Image courtesy of Sumo Newcastle

Designing Hood: Outlaws & Legends — a class-based PVPVE heist game set in the medieval era

Over the past 15 years, Andrew Willans has been part of developing a range of AAA titles while designing at companies such as 2K, Ubisoft, CCP, and now Sumo Digital. Career highlights include Driver San Francisco, Watch_Dogs, The Crew, Tom Clancy’s: The Division, EVE: Valkyrie, and the BAFTA-nominated (for game design) Grow Home. He is currently game director for the multiplayer medieval heist game Hood: Outlaws and Legends.
Developed by Sumo Newcastle, Hood: Outlaws & Legends is unlike anything you might have played before. It’s a class-based PVPVE heist game set in the medieval era. The British developer designed the title to be a mixture of Game of Thrones and Payday set in the Robin Hood universe. Shacknews praised its innovative blend in its review, stating, “Hood: Outlaws and Legends is an excellent change of pace from your standard online multiplayer games. The PVPVE style works surprisingly well, creating new obstacles and offering unique strategies at every corner.”

The game places two teams of four against each other and the AI across three stages: keep infiltration, breaking into a vault, and escaping with the treasure. To facilitate these distinct phases, Sumo Newcastle created a variety of different-sized maps that offer numerous entry points and pathways coupled with different classes with wildly different abilities that excel at different scenarios. To add replayability to the mix, the game features random elements that mix up the location of keys, where enemy AI patrol, and more. These myriad options enable players to approach missions tactically, stealthily, or bows blazing. 

To see how Sumo Newcastle approached developing Hood, we interviewed Andrew Willans. The game director talks about the projects that influenced the title, shares that the team leveraged Unreal Marketplace assets to prototype the game, talks about some of the challenges that come with innovating in the space, and more. 
Image courtesy of Sumo Newcastle
Hood: Outlaws & Legends features a unique multi-staged PVPVE premise. What inspired it? 

Game Director Andrew Willans: We’re a healthy mix of PVE and PVP fans here at Sumo Newcastle. Much of our development expertise has been in creating competitive multiplayer experiences, but we’re equally as passionate about co-op games like Vermintide 2 as we are about adversarial shooters such as Call Of Duty (to name just a couple of our favorites). I think we really wanted to have our cake and eat it, and games like Hunt Showdown had proven there was an appetite for mixing the genres. We had lengthy design discussions about a more traditional face-off between the teams, or even having a team of outlaws face off against player-controlled State guards, but ultimately, we were more excited about treading new ground.

The game is loosely based on Robin Hood and set in the medieval era. Why was that premise and setting right for the game?

Willans: The main word that comes to mind here is gritty. When we started concepting for Hood, we had a head start on defining the tone thanks to some wonderfully dark and mature character art we inherited from an unused concept within the Sumo family. When pitching to publishing partners back in 2019, the market was flooded with hero shooters and battle royale clones that all seemed to embrace a very friendly, almost cartoony aesthetic. We wanted to stand out from the crowd, appeal to a slightly different audience, and stay true to the gamer DNA we had in the initial pitch. We asked ourselves, “What if Game Of Thrones did Payday in the Robin Hood universe?” 
Image courtesy of Sumo Newcastle
Hood: Outlaws & Legends launched with four playable classes that include the sharpshooting Ranger, support class Mystic, stealth specialist Hunter, and melee-focused Brawler. Considering they all offer vastly different skill sets and special abilities, can you share your approach to designing and balancing them? 

Willans: As a starting point, we considered the two ranged characters as appealing to our more hardcore players. Anyone should be able to play a role and have fun, but to really get the best from either the Ranger or the Hunter, players will need to refine their input skills for long-distance and tracking shots in the case of the Ranger, or strategy as you remain unseen and chain assassinations when using the Hunter. It's been interesting to see how initially a lot of players considered Marianne (the Hunter) to be the weakest character, when in actual fact (and in the right hands), she’s arguably the strongest.  

The melee characters present players with a more immediate and accessible experience. If you are more into a run n’ gun playstyle, then the Brawler (John) is a perfect choice for mashing! The Mystic (Tooke) requires a little more caution before launching his heavy attacks, but for a healer, he can still pack a serious punch and leave opponents in a stunned state for longer. Both the melee outlaws use a block instead of the defensive evade used by the ranged characters. This was a deliberate decision to further differentiate the playstyles. Block is easier to time than evading, but mastering the parry is key to getting the best out of both Tooke and John. 

