[Music] hi everyone welcome to evolving god of worst combat for a new perspective thanks for coming by on the last day of gdc after what was i'm sure an exciting but very tiring week as a quick show of hands who here played god of war oh [ __ ] all right well that's cool so who here played on the give me god of war difficulty i apologize for your pain so my name is mihir chef and i'm the lead combat designer here at santa monica studio joining me during the q a will be our gameplay director and former gameplay engineering lead g trough together we worked alongside a host of talented gameplay designers and engineers in evolving our combat systems on 2018's god of war though i'm the one up here presenting today everything i'm about to go through is the culmination of a team's effort we've got a lot to cover so let's get started for those unfamiliar with past and present god of war here's a small video to give context to the evolution and combat that took place god of war is a third person action-adventure game series where you play as the spartan demigod kratos and use your legendary weaponry to overcome obstacles in a visceral delight clearly the two games shown look very different but the gameplay design underneath both of them shares common dna to talk about combat evolution we first need to look at the past what was the traditional core of god of war's combat for over a decade the series combat has focused on a few central pillars fundamentally we wanted to adapt the intensity and finesse of japanese action games to the west in doing so we aim to expand the audience by increasing the accessibility of gameplay and bringing over-the-top bombastic presentation to the forefront kratos was a character all about offense and the power fantasy of our combat represented that by making fighting your enemies feel like playing with your food and of course all this was done from a pulled back game controlled camera that framed all of the action in a way that was larger than life each game in the series built upon the foundations of the original by refining combat mechanics and escalating spectacle to epic proportions with this new god of war however the overall vision demanded a more grounded and intimate tone this applied to all aspects of the game including combat and there was so much new with this vision the norse mythology a new million range weapon and leviathan axe and a new player controlled close camera that never cut and of course i'm skipping over a brand new companion if you want to learn more about that and missed hayato yoshidomi's talk yesterday check out his talk raising a trace for battle when it hits the gdc vault so the idea behind the new camera was to create a more visceral experience that brought you closer to kratos and to make the combat more deliberate and unflinching this would fundamentally change the combat experience of the franchise and it had to be better well that meant two very clear things we knew we couldn't refine and iterate on the combat formula that existed like previous titles had done instead we would have to re-imagine combat for this new style of game and it would have to be largely governed by the camera and in this new style we also wanted to retain the core identity of what fans know and love about god of war's combat and these goals drove us insane right like at the time these two things were completely at odds with each other fulfill an offensive power fantasy with the fast intense action but for a more grounded game with a super close camera and it was really funny seeing people's response when we eventually showed the game for the first time at e3 you know people thankfully responded really positively overall but many kept saying uh i don't know they should have pulled the camera back you can't really see anything the internet was experiencing what we agonized over for years but let's roll back a bit to understand what i mean by that to know why these two goals were a problem we have to dive in a bit deeper into the tangible elements which define the combat identity of god of war now obviously you're doing a lot of things while playing the game but i'm focusing on what you're doing most of the time which is fighting basic grunt enemies such as the dragger at the heart of god of war's combat design for moment-to-moment gameplay lies four tenants responsive and combo oriented gameplay leads to varied and exaggerated enemy hit reactions satisfying and powerful attacks move you forward physically through the space kratos can handle and often excels when he fights multiple enemies and finally game plan controls are accessible hassle-free and are on the player's side now with that in mind most other action games with similar types of camera perspectives are designed to focus primarily on one versus one combat with a lock-on system not on fluidly fighting multiple enemies together they tend to have combat loops which promote relatively defensive and cautious player behavior compared to traditional action games as it's harder to see your surroundings the player's attacks typically don't translate them forward through the space in order to help with the camera framing and it keeps combat aimed at gauging the distance and timing of attacks as opposed to manipulating fun enemy hit reactions and this is all for good reason right players love these games and these types of combat systems make sense for a closer camera since the combat is designed around it unfortunately this was not the case for us and our combat identity we needed to stick to our special ingredients if we wanted to deliver a different but true god of war experience we needed players of all skill levels to be confident and offensive and it didn't seem like this was possible with the proposed change in fact in the beginning many of us on the combat team were really opposed to the new camera for this very reason but the vision was set and after many initial months of squirming and fighting about it we trusted in the vision and committed to the thought that we were going to make it work it wasn't easy her early attempts of making the camera mesh together with combat did not go well as you might imagine frankly speaking it was terrible to play and it was almost impossible for players to make sense of what was going on around them i mean can you guys tell what the enemies are doing in the video because i sure can players just panicked evaded around constantly and mash buttons without a semblance of strategy because the combat systems were fighting the camera in a way that made game play extremely clunky to solve this we had to re-evaluate our combat late into the fall of 2016 and go forward with the strategy for players to play confidently and offensively we needed to remove as much clunkiness as we could after the initial adjustment of getting used to the game the difficulty needed to come from the game itself and not the controls in camera if we didn't get this right players would be discouraged to experiment with our combat and would maybe even rely on the pattern of play that they're used to in some of the games i showed before so with these challenges and goals framed we broke down the core combat systems and identified the aspects which we needed to solve tracking targeting and engaging enemies and i'm excited to go through them with more technical detail and share some of what