next week for all you Europeans, Canadians, and other non Yankee folks)?

[zach]

I feel I should reply to this, but as I only played Portal of The Orange Box (which I'd be inclined to say HAS a tutorial; the first twelve test chambers), I haven't really anything to say.

I like the thought of a graph with complexity and usability as the two-dimensional axes, that would be a bit like this post at indexed:

I'll probably reply to this again later - that is to say, "when I think of something that will make me look smarter".

[shinygerbil]

The entire of Portal was a tutorial. It was kinda disappointing, but still worth it.

/me goes to clear January's schedule

[Xocrates]

I agree with the above. The fact that the tutorial is incorporated in the gameplay is not exactly the same as saying it doesn't have one. In most games the tutorial will teach you the very basics as control and stuff, but then the single player campaign is usually a huge tutorial where you slowly learn more tactics, and how to use certain units/weapons/objects etc.

I would like to exemplify with Dawn of War. The original had a tutorial and one campaign, but out of four playable races you could only play with one in both the tutorial and campaign, meaning that for a very long time that would be the only race you would be playing while slowly experimenting with the other ones (which meant getting stuck reading the tooltips/manual for every freaking unit, building, and upgrade the game had). Fortunately they fixed that in the first expansion.

So, long story short, feel free not to have a "tutorial", but if you're going to do so then please add a single player campaign.


I should probably add that from the orange box I also only played portal, so I'm unaware if TF2 has a single player mode or something. But one should probably point out that FPS have very standard controls so I doubt a big tutorial would be needed.

[jelco]

Chris wrote:

The current game industry masters of usability are undoubtedly Valve software (who might very well have rekindled my love of PC gaming single-handedly, such is the quality of the Orange Box).

I'm not so sure about this. Valve has shown to make great games, but I wouldn't say they're top of the bill. I can't say who is, but it just doesn't feel right to give Valve this credit.

Chris wrote:

In particular, Team Fortress 2 is an extremely complex game on paper (as far as First Person Shooters go) – nine radically different classes with 3 different weapons each, with dynamic objectives moving extremely fast – yet somehow the game is trivially easy to get into and have fun. I’ve never once wondered what sort of enemy I was fighting, or what his strengths were. Valve are actually teaching you the mechanics of the game in the very animations that move the characters around.

TF2 certainly is an excellent example of how a game can be easy to learn without something like a tutorial. In fact, the game doesn't even have single player, and you purely learn by randomly selecting a server, getting killed a few times, and then it suddenly all dawns on you. The game is incredibly simple to learn, and is fun to play when killing time. However, the games can also get incredibly good if the right players are put together, and once again these people learn everything purely from experience.

Chris wrote:

You’d be amazed how basic left/right mouse control can confuse the player if you don’t make it intuitive and well explained, and it’s a big step to accept that this is your fault as a game designer, not their fault as a player. It’s one of those areas of game design that is totally invisible when it’s done extremely well, and fatal when done badly.

I can imagine this. It's kind of like the work of companies like the CIA: their failures attract publicity, but you never hear of their successes. With basic features like this it probably works the same - people only notice it's there when it doesn't function properly.

To people outside the developer's business - including myself - it's probably hard to understand the complexity of creating a feature that has become so common that every games thinks of it as standard. However, I have experienced similar problems with my own programming (in PHP). For example, the ladder for DEFCON I created had a slight flaw: the same match could be entered twice. So I was asked to implement a duplicate check to prevent the same match from being entered multiple times (since it's nearly impossible to have two games with exactly the same outcome). However, the dynamic nature of the system makes this much more complex than you'd think: the order of the players can differ per input (say you first enter a match by entering the scores for A, B and C, and then another enters the match with another order like C, A, B). This makes the problem so much more complex that I didn't want to make it at first. Eventually it turned out to be a ten-minute job, but it was quite difficult nonetheless.

Since I've had that problem, I have learnt to watch everything from a programmer's perspective (as far as my knowledge allows me to). I've learnt quite a bit about programming - I've recently started learning C#, and am quite experienced with PHP already - and it's become easy for me to imagine what a function would look like in code. It's been a real revelation to me.

Chris wrote:

We should be starting our beta test process in January, and this will give us yet more feedback on the controls and from that we’ll make yet more refinements. It’s telling that Multiwinia is the first game we’re planning to release that does not include a tutorial (apart from the first Darwinia demo, but we shall strike that shocking mistake from the record in the interests of our personal sanity). Our aim is to reach those levels of usability that Valve have mastered – no tutorial is required, and most of the controls are obvious to the player, even though hours of work goes into making that the case.

The key to such usability is intuitive controls, as you said yourself. Darwinia was very special with it's input method in the task manager (the drawing technique) - and the difficulty players experienced resulted in a possibility to switch to a simple key-command method. That, of course, needs something like a tutorial to teach the player how to use this system. A classic RTS model - like the CnC series - simply uses a building to produce units, and you can give them commands with simple mouse-clicks. This is easy to learn (if you don't already know how it works from older games) because the player can just go around clicking stuff and see what happens. Darwinia didn't allow such freedom.

However, a game which uses these classic methods loses depth (as you put it). After a comparison with the Orange release, it would seem that you can use clever game mechanics while still working fairly intuitively. Do realize though, that an FPS is fundamentally different from an RTS. In a first-person game you directly control a character with every single key-press or mouse-click. In a strategy game, you first have to choose which units are going to get orders, and then give the orders themselves. The result is that a first-person game allows for loads of experimenting (because you immediately see what action is associated with a command) while a strategy game requires much more guidance to show the gamer how to select the units to be controlled in the first place. I think this is a very important difference, and it requires much more knowledge (and the ability to imagine how a player uses the game) to be able to create a game that works this way.

Chris wrote:

We shall see how effective this new viewpoint is once people start to play the game. Soon!

If I get the chance (and if I have time) I will definitly help you support this new viewpoint once the playtests start!

Jelco

[KingAl]

jelco the galactaboy wrote:

Chris wrote: The current game industry masters of usability are undoubtedly Valve software (who might very well have rekindled my love of PC gaming single-handedly, such is the quality of the Orange Box). I'm not so sure about this. Valve has shown to make great games, but I wouldn't say they're top of the bill. I can't say who is, but it just doesn't feel right to give Valve this credit.

Purely framing it from the perspective of usability, though? Considering the Orange Box both presents potentially complex concepts and combines it with great intuitiveness, I think it's fair to say. Contrast it with S.T.A.L.K.E.R., or even BioShock - wherein rather than working on elucidating concepts better, they merely simplified the concepts themselves (if one makes the fairly easy assumption that System Shock 2 was a starting point design-wise).

[Rkiver]

Beta for January, should give me something to do when I am made redundant.

[frenchfrog]

Xocrates wrote:

I agree with the above. The fact that the tutorial is incorporated in the gameplay is not exactly the same as saying it doesn't have one. In most games the tutorial will teach you the very basics as control and stuff, but then the single player campaign is usually a huge tutorial where you slowly learn more tactics, and how to use certain units/weapons/objects etc.

I just began playing Half-Life 2 Episode 2 (and I just finished Episode 1 before that) and it's a big tutorial on all aspects. It's very well done in 3 aspects, it's gradual, use explanation by example instead of by description, and ask for player creativity/intuition. Let me details them:

Gradual: Start with the basics like jump, crouch, use of basic _weapons_( gravity gun ). For the gravity gun in itself you learn to use it to destroy thing blocking your way, throw stuff at people. Then come the _puzzles_, you get shown basic gravity, lever effect, weight tricks, ... Same goes for the latter puzzles, always a bit harder then the last one, one more _trick_ is needed to solve it.

Explanation by example: First they use simple puzzles to teach you the basic concepts. But more importantly you are often shown a working or part of a working _machine_/solved puzzle before you need to solve a not-functioning one. (ex: They show you that gaz barrels can explode, you are shown a working computer with a connected power cord before needing to reconnect a power cord, you are shown that some _webs_ can be open by shooting into them, that valves can be opened, ...). And it sometime leads you to a very interesting question: 'How do they that?' How do I do that myself!

Creativity/Intuition: By the 2 previous items all puzzle are _explained_ to you, you then just need to put the parts together to solve the puzzle. I think this is the fun part, this is where your creativity is called for, the job is not done for you but at the same time you have all the needed tools.



I think this also have a really nice side effect: you are much more attentive to the world around you since you know there is always something to _discover_.



At time, it gets a bit tedious, greater subtlety would have been great. It's like in the movies when you see a knife/car/escape route at the every beginning and you say to yourself: "Ah! The hero will surely use this at the end of the movie to win over the bad guy". The other thing missing in this scheme is that difficulty rating don't apply to _puzzles_. It only apply to _beast health_ which is only a small part of the whole Half-Life 2 experience.

[shinygerbil]

frenchfrog wrote:

It's like in the movies when you see a knife/car/escape route at the every beginning and you say to yourself: "Ah! The hero will surely use this at the end of the movie to win over the bad guy".

I agree wholeheartedly. I wish games would move away, but in many cases it's more a case of 'could' - for a game like Half-Life 2, to make such things as, say, an object which is crucial in solving the puzzle actually subtly blend in to the suroundings, it is then a much more complicated task to design a level, introducing many unnecessary items and scenery which would only clutter up the game.

At least we've sort of gotten away from the stage in games where items/objects which were static were very clearly and visually part of the 'background scenery' model, while objects with which you could interact were often lit using a different light source, used a radically different colour palette, and just generally stuck out like a sore thumb. That always used to make me so very mad.

[Feud]

shinygerbil wrote:

At least we've sort of gotten away from the stage in games where items/objects which were static were very clearly and visually part of the 'background scenery' model, while objects with which you could interact were often lit using a different light source, used a radically different colour palette, and just generally stuck out like a sore thumb. That always used to make me so very mad.

I always found this funny when watching cartoons as a kid. I didn't understand then how they were made (with cells and layers and all that) so I kept wondering why the backgrounds were so different from everything else.

[byrneda]

Well.
There you go.
The reality that just because someone says one thing, someone (or something) comes along and blows it away.
Way to go Chris. I likes what I read.

[martin]

I hope you realise, chris, just how evil it is to put up a new picture on the front page without a link behind it? ¬¬

I soooo want to play multiwinia without being reminded every time I come to the IV forums!

[jelco]

You know, I always go to these forums directly, so I hadn't seen that new picture before you told me about it. It has cleared up one thing though: brice has won the naming contest so far.

Jelco

PS: To be honest, I thought I had posted that name myself. However, I was a little off.

[KingAl]

I had a sneaking suspicion that the name 'Multiwinia' would prevail (assuming that's what this indicates...). I'd update the Wikipedia page myself but hotel-internet-on-TV is just downright painful, so - hop to it, people!