What Went Right

Very interesting. I'd love to know more about the sample group.

Fascinating stuff. There really is no substitute for hard data.

One thing I wondered about was the use throughout the article of the word "engagement." This term was used most of the time to refer to physical engagement, to grabbing a player's attention through on-screen action content that stimulates physical excitement responses. The one exception to this usage referred to Half-Life 2's puzzles as an "emotional break from action" that generated their own distinct form of positive response. (Actually, calling HL2's puzzles an "emotional" break seemed to me to be a rather odd description for puzzle play, which I see as being more about achieving intellectual satisfaction than as gameplay for generating and relieving emotional tension.)

Other than that usage, however, "engagement" usually seemed to mean just adrenaline production. For the shooters being studied, that makes sense. But what about other kinds of games? Do other kinds of games produce less engagement? Or can they produce equal levels of different kinds of engagement?

I'd enjoy seeing the results of research by EmSense into those questions. In addition to the physical engagement typical of the shooters studied here, aren't there computer games that elicit intellectual engagement (such as 4X games) or emotional engagement (such as Petz or Façade)? Given that "cognition" is already said to be measured in some way, I'd think EmSense might be able to evaluate intellectual engagement... but what's the proper sensory metric for detecting whether emotional engagement is high or low?

It would be interesting to see what makes one game that focuses on a particular style of play more engaging than another such game. For example, could EmSense's data-collection methodology explain why Master of Orion 2's gameplay was popular while that of MoO3 was not?

Perhaps even more interesting would be to see whether games that hit on all three cylinders are typically more engaging to more gamers than games that provide just one mode of fun but do it well.

Finally, what about the possibility that some gamers -- whether innately or through experience -- are wired to enjoy one form of play-stimulus more than other forms? Would EmSense's research provide support for the theory that there is a biological basis for different psychologies of play?

I hope we can look forward to reading more research results from EmSense.

I dont disagree with this article and in fact I find it very interesting. However in regards to the close combat, I believe a big part of the excitement associated has to do with the simple strength of the weapons. Gears of War, Halo, and Call of Duty 4 all have melee weapons that instantly kill the targets. If that power was downed, I think the weapons would be a lot less exciting to use. Likewise, I believe that other powerful weapons such as Rocket Launcher probably have similar excitement ratings, and if they didnt I would suprised. (That also would be a pretty interesting piece of information.) I can also name a few experiences where I didnt have fun with the close combat weapons (notably in Gears). I dont believe that the study is flawed, but I dont think close combat itself is inherently engaging.

Really though, interesting research.

Maybe many people get nervous when they use melee weapons, because the probability to die is much higher and you often loose control in melee. I think in many fast action oriented games the control scheme for close combat is not very satisfiying and most fps players know that...

Another large attribute to close combat being a high intensity situation is that the view pane is often obscured with the enemy so close so that provides a natural feeling of claustrophobia, in the same way generated by level paths that move from the open into dark or enclosed areas where sight lines are obscured and limited.

John - melee attacks could be downed if the audio visual payoff were still juicy enough, and strategically viable enough to use. Doesn't just have to be about damage. All you're getting with massive damage is an immediate, distinct, noticeable, strategically useful payoff.

Melee systems are clearly horrible when you're just two blokes running in circles, prodding each other with spoons, which is what happens when you nerf their damage. However, damage isn't the only affect you could have from the attacks. Add some juice some other way.

I'm curious to see how this is applied to games where the obvious lower brain stem fight-or-flee mechanism isn't at the core.. like puzzles or adventures..

I'm wondering if you could use this technology to graph more subtle sensations such as confusion, delight, surprise, frustration against a timeline for a game without a core fighting mechanic.

I simply have to point out that you consistently talk about engagement, and in the final paragraph of "Keys to Engagement" seem to equate this engagement to fun.

My point lies in game mechanics such as Machine Gun Turrets and Sniper Rifles.

It's mentioned that Call of Duty 3 allows a player to distance themselves, and be protected, while using the sniper rifle. While this is true, I personally often find Sniper Rifles completely useless in most games because not enough cover is provided. The Sniper Rifle is one of my favorite weapons, because it's tactical and hands-off - I can take my enemy down without him even knowing I'm watching him.

Machine Gun Turrets can be equally frustrating. While it's more engaging to have a small turret in the open where the player can be hurt by using it, being unable to use these weapons because it's too likely to cause failure turns them into a tease. The problem is worse with turrets that have warm-up times before they'll fire.

My point is that, while some experiences may be more engaging, they can also be more frustrating. They cause the gamer to pay more attention, but sometimes piss them off. Special care must be taken to provide fun -and- engaging play.

I'd be interested in seeing more data on the conclusions drawn about close combat. While I think it's clear that the payoff from a melee hit (massive damage, visual/audio rewards, and often a completion of the player's goal to kill the target) does cause an increase in engagement I think the article might possibly be missing out on a piece of the puzzle.

Does all close combat cause an increase in engagement or is it simply melee? The way the article is written would first have you believe all close combat is engaging, even close combat where you fire your gun instead of melee. However the article then goes on to focus on close combat that is centered around melee. Since the article doesn't really provide a distinction between the two in terms of data it's left me wondering.

My 'instinct' (hardly a match for hard data I know) would be that any close combat, melee or with a gun, would cause an increase in engagement. When a player is engaged in close combat the perceived threat is likely to be more than when engaging a foe from afar. On top of that the player now has to actually be more focused to actually succeed. In close quarters combat the player is at risk of their target quickly getting out of their FoV or aim which adds to the danger and excitement of the encounter.

I'd really be curious to see what the Aggregate response of an encounter at close quarters without melee is compared to one where a player does have a strong melee option. My estimate would be that simply being in close quarters with a high level of perceived danger would be enough to cause a large amount of engagement and the presence of a rewarding melee option (like the lancer) simply builds on top of that engagement.