we learned in the process if you would like to hear more about the design process covered in this section jason mcdonald our design director did a great talk on this yesterday called taking an axe to god of war gameplay and it should be up on the vault soon for you to check out let's start with tracking enemies and by which i mean everything related to enabling players to be aware of threats in the fight in order to track enemies the player has to have an idea of who the active threats are you could be fighting eight enemies at a time but which group should you really be paying the closest attention to aggressiveness affects how enemies will behave and where they will position themselves our enemies can be either aggressive or non-aggressive in a fight think of it kind of like a scene in a kung fu movie aggressive enemies are those who are actively trying to attack the player in the front while non-aggressive enemies hang around in the back and wait to become aggressive so how do we select who is aggressive at any moment and who isn't we mostly updated our old aggression scoring and token system from god of war ascension but accounted for the new camera we evaluate all enemies in the fight periodically on a short timer and give them a numerical aggression score built by the following criteria in order of importance can the enemy become aggressive and this is to ensure that they are not in a hit reaction or in a similar state where they're barred from becoming aggressive the second is what's their regression priority and this is a number set by designers per enemy type along with the range in which the priority is applied so a high priority like a really big ogre who's super far away doesn't out prioritize a lower enemy like the dragger who's already close to the player the third is is the enemy currently targeted by the player and i'll go over this in more detail soon in the targeting enemy section of the talk and finally what's the enemy's action rank and this is a calculation that looks at whether they are on or off screen their angle from the camera and their distance from kratos the enemies are sorted with a score and are handed aggression tokens from a fixed pool in order each enemy type can claim a different number of aggression tokens to signify size or importance and after the max token count is reached all others in the list become non-aggressive in this example the player has a pool of 14 tokens in the table the outline column on the right denotes how many tokens that enemy claims after the third enemy the pool of 14 runs out and the remaining enemy is set to be non-aggressive the outline column on the left shows who is aggressive and as you can see only the blue enemy was set to non-aggressive because he's missing the asterix and you can even see this in the image he's standing all the way in the back careful tuning of these values helps to ensure the player always has a manageable threat at any time these values are also changed for difficulty to either increase or decrease the overall enemy aggression once we know which enemies are aggressive and which are non-aggressive we have to figure out where they want to be in the fight space positioning needed a complete overhaul and was built from the ground up with this new camera in mind in older god of war games enemies use their aggressiveness to determine whether they should stand near or far from the player effectively surrounding kratos in two circles there was no notion of nicely spreading the enemies out or caring about one area more than another the pulled back camera allowed you to see everyone at all times so it was easy to perceive your threats and non-aggressive enemies occasionally moved around the perimeter of the fight just to keep the scene a bit more dynamic now however enemies could be on screen or off screen and where they stand relative to your perspective is extremely important the camera perspective makes depth perception more difficult to gauge distances and enemies in front obscured those who are standing behind them as such we needed to make sure enemies spread out comfortably early on in the project we tried to find the optimal positions for each enemy using a weight-based positioning system we use the fixed size circular grid around kratos which would update periodically as he moved around each grid position would evaluate itself by checking if it was on the nav mesh and if it had a valid straight path to kratos valid positions were then weighted based on many different heuristics there were a few valuable learnings early on from using the system but over time it became clear that we needed another solution the weight structure became monolithic complicated and messy making it almost impossible to debug effectively which you might be able to guess by looking at that weights kept constantly getting invalidated due to both player and enemy movement and enemies frequently bunched together because of this most importantly we saw from early internal testing that players immediately got scared as soon as an enemy went off screen and shut down any kind of offensive or confident gameplay that they were engaged in it didn't look like god of war we needed players to build a mental map of enemy locations in the fight so they could feel comfortable in combat after almost a year of iteration we scrapped the system entirely this time driven by the madness of managing the complex weights we changed the goal from picking optimal positions for each enemy to something far simpler prevent enemies from getting into bad positions this led to a strict zone constraint based positioning system the system required each enemy to have two elements a positional constraint which is basically where do i want to stand and the separation constraint which is how much space do i take as long as these constraints were met no position was better than any other both constraints would take the form of zones and would be fully data-driven these zones required a minimum amount of navmesh to be available otherwise more lenient fallback zones would be used instead of using a grid that moved around kratos we generated a grid on the nav mesh in world space the positional constraint we use for aggressive enemies was this kind of crescent shape in front of kratos to keep them on screen and you can see that in the image in red this particular shape played well with the camera's perspective and prevented enemies in the back of the zone from standing off screen as you move forward non-aggressive enemies would just stand around kratos in a white circle similar to old god of war since they didn't care about being on or off screen separation constraints the one that specify how much space an enemy takes had different shapes the aggressive enemies simply used a radius to define how much space they took so no other creature was allowed to stand in their circle non-aggressive enemies however had a wider angled separation constraint that kind of looked like a slightly curved rectangle this shape was helpful to spread out the enemies laterally in the back but didn't take up as much space since they could be many more non-aggressive enemies at any one time than there could be aggressive ones this shape also meant that the enemies in the back could be seen behind enemies in the front you could see how this led to enemies being clearly perceptible by these two images which show the same exact enemy layout but from an aerial and a in-game camera perspective this approach had much less nitty-gritty control but yielded far better results the zone data was easier to author and debug and the simplicity of solving the constraints led to both less frequent and less noisy invalidation the constraints also caused the enemies to spare out in a way that was much more visually pleasing however it still wasn't perfect by making all aggressive enemies want to be in that crescent shape in front of kratos many of the off-screen enemies behind him who had recently become aggressive would walk directly into the back of him while trying to get on screen in fact off-screen enemies in general were still really hard to track you didn't know where they were and at any time one could turn aggressive and be a new threat from a blind spot players still didn't have a mental map of the fight and they couldn't play confidently and we relied on this version of positioning for a long time on the project in fact it was only during the last few months during intense play testing that we re-examined it and tried to address remaining issues that we knew existed we had to figure out how enemies could position themselves in a way that you would just kind of expect as a player we employed new techniques to help preserve enemies as being either on or off screen and that's to say that an on-screen enemy should only go off of the screen if the player caused that to happen similarly an off-screen enemy who turned aggressive should not be forced back on screen until you've looked at it and this was all done in logic that assigned the constraint zones to enemies another technique we used was to keep off-screen enemies in their same world-space-oriented quadrant relative to kratos's position and this was just another zone constraint that they had to satisfy this way enemies that you weren't actively keeping track of would stay relatively in the same area as you've last seen so for example an enemy currently in the back left of the player's view would stay in that general area so if the player turned right and put that enemy off screen they would know that the enemy was now directly behind them so here's a demo of the system with enemies who are not attacking for debug purposes but aggressiveness as a system is still on those are two different concepts i'll commentate on what's happening but the main takeaway is to notice if the enemies move in a way that you would expect take a look so here both the green and yellow enemies are aggressive and want to stay on me and you can see that as i move around they try to stay in that crescent shape when i move past them they're off screen and their zone shape changes to be a much wider circle around the player due to the aggression scoring the pink enemy who's now on screen becomes aggressive the greed enemy stays on kratos's right side even while off screen due to the quadrants and similarly the yellow enemy stays to his left the green enemy is still treated as off-screen until kratos turns around but once he does it's back and in that crescent shape again an aggressiveness switches between enemies super frequently with all that's happening in combat but overall enemies generally stay where you'd expect them to be these key improvements along with many other small changes led to significant outcomes we were able to solve the on-screen off-screen problem and prevent enemies from standing in awkward positions quadrants were a huge success in making enemy positions more predictable and were especially helpful in more complex environments overall these changes helped our goal of allowing players to make mental maps of the battlefield and begin to play confidently and this was evident in play tests i mean players whether they were aware of it or not would start to feel confident enough to take risks and continue to be engaged offensively even when handling multiple enemies to help players further we wanted to make them even more aware of off-screen enemies and what they were doing without it being so distracting that it shifted attention away from threats immediately in front of them we knew we didn't want to do something as overt as a radar or mini-map so we experimented with offscreen indicators most players just need to know if off-screen enemies were attacking but as they became more and more confident with the game it was helpful to get an idea of where even nearby idol enemies were to best maximize their combat strategies this information could encourage them to be more offensive instead of hanging back all the time early on we tried flashing the edge of the screen closest to enemies off screen but players thought they were damage indicators instead not ideal we tried many other implementations which involved the edges of the screen but finally settled on changing them to be arrows which circled around kratos and pointed more clearly towards the enemies themselves the red indicators are for incoming attacks and the more subtle white arrows just show nearby idling enemies and it was important to design the white indicators in a way that weren't distracting for players who weren't trying to use them range attacks actually had a slightly different purple color for players who paid attention but again the goal was to not overload information if you saw a reddish flashing arrow you knew something was coming and it's important to point out that the work done on the positioning system described earlier in terms of reducing the noise and invalidation of enemies made these offscreen indicators far more stable predictable and easier for players to process our indicators help players track enemies off screen but what about enemies that are right in front of us controlling the camera itself with the right analog stick in the heat of battle is an added complexity this time that we needed to help make accessible for a wide audience looking to play an action game we tried to find a solution for tracking enemies without relying on the player to constantly manage the camera by implementing our own version of a camera strafe assist the camera will automatically try to adjust and keep relevant targets and threats in view when the player moves left to right to do this we wait enemies to find a center point for the camera to focus on in particular evading laterally really showed this behavior in action we also eventually added a behavior to reorient the camera automatically in the direction of kratos attacks to ensure that we reduced frustration for players struggling with controls as many of you likely know anticipating the player's intentions with how they wanted the camera to behave is a difficult task and we didn't always get it right 100 but our play test proved however that these camera assists dramatically improved gameplay for most players and those who found them frustrating could just turn them off in the settings menu now that we've gone through systems that help players track their threats let's look at others which help players choose who they wanted to attack by diving into targeting in a common fight scene like this which enemy should kratos target when the player presses the attack button should it be the bandit on the right because he looks like he's the closest what about the one in the middle because he's in the center of the screen or should kratos just attack directly in front of wherever he's facing without aiming at anyone in particular this is a pretty complicated problem target selection is an extremely important system for us since we need to allow players to fight multiple enemies at the same time fluidly due to the one versus many fantasy it's a core part of our combat identity we needed to make this possible without requiring players to use a lock-on system which was a directive all the way from the top a new challenge was also that the leviathan axe enabled both ranged and melee combat on the fly and switching between them should be seamless a lot of time was spent on this problem in particular in order to get it feeling just right let's start with melee targeting in old god of war games with the pulled back camera we relied heavily on the left analog stick to decide who kratos should target with his attacks the problem with doing something like that for the new game is that we didn't want kratos attacking into the camera if you pulled back on the stick one early approach we took was to kind of make it like a shooter and have it be entirely dependent on the reticle point in the center of the camera and this didn't work it wasn't intuitive for our style of game and players kept complaining you know like why can't i hit that enemy he looks like he's right there we weren't able to get their intent across and the result was gameplay that felt clunky the solution ended up being a mix between both the left and right analog stick as well as other factors we'll take a look at it in action you'll see a green circle on whoever kratos's target is at any time so in the beginning all the enemies are out of range and kratos doesn't have a valid target as soon as an enemy comes into range it becomes the initial target and you can see that with the green circle the target will change as the characters move around due to a lot of weights that are determining the targeting system [Music] if we look at only the right stick and move that around you can see that the screen centeredness alone will not pick the best target but characters completely off screen can't be targets and that's probably something you can imagine [Music] so inversely if we only move the left analog stick the stick deflection drives the intent and you can see the target change as kratos walks around and so as you saw in the video we still tried to retain the left stick as much as possible as being the main way to show intent and because it kind of felt familiar to players and i'm not going to go into them in too much detail but for those curious here's a list of the factors that played a role in target selection for reference and feel free to take a picture if you want as long as there is a viable enemy candidate the targeting system will always choose it over not having any target at all you can't choose to attack into an empty space instead of a valid target when there are no possible targets kratos will always attack directly forward along the direction of the camera you can't steer his attacks left right or backwards and it's also important that we fulfill the fantasy of being kratos right like we don't want him to miss his attacks clumsily when enemies are right in front of him so as with past god of war games if kratos has a target all of his attacks are automatically rotated to face them one interesting issue that we saw within the studio during the development was that people control the camera really differently corey barlock the game director always had the camera flat aimed right at the horizon line to get the best cinematic presentation those of us who played the game correctly of course angled the camera down towards the ground so we had a reference point for judging distances having gameplay that was on your side meant that regardless of how you played we knew it felt terrible to attack and miss an enemy who looked like it was right in front of you even if it really was a good distance away we re-used an old god of war ascension feature called suck to target on most of kratos attacks to help with this issue each attack specified a range speed and stop distance which the engine would use to suck kratos towards his target this would make it significantly easier to hit enemies but as with most older systems suck to target had problems from this new camera perspective the behavior often felt jarring and disorienting and this was especially true if an enemy was on the side of your screen and cradles suck towards it laterally since the same distance would appear to be much longer because once again depth perception worst of all was the case where you would be sucked in only to still miss your attack because the enemy happened to move right out of range during that time so the old sector target had problems we had to update and change the behavior to help ensure that you reached your target at a specific time in the animation and this is kind of like a simplified form of motion warping to help with the jarring lateral suck problem we scale the range of suck to target depending on the target's angle to kratos the wider the angle the lower the range this way kratos would miss his attacks more frequently against enemies away to the side of the screen compared to enemies in front of him even if they were at the same distance away and we were pretty happy with the results some of the examples i stuck to target in the video are obvious over large distances but it's actually happening much more frequently than you might realize you'll notice it more when the camera is aimed downwards towards the ground as a fun fact we actually toned down the suck distance dramatically for our hardest game difficulty gimme god of war because we wanted players to pay more attention to their positioning in combat once again i apologize for our range combat targeting we look to shooters for guidance on building both aim friction and zoom snapping aim friction refers to the system slowing the reticle over targets as you pass them by and zoom snapping is a system which snaps the reticle directly to a nearby target when you pull aim as you can see both of these systems were tuned very liberally as we wanted to encourage players to mix ranged and melee attacks easily without hassle in doing so we made it on the player's side kratos could both swing and throw while attacking an enemy fluidly alright i know i know i talked earlier about how other games use lock-on and that it isn't very god of war and i meant it it was a directive from the top to make the game completely playable without a lock-on system since managing lock-on can be cumbersome and frustrating for many players and honestly for most of development right up until really close to ship sorry g we didn't have it at all in the end however we realized through playtest feedback that some players just always wanted lock-on partly due to the fact that they were already familiar with it from other games and kind of expected the same players approached the game quite differently with lock-on and that's okay but it also meant that we needed to develop a lock-on system late into the project that could work even if the game wasn't designed around it and many players might have experienced frustrations with using the system for this very reason we used our already existing melee target as the initial locked target when the lock system when the lock-on system became engaged so the transition would be smooth the right stick could be flicked in a direction to change the target selection in screen space as opposed to cycling through an ordered list to make it a bit more intuitive when fighting multiple enemies we decided to allow players to adjust the camera's elevation even when they were locked onto a target unlike most implementations of the feature so they could still see elevated points of interest or even just take in the environment without losing their target one particularly interesting feature we added to lock-on happens when you pull aim we allow you to aim around freely but retain your lock-on target so you'll return to it when the aim button is released keeping with the goal of supporting fights with multiple enemies this feature allows players to handle two enemies at once even when using lock-on as they could focus primarily on their locked on target and occasionally throw their axe at the other now that we've covered systems that help players track and target their enemies let's focus on those which help players to continue to combat them once they had already been engaged to better understand this let's take a minute to talk about animation translation and how it affects engaging enemies in combat i mentioned during the introduction that having exaggerated and fun enemy hit reactions as well as strong attacks for kratos which moved him forward through the space we're both integral to our combat system's identity these two concepts play into each other due to animation translation god of war enemies translate backwards substantially during hit reactions as a result kratos is always moving forward to keep up eventually kratos and the enemy keep moving in one direction and cover a large distance over the course of a combo one problem is that animation translation like this was challenging with the new camera it's easy to knock enemies off screen to the left or to the right and zooming forward through the world is disorienting as you see the environment pass rapidly by you there's also the fundamental issue that in this perspective moving forward is naturally disadvantageous since you are reducing the amount that you can see of your surroundings on the other hand moving backwards is naturally advantageous as it allows you to see more and it's one of the reasons why players are naturally drawn to kiting or defensive gameplay patterns with this close camera we had to be careful with how we balanced forward movement of kratos's attacks which really helped them deliver their power and satisfaction with the disadvantageous state that they often put the player in because of this we had to revisit how we author translation in many of our animations we reduced the translation of basic enemy hit reactions substantially from how they were made in the past and even reduced the translation of some of kratos attacks but this actually worked out pretty well since sucked to target could still allow him to reach enemies with his attacks and it also provided enough movement to give the satisfying oomph we needed but he wouldn't go zooming across the screen in an empty room anymore playing around with all this took a lot of trial and error and having the functionality in our design strips to procedurally scale the translation of animations was incredibly helpful for testing outcomes with fast iteration speed this video shows the problem of highly translating reactions causing enemies to go off screen the reactions themselves are fun and satisfying but it's extremely frustrating to not be able to follow up on your target with the added insult of having your target switched involuntarily back to another on-screen enemy and even worse the enemy that you just beat up is sure to return as an off-screen threat in a few seconds like let me make this clear we were punishing you for attacking with your combos and causing fun enemy hit reactions and that's the exact opposite of what our combat should be about all right time to talk about one of my favorite features that we called strike assist and how it helps solve this problem in a stroke of pure brilliance our then lead gameplay designer asked the profound question what if we just kept enemies that were hit on screen seriously check out jason's taking an axe to god of war gameplay talk when it comes online there's even an og god of war trophy named after him right right anyway after back a bit of back and forth jeet and his programming team came up with something really cool and it looks something like this with strike assist and bear with me for a second attacks cause their victim to snap to an initial orientation at the beginning of their hit reaction such that their adjusted trajectory will now go towards a blend between the camera's facing vector and their original trajectory and if all that was too confusing all that means is that enemies are now pulled onto the screen if they were going to go off of it the larger the delta the larger the correction each attack can specify its own blend so some attacks can say make the enemy go right to the center of the camera or just pull it close enough so that it stays on screen naturally the farther the translation of the reaction the more you'd see the effects of this there's a lot of complexity to this feature and how it was used but that's kind of the basic overview with strike assist enemies stopped being hit off screen successive attacks in a combo on the same target would pull it more and more into the center of the camera and since if you remember kratos rotates to face his target when he attacks this eventually led to both kratos and his target translating back down the camera's direction which was great for framing as a bonus this made switching between melee attacks and range targeting easier as zoom snapping worked better if the enemy was already near the reticle which is of course at the center of the screen this feature led to a bunch of other new discoveries as well see players have control of the camera so they could use this to aim where they wanted their enemies to be hit to be clear this was huge building up combat systems which this new camera had led to a new type of moment-to-moment gameplay and expression of player intent and skill for example strategies such as intentionally hurting enemies together before pulling off a massive runic slam attack or moving the camera to direct an enemy to fly off the side of a cliff with the last hit of your combo were now possible and it was so intuitive through just looking at what you wanted to do that most players had no idea that this was even happening at all strike assist may have possibly been the most important aspect in making the new combat system come together it was a huge victory for the team apart from strike assist we used a few other tricks to help players engage their enemies by allowing them to keep up the offense while enemies were already playing hit reactions many of the hit reactions themselves would fire invisible collisions allowing enemies to bump into each other and even hit enemies to their sides despite their sometimes being no visual contact this rewarded players by helping them stay safe since enemies who were bumped were sure not to attack until after they had recovered another trick was disrupting enemy aggression the close camera naturally makes kratos and his target take up a large part of the screen real estate and obstruct the player's vision we realized that we would have to incentivize offensive gameplay even more if we wanted players to feel confident despite being able to see less remember the aggression system the topic i covered earlier about how we choose which enemies are active threats we made some simple tweaks to it to handle this if an attacking aggressive enemy gets interrupted and plays a hit reaction it retains its aggressive tokens temporarily instead of relinquishing them immediately for others to use for a short period of time that enemy would block other non-aggressive enemies from taking these tokens but couldn't do an attack itself since it was already in a hit reaction and basically that effectively gave the player a power play every time they were being offensive if they extended their offense for too long without worrying about other threats they could be punished but for a window of time they'd have the advantage in most scenarios on our normal difficulty say against the pack of draugr only two enemies would be allowed to be aggressive at any one time this meant that the player just had to worry about one other enemy apart from the one kratos was already engaged with which isn't all that bad to keep track of enemy air reactions specifically would prolong this effect for more time giving you yet another bonus to keep your combos going despite the threats around you speaking of at first we didn't have enemies getting launched or juggled in the air due to the new grounded tone however juggling enemies is part of the playing with your food identity we needed to retain so we have to solve juggling for our new combat system and close camera the largest issue is that enemies could now be constantly hit upwards until they are above the camera and you can't see them anymore not to mention the leviathan axe had limited melee attack range and since we removed the jump button from this entry in the series there would be no way for players to stay engaged on a target like that to make matters worse juggled enemies took up a lot of real estate on the screen and occluded nearby threats so players rarely felt safe while attacking them to tackle the first problem of enemies exiting the camera frame we implemented a float height system which could be used to tell an enemy how high it was allowed to go relative to kratos if the root joint of the enemy ever went higher than the specified float height some corruption velocity would be applied to keep it in bounds and this could be tuned per enemy hit reaction by designers in script we tried to make this look as natural as possible while still keeping enemies low enough where they were in range of the axis nelly attacks it's worth noting that to get this feeling right we actually had to animate the character's falling instead of relying on gravity or physics since the trick was that the bounce of juggling always had to feel good enemy air reactions also warranted the same technique we used on ground reactions by having invisible impact collisions that hit other nearby enemies except we made them much much larger than their grounded counterparts enemies who could fire range projectiles from afar however were still always a threat so players still have to be smart about when they should and when they shouldn't juggle an airborne target so after the long journey of development what did this process yield for us in the end was it possible to make a combat system within a close and intimate camera feel like an honest evolution of god of war not all players might have shared the same experience and there is still much room to grow in all the aspects i've covered in this talk but by and large it was rewarding to see most players engage in our combat in an offensive and confident manner despite the many issues posed by the new camera by keeping our combat identity at the forefront of all of our systems for tracking targeting and engaging enemies we were able to work towards gameplay that felt satisfyingly familiar yet fresh from a new perspective the result was a more deliberate and unflinching god of war combat experience that felt authentic to the franchise's legacy and this line of thinking and problem solving in terms of breaking down your gameplay's identity and keeping it in mind at all time during development can apply to other ips in any genre which are similarly going through a challenging reboot or reimagining reaching our results took a very very long time of more than three years of constant iteration and testing and this was a crucially joint problem between our gameplay design and programming departments and involved many talented people we needed designers deeply involved with the technical implementation and programmers invested in the design context and player experience and this was true for both on the ground work as well as department leadership as a result these two departments work very closely with each other both in terms of feature development and proximity at our studio finally the close camera seemed counterintuitive to almost all aspects of god of war style melee combat when traditionally good design principles conflicted with vision the team took a difficult risk and committed to the latter the process was challenging and there was a lot of doubt throughout development but it was what the game needed to shine as a whole what combat needed to evolve in a bold new direction and at the end of the day it's what made the journey absolutely worth it that's the talk if you're excited by anything i've gone through today and would like to join us on our next journey please get in touch we're hiring across many departments including combat design and gameplay programming and we'd love to hear from you thanks so much it was a pleasure [Applause] and now uh jeet and i can open it up to questions and these are bright lights yeah this work thanks for the talk i noticed in the valkyrie fights a lot of your priorities are opposite like offensive is usually punished and they love being off screen so you know what was the process there sure i could talk a bit about that so the valkyrie fights are kind of like the the most challenging combat encounters and they're all apart from one uh you know 1v1 encounters so because those are testing mastery um camera control is actually part of the mastery that we're testing in those fights a lot of the core fights that you're doing throughout the game with like let's say dragger it's really just like getting in there and fulfilling that traditional god of war fantasy but with the valkyries it's like we want to test your mastery of your combat suite as well as the control of your camera um and so with that's kind of with regards to the second part of your question for the first um and this is true also in past god of war games that that kind of punching back feeling of playing with your food doesn't necessarily extend to all enemies in the cast it's really what we focus on what we call like the grunt class of characters and other characters that are beyond that that are less kind of like playing with your food the fun comes more from the challenge and the kind of pattern of play that they provide thank you yep hey really great talk thank you very much so we heard a lot of good stuff about the new weapon the leviathan axe i'm curious about everyone's favorite the blades of chaos what did that have to get reinvented much for this and did it affect some of the gameplay new features that you talked about sure um so the question is you talked about the leviathan axe but what about the the blades of chaos uh for this game and yes it had to get relooked at uh a lot uh i don't actually cover that in this talk but for a long time uh the the attacks with the blades and game we wanted to shift that up from the different camera and they actually had the blades in hand the entire time and not actually swinging out so uh you unlock them later in the tree but the delay attacks where you're just swiping with the blades in hand that used to be the core kind of bass combo with the blades uh because it kind of played better with the camera uh but you know after we messed around with it for a bit we were like all right this doesn't feel like the blades right so we we kind of swapped that and kept the original move set um on the on the camera but a lot of the systems that we have actually extended once we had made them work with the axe they actually extended to the blades there's a bit of tuning and differences that we do but we were able to reuse a lot of it yeah and so just from a sort of technical point of view compared to the old god of war blades the the new blades had to be completely redone for this project looks beautiful by the way great work thanks hi i was wondering uh if you guys did any sort of partial uh rendered or not rendered kind of definition for larger enemies to determine whether or not they were on screen or slightly off screen in terms of the targeting and things like that so a lot of the targeting stuff is done we have i mentioned in the talk but there's like a root joint on every character right um and so a lot of the times systems are doing a pretty naive check in terms of is the root joint kind of on you know on screen we do have a rendering check as well different systems use different checks um and actually for positioning uh i didn't get into this in the talk but we don't actually when i say enemies are on screen or off screen it's more that are they in a frustum in front of kratos's facing direction uh because the camera moves so much that often uh you would invalidate the checks like instantly um so it's a good question because it's actually pretty nuanced like like checking whether enemies on-screen or off-screen i think there's like three different things that we do to to check that depending on the system thank you and i think like different different systems need different uh different ideas of what on-screen really means like for positioning you're going to want to do something different and then like me here mentioned with the camera strafe assist that required a completely new version of what onscreen really meant so i think it is about trying to make sure the solution applies to the type of problem you're solving all right thank you hi um i'm wondering from a production standpoint how you guys may prioritize what you feel is important in combat right is it like some guy's running into room and he's like i want to be able to kill 10 jogger in in 10 seconds like how do you know what how do you deal with that when there's different enemy types like where do you guys pull that inspiration from sure that's a that's a really good question um so at our studio we we have the benefit of kind of a lineage of working on these kind of like melee action games and so in a lot of like directorial and lead positions we have people that have you know gone through i mean we have many people that have been there from from god of war one right that are still in the studio and so because of that there's a bit of a history and a philosophy to how we design and the first thing on combat is that in a test room you know with one enemy who isn't fighting back that should be a blast like it should just be fun to beat up that character and nothing else matters until we you know we haven't reached that and that was true on the project i mean we didn't have that for a long time on this game uh and that was very frustrating because people would be wondering about you know hey the rest of the enemy cast like why aren't we building that it's like well the jogger isn't fun to hit so why even bother doing all this other stuff if that's not going to be good right um yeah i think i think he yeah mia pointed out the fact that like we you know focus on the basics first get your core working really well uh and then in parallel start to look at the finesse stuff so like he mentioned like for a long time we were just trying to get the basic attacks and the basic sort of drug or core loop working well and then there was a moment in time when we decided like even from a production point of view that we are going to now start really focusing on combat feel and that's kind of where a lot of these ideas came from so this is where like jason and i would sit down and and uh play the game a lot and start to you know really try to break down feel issues and at the same time we're starting to do heavy sort of play testing as well so having sort of you know very very scrutinized uh breakdowns of reviews just strictly around combat feel play testing heavily and then sort of also the design team doing a lot of internal reviews sort of gave us a really good perspective of what should be prioritized and what are the next steps that we're going to take thank you guys hi thank you so much for your talk it's very cool my question is more on like how do you handle ranged attacks or arranged enemies did you use the same system in terms of positioning and how much did level design play into this uh into the combat design and vice versa sure um the second part of your question are you talking specifically for ranged enemies or just in general in general okay so the first part of your question um pretty much every enemy uses the same fight positioning system that we have um but the zone shapes can be changed depending on the enemy we do cheat a bit with some of the range enemies where we actually have a behavior that we just call turret mode internally and those are enemies that are stationary they just spawn in a position and they never move so in some areas we're able to kind of pepper them around especially with a bit of verticality but otherwise really every every enemy just follows the positioning system even if they're arranged the aggressiveness system has some takes that into account so even enemies that are in the back if they're ranged they're allowed to kind of like you know their attacks can kind of seep through um from the second part of your question we actually have an encounters team between the combat design and level design team and the encounters team are you know they're focused on what will the fights be what's the enemy composition um they're kind of working very closely back and forth the the exceptions are when we have boss fights that's kind of when combat designers and level designers really work closely together on a specific feature cool thank you did you ever want to uh explore the area of giving players a rewarding way to attack an enemy that is off camera or behind them figure so sorry the question was uh did we ever explore giving players a rewarding way to