It’s been nearly a full month since my last post. I had to go to Paris for a meeting and I got the swine on my flight back to Montreal. Nailed to my bed for four days and out from work for eight was enough to bring my brain to new levels of weirdness where old thoughts from Far Cry 2, new ones from my new project and a few others for a future talk of mine were all mixed up together. So I now have to put some of those in here in order to think clearly again.
So today I decided to talk a bit about improvisational play which is from ‘my old thoughts from Far Cry 2’ brain section. As presented by Clint Hocking at GDC 2009, improvisational play was observed in Far cry 2 as a result of our inability to provide a game structure where the composition phase is balanced with the execution phase. In short, as the game was slowly getting together we realized that during combat, players were often confronted to step out of the execution phase as a result of an unpredictable event. This forced them to readjust their direction by going immediately back to the composition phase where a new plan was required. The fact that this was happening several time during a single battle clearly demonstrated that players were often improvising on the fly instead of executing their plan as intended. They were not dominating the situation through intentional play; they had to constantly readjust to combat events.
There is no doubt improvisational play is a positive design side effect of the Far Cry 2 experience. As Clint said, we were embracing what the game was trying to be and I would say with a bit of luck it ended up being a good thing. But it was not design that way so there’s still a dark side to it.
The amount of FC2 players who actually understood what was going on when they experience improvisation is quite small in proportion. This means countless players are asked to readjust on the fly without understanding why. Clint mentioned in his talk that improvisational play can be experienced without mastery which is true but it will not be appreciated if it is not understood.
What do I mean by understanding it? I am talking about the ability to identify the cause behind an execution phase break. In FC2, improvisation can be caused by several systems which are very hard to predict. Malaria, weapon jams and wounding are those Clint used in his presentation so I will stick with them. Each of those systems requires constant attention from the player. Malaria attacks are predictable but in an unexposed way and the player can make them go away by getting more pills through side missions. Weapon jams are all about weapon management. Weapon shops and safe houses are locations where it is possible to grab new weapons in order to avoid going into a fight with rusty unreliable guns. Wounds are quite predictable but the speeds at which the health can deplete make it hard to fully anticipate. There are a few criterias here that make those systems efficient at generating improvisational play.
First of all, they are negative feedbacks that will confront the player with new challenging reality on the spot. Secondly, they can be avoided which means the player ultimately has the responsibility of his own situation. Finally, they're designed to be uncertain so players constantly fear them and feel the need to take them into account in their composition phase (planning). They are working really well together and are responsible for magical moments in Far Cry 2... If understood properly. The FC2 design team had a lot of fun playing the game because they understood the mistakes they were making and how it was forging their experience. For example, it happens to me quite a few times that my gun jammed while I was in perfect control of the situation and I often smiled while saying to myself: “Ho Shit, I knew I should have passed by the weapon shop before coming here”. It is really cool to fully understand why you are in a given situation and it is also essential to be comfortable in the improvisation mode.
Here is something from punch-out to illustrate my point a bit better. Punch-out uses simple timing variation to challenge the brain. For example, the opponent AI will crouch exactly the same way for the right and the left uppercut but the right uppercut is launched 3 times faster than the left one. Often players will react too fast to the left punch and get back in the danger zone as the punch is launched. At this very moment the player says to himself: “Ho shit I knew it was the left one why the hell did I dodge so fast?” and at this very moment he dodge a second time and barely manage to avoid the real hit. This moment is exciting because the player adapted to a situation he understood which made him feels pretty good.
Now it is easy to say that Punch-out is fully deterministic and therefore is a bad example for improvisation but that would be wrong. When our brain is asked to react at the split-second level it is pretty similar to improvisation. Also, in the punch-out example the player is not aware of what will come next and he makes a mistake but manages to fix it based on his understanding of the situation. I believe this is where the true fun of improvisational play lies. FC2 explores improvisation in much deeper ways but it also fails to teach the player properly the crucial rules. Weapon’s reliability is shown through the gun shaders, the sound and some tutorial explains the importance of it and how to keep relatively new guns at all time. So we did try but it was not fully embedded in the progression of the game. Still to this day, I regularly confirm with various FC2 players that most people don’t fully get the reliability rules. When I talk to them about the weapon jam they see it as a frustration generator more than anything else.
Improvisational play remains heavily intentional because players still make risk analysis unconsciously leading them to decisions that will define the probability of an eventual improvisation moment. It is crucial that each variable relating to these risk analysis are properly exposed to players so they can perceive what is going on and appreciate the level of improvisation required.
Improvisational play was always part of our reality. In a way, Clint’s observation and his appreciation of this design structure is the same as Ralph Koster conclusion in A Theory of Fun when he mention the need to provide more complex patterns to players to avoid domination on their side which leads to boredom. Whether it’s the organic pattern behind the four ghosts in Pac-man, a strong opponent in street fighter, the uncertainty behind poker probabilities or a music improvisation, it is always based on the level of understanding of its participants. So I guess all of this is a very long way to say: “There is a higher level of play in Far Cry 2 heavily driven by improvisation. Unfortunately, this level is profoundly hidden behind inaccessibility…” There are definitly players who manage to reach it and enjoy the hell out of it. Now it is our job to make sure there are more of these guys for the next games we'll ship...
For extra information on the excellent talk about improvisation by Clint Hocking, look here on his blog.
Jonathan
Nice post. In my personal experience with Far Cry 2, I was always one to, for example, head to an arms dealer and switch out my weapons before embarking on a mission. As such, I almost never ran into the gun-jam problem.
Similarly, after purchasing the camo-suit and silenced weapons, I found I could almost eliminate the wounding problems as well (almost) by staying back, picking off soldiers one by one, and by regaining health at arms dealers between any skirmishes. Being a stealth player, it seems, kind of eliminates the improvisational aspect of the game (to an extent). I'm not complaining, but it is interesting to ponder upon.
Posted by: JPLC | 09/14/2009 at 12:50 AM
Prep a mission through weapon and supply check, and recon the mission objective's location and potential threat should have get most of the uncertainty out of the way, and that is including the road patrols and studying the pattern of behaviour of tangos on the location.
Posted by: tourist.tam | 09/14/2009 at 08:19 PM
This is a really interesting twist on something I've noticed as well; the signal to noise ratio when communicating any complex (second order) game system. I've not yet played FC2 and from the sounds of it, it has a tendency to work at a fairly high cognitive level which, from the sounds of it, makes it a microcosm for the analysis of this concept. The player metrics for this game must be pretty fascinating to look at! How was this functionality intended to be communicated to the player Jonathan if you don't mind me asking?
It seems that the threshold for when signal (the way a system explains itself or is explained by the designer through a given process) becomes too noisy to be clearly understood (the ratio becomes unacceptably unbalanced and requires the player to either reject the system or continue using it with a much higher error rate) is something we often tend to base on insider knowledge - our familiarity with the system from its inception. It also sounds like this becomes a much more pronounced usability problem when the system concerned isn't directly user facing, but is instead a second order event (a property of a user facing object).
Posted by: Martin Herink | 09/18/2009 at 07:47 AM
To the first two comments I simply want to specify something. It is true that uncertainty can be significantly reduced through clever long distance play or good use of stealth upgrades but it will not get rid of it. There's always a ratio for the gun to jam anyway and even if a given fight goes as plan, you might have less ammo than predicted or something like that and you will have to change your mid level plan. It is more and more obvious as you crank the difficulty level up but it is still there at normal. It is unlikely to happen, but never the less possible.
Martin: We intended to teach the system by altering the weapon's shaders and make them rustier as its reliability goes down. We also have a pretty bad text tutorial early in the game, the kind that most players pass pretty fast without reading everything. Everything is there, but it wasn’t good enough. In the end when managed properly, the system is not too much of a challenge but it does alter the way you plan where you go when and that was the interesting part for the level design. This renders navigation choices meaningful which is good for the open world structure. Unfortunately it was not properly explained and it was also a bit tedious for some players. It is pretty fun for me to play it at higher difficulty level and feel the impact on my mid level game. But it is definitely implemented in a very hardcore way which I am guilty to love but clearly isn't the best way to go. We now know better for the future :).
Posted by: Jonathan Morin | 10/04/2009 at 09:00 PM
The 'Improvisational gameplay' as you put it was not a problem for me at all - I thought it was excellent, and it added to the immersion. The problem with FC2 was that there were so many things that broke the immersion and nearly ruined the game; for example - the respawning guards at the checkpoints.
Posted by: Benjamin Thatcher | 10/21/2009 at 06:56 PM
Benjamin: I am not saying improvisational play is a problem. It was simply poorly supported because so many players didn't fully understand the possibilities and the mistakes they were doing. For them it was like a musician who improvised without any comprehension of the notes he uses. Your example of the checkpoints is an immersion breaker for you as a player (and probably 80% of players out there) but it is equally problematic to not understand why your gun jammed and what you did wrong.
There are many problem in Far Cry 2 as there are in most game. Lack of AI persistence is one of them and lack of system understanding is another. Several systems and rules are better off unexplained in games, but there were quite a few systems/rules in Far Cry 2 that should have been clearer...
Jon
Posted by: Jonathan Morin | 10/23/2009 at 09:10 PM
The main disappointment for me in Far Cry 2 was some missing functionality which was clearly supposed to be there but doesn't work, like using the flare gun to attract guard patrols.
Having said that, Far Cry 2…What a great game! Any plans for a sequel or expansion? Or maybe a proper SDK?
Posted by: J. Santo | 11/06/2009 at 03:39 PM
J.Santo - You know I cannot answer those questions :). It's true that Flares had bigger plans and we didn't manage to make them happen. The one flare idea I wish would have made it is the reinforcement system.
The idea was to first of all have AI uses some kind of launch platform to throw flares in the air which would alert reinforcement troops in the field. If the player managed to prevent AI from going there they would had prevented the call. Also, we wanted to have two flare types (one per faction)making it possible to use your flare gun to call enemies from the opposite faction which would have caused some cool fights out there.
That was one of those thing we lacked the time to make. Far Cry 2 was a massive game and they were quite a few things like that...
Jon
Posted by: Jonathan Morin | 11/07/2009 at 08:46 AM
Understood Jon—thanks for clearing that up! Far Cry 2 was one of the most memorable gaming experiences I've had. The AI, the animation and the also the graphics engine and sound were all fabulous. I am looking forward to whatever comes next!
J
Posted by: J. Santo | 11/08/2009 at 12:02 PM