@Fazoum Pippilnaburugh

Absolutely you can. Psychology research has been studying these things for decades now. Further, research in Augmented Cognition is looking at the use of physiological measures to dynamically change simulation parameters (reduce/increase workload based on perceived stress, frustration, etc.)

There is still a lot of work to be done, in so far as identifying the unique physiological signs for a particular feeling. Does this increase in heart rate coupled with decreased breathing mean excitement or frustration? What pattern can be seen in brain alpha and beta waves? These are questions that are complicated and very likely context dependent. But researchers are looking into them.

In particular, you might be interested in what is known as a breaks in presence. Substitute presence for the "immersion" or "engagement" terms often used in gaming, as presence research encapsulates both. Basically, a break in presence is the point where the person realizes that they are no longer in the virtual world.

See the following paper if you want to learn more.

Dustin B. Chertoff, Sae Lynne Schatz, Rudy McDaniel, Clint A. Bowers: Improving Presence Theory Through Experiential Design. Presence 17(4): 405-413 (2008)

http://grassrootsgamemaster.blogspot.com/2007/03/dead-game-designers-society.html

It would be interesting to elaborate on close combat/excitement aspect on games that don't just encourage it but that actively make it a part of the core gameplay experience or where it's something more than merely a high damage/high risk weapon and/or where close combat is emphasized over medium range and gadgetry. Games like Left 4 Dead and Mirror's Edge for instance.

In the former, melee is actually pretty ineffective barring a sneak attack. However, it provides the very useful ability to push away zombies in a wide swath. The combat itself is, for the most part, heavily focused on very short ranges. In the latter, close combat is emphasized only as a way to avoid combat - disarm an opponent and then run.

Boredom, anticipation, fear, engagement, relief. Rinse a repeat.

For narrative designers, this kind of research is invaluable. In particular, it shows that plotting out the intensity curve, which narrative designers do, is essential to player engagement.

Excellent and fascinating article...even if the results don't seem very surprising! We've known for years that the most engaging moments in games are when the outcome is unknown - particularly if it's at the end of the game. Engagement from a weapon or vehicle that makes the player over-powered and highly defended versus one that is not demonstrates this clearly. The difference here is actually being able to demonstrate it in a lab - and actually providing a checklist that we can apply to other games.

Let's take a 2D shmup like Metal Slug. We can learn something from its design using this formula. Metal Slug (the whole series, but I'll be talking about the first game in particular) has high-powered weapons that increase engagement by giving the player considerably greater firepower (and sometimes, like the flame thrower, producing interesting effects when used). Everyone's favorite weapon, the Shotgun, is also a close-combat weapon and has VERY destructive firepower, but also slow rate of fire (Metal Slug also has the combat knife for close-range). Unsurprisingly, grenades are also fun to use (and therefore engaging) due to their high firepower, splash damage, and spammability. Additionally, simply getting a high-powered weapon or lots of grenades does not guarantee success, as the player may die in one hit at any time and lose the weapon.

However, there is also tactics involved with weapons use, something that can be whittled down to player choice. For instance, in Eternal Darkness (a 3d action game), I have a choice between chopping off the zombie's head or severing its limbs; I can also use projectile weapons or melee weapons. In Splinter Cell, I can choose which route to take through the level. This lets players develop a strategy that is most effective by giving them meaningful choices.

In Metal Slug, the player gets to choose which weapon to take down the enemy with - the flame thrower has limited range, but it can penetrate riot shields; the heavy machinegun fires rapidly and at a diagonal, but has comparatively lower damage. You can also destroy the tanks with spike shields by throwing a grenade onto the back or by taking out the shield with a lot of firepower. Finally, in Metal Slug 3, I can also choose which path to take through the levels, adding replay and engagement through meaningful choice.

Metal Slug also has turrets and vehicles. Turrets have high firepower, but they provide no protection and turn the player into a stationary target. Driving the tank gives the player an extra level of defense and greater firepower, but ultimately does not guarantee success, as the tank can only absorb three hits before it is destroyed, and enemy tanks and boss vehicles have tough armor. Finally, the player is rewarded for surviving a level without getting killed through prisoner bonuses and also keeping the tank until the end.

The use of death animations is also important for reward: you want to provide the player with something fun and interesting beyond the simple goal of killing something. Sprays of blood and debris (both present in Metal Slug - and in vast profusion) adds to this by creating spectacle. Death animations in Metal Slug also contain a unique blend of slapstick, and the variety of animations and how they are triggered means soldiers don't always die the same way - so I would add variety to engagement as well. Ragdoll physics tend to do this as well and can create humor when a large number of enemies is annihilated at once by an explosion.

Regarding pacing, the most obvious element is minibosses and end bosses. Bosses are inherently difficult and take many bullets to defeat, producing a more frenetic battle. Additionally, Metal Slug creates screens full of enemies (such as the dozens of soldiers coming out of the caves in the final level), each providing its own challenge. Metal Slug 3's final stage is one great roller coaster, taking the player from jetpacks and airplanes to rocket ships and an alien spaceship. It also provides nice breathers between sections. The unique thing to learn from here is that each screen tends to have something different inside it, creating clear progression and variety. While most FPS games don't have bosses, they do contain rushes of enemies or several tough enemies that help keep the outcome in doubt (a boss fight is inherently doubtful simply because the enemy is so big and powerful).

While a 2D shmup may not be as popular a genre as the FPS, they clearly have similarities and applicable designs. Doug Church says that for design to be a practice, we should be able to take something from a 2D shmup and apply it to the baseball game we're making. Shoot to Thrill provides just that kind of formal abstract design tool that we so desperately need.