handle off-screen enemies yeah sure like i i think of like the enemies in dark souls who can fall backwards yeah i mean we tried we tried uh allowing kratos to start sort of attacking uh much white like getting outside of the the main camera frustrum and for the most part like i would say the decision was more that it just didn't look very good even from a fidelity point of view sort of having kratos completely turn around and start angling you know laterally through the camera uh that was one of the main reasons i would say that we decided to avoid that yeah the other thing is that we we tried a lot of iterations on the quick turn so that was like we had that in the game where you could just kind of quickly turn around um and for a long time we actually had a version of that that included an attack so you would do a quick turnaround and attack an enemy um but it felt kind of weird right because it kind of just auto did some auto seeking to find that target so if an enemy was here you'd kind of like run around and then personally like you know it would you'd automatically attack for you uh and that's not really a pattern of play that we were trying to encourage um so we doubled down on like no like camera is a big part of this game and part of learning the combat in this game is learning how to use the camera thanks hi um you seem to have very clear objectives of what you wanted to achieve with the combat even though you iterated on the techniques but how true was that is it that artifact of hindsight in the speech how long do you take for for you to please start clear your objectives yeah so jeet kind of alluded to this um earlier you might have more to say on this but uh for the first couple of years on this project there was so much that was new and changing and and it was very hard to move forward because we were questioning a lot of what we were doing um but i i said in there that in the fall of 2016 um so this is after we had done the e3 demo and people you know we people saw the vision and they liked it we're like oh man let's you know like we could do this but everything still doesn't feel very good and that's when we started to kind of step back and do these internal combat play tests on the team um and that really you know when we started getting a lot of data we were like okay this this is interesting and that's when jeet and jason really like started meeting with corey really frequently on being like no let's let's break this down let's actually come up with a strategy so obviously in a gdc talk you know i'm going to try to like couple things together but the idea wasn't far off like it was a very targeted approach and like let's let's look at just targeting how what are we gonna do about targeting just get the intent right okay what about juggled enemies this air combat feels terrible you know like it was looking at those different things and tackling them and like to to mere's point i think maybe in terms of like pillars i would say like a lot of ideas kind of came back given that we started getting more confident in our ability to achieve and execute them so like for example like juggling it was something that again very much lines up with the idea of play with your food but you know there was a lot of excuse me resistance to it maybe early on because you know we were still trying to just get the base stuff in a good place so i think the core tenets of kind of if in the in the very first early slides where we kind of hear laid out like what is the vision of god of war combat as a whole i think that still stayed true pretty much like we knew that as as being a fundamental kind of vision of what this would be it was more like how do we really execute it how do we realize it under these new uh sort of in this new sort of framework and then finding our way through that was you know practically kind of all the solutions that we came up with thank you hey guys um i was you guys briefly touched on it but um how did all that kind of stuff change on the harder difficulties like positioning and aggressiveness and how do you make that fair again go no let me take this [Laughter] i could give a whole talk on this uh uh i we spent like two months straight towards the end of the project working on the difficulty tuning um which was not accounted for so sorry production uh yeah so so we um a lot of us on the comment team are really big devil may cry fans and so whenever we used to play dante must die we we love the way that difficulty was handled and so we you know are at the time our lead gameplay designer jason mcdonald he was like you know we're going to make a hard difficulty and we're going to do it right like we're going to do it where it feels like a different game and it's a different combat puzzle um and so man i can't get into the details too much right now but like almost man like almost everything in the combat system is different for those difficulties uh far more than people would expect like we even treat the stats differently on the different difficulties like the stat tuning curves are completely different so um your choices in what you choose to equip matter a lot more um that's just an example of something you might not have expected but yeah it's are they like more aggressive or on the aggression side yes so like remember that token slide that i said um we changed the amount that enemies take so let's say like you know if you have like 10 tokens available and an enemy takes a four then it's like okay well only two you know can be aggressive at any time we would shrink that down to three so then you could have three enemies on you at any one time like that's an example but in a more complicated way we actually change up the attacks that enemies do on harder difficulties so um and we we kind of make combo strings possible with enemies that they weren't able to do previously as well so it's it's like a multi-faceted approach to overall change the experience good thank you thanks i think we might have time for just this last one uh yeah very quickly i i hope i didn't miss that in the presentation which was great by the way the amount of token that is available to the player is it dynamic is it where where does this fixed number come from uh so the question is the the tokens that that the enemy has that the player has for aggression um is that dynamic and the answer is actually no but um in practice we change it on the on the enemy side yeah so not on the player side that number is a funny number it's it's a pretty legacy honestly like it's just a thing with god of war even in past games where there would be this number and we just kind of stuck to it i don't know we didn't really change it too much but we would kind of change the enemies the amount that they took but it wouldn't be dynamic in the game it would be more that per difficulty we would change it within the same difficulty it never changed you can think of it as like being statically defined on a per kind of uh mode basis okay so it remains the same throughout the game no matter your level it's just like okay it's 12 now yeah for a specific cast member it remains the same okay thank you very much cool and if there are any more questions i can take it in the in a wrap-up room 30-20 all right so yeah i'll head over to 30 20. if there's any more questions i could take them there thank you so much guys