We spent a lot of time in combat gyms just purely testing the group dynamics of combat, hit reactions, stun states, stamina, hit points, and more. Obviously, nothing beats a live game for collecting information and now that we’re released, we are pouring over the incoming data to validate our tunings. This will be an ongoing process, which is the norm, and any tweaks we make will now be guided by data rather than opinion.  

Hood: Outlaws & Legends features three stages of play that include infiltrating a keep, breaking into a vault, and escaping with the treasure. Was there a lot of iteration involved in nailing the gameplay loop?   

We were always happy with the distinct stages of the heist. Narratively, this is what makes a great heist movie, and this three-act structure translated perfectly to delivering gameplay that required different amounts of stealth versus action at the right times. Our biggest iteration around the core loop was in how much information we provided to both teams. In early prototypes, we never informed the other team of their opponents’ progress. So, it was possible to steal the key, open the vault, and then extract the treasure without the other team even realizing it. This was fun from a “perfect plan, perfectly executed” perspective, but it also made for some pretty lonely games if you didn’t manage to locate the enemy outlaws. 

We found that the most intense and fun moments came from when the two teams encountered each other, so we experimented with various amounts of UI to keep both teams updated. We’re happy with the three-stage update messages for launch, but we’d still love to explore how a more hardcore variation of a heist could play out post-launch. In that instance, stealth would be pretty much essential to victory. 
Can you talk about how you developed the game's AI (and nearly indestructible sheriff boss) that challenges two competing teams of four players while keeping things fresh on repeat playthroughs? 

Willans: This was a constant juggle between technical restraints and gameplay. On the one hand, we needed to provide a noticeable consequence when a player alerted a guard, but at the same time, combat between too many targets can quickly become overwhelming and prevent you from properly defending, evading, or even landing a counter-attack. Take this challenge and multiply it by eight (for all players) and you can imagine it took some serious iteration to find the sweet spot. We used different guard archetypes to provide different levels of challenge, so a single bowman can present a serious threat to a melee character because they will need to close the distance to attack them. In the case of the ranged characters, their arrows and bolt cannot damage the armor of the knight archetype, so they need to escape danger or use gear to distract them in order to perform an assassination.  

To keep things fresh, we have multiple posts and patrol paths that alternate each time a map is played, and, of course, the AI can be led pretty much anywhere. So, once the heist is in play, there’s very little you can predict other than the AI’s starting position and the fact that the treasure vault will be guarded by a knight. Even the resistance you will encounter during extraction is based on whether or not the AI has seen the treasure chest as you take it to the winch. Stay stealthy if you want an easier time, but trip the alarms and you can guarantee to see more state squads descend on you as you try to bank the loot. 

Can you share your approach to creating the game's levels? Are there any unique design elements to consider when you've got PVPVE gameplay coupled with different character abilities?

Willans: We needed each map to be engaging for multiple playthroughs, so even though some maps seem pretty huge in terms of sightlines, which is better for ranged characters, they are also incredibly dense as you enter the main areas in pursuit of your objectives, which is better for melee characters. Some maps contain three treasure buildings, each with multiple floors that could spawn a treasure vault. Ensuring variety was critical to creating a level playing field for players. It's nearly impossible to predict where the objectives will appear, and even when you do get lucky, there are many dynamic systems that ensure no two heists play out the same way. 

There are also some more obvious level design ingredients that are directly linked to a character's abilities, such as the portcullis gates that only John can lift, or the rope ladders that only Robin or Marianne can shoot down to open up alternative traversal opportunities.  
Image courtesy of Sumo Newcastle
Players can often use stealth or go in bows blazing. Can you talk about your approach to creating a sandbox that facilitates different playstyles?

Willans: “Embrace the emergent gameplay” immediately springs to mind. Throughout development, we balanced between stealth and action to try and find something which embraces the fun, but isn’t too restrictive or open to trolling behavior. The biggest challenge when creating a purely multiplayer stealth game is what happens when someone on your team doesn’t play the way you intended them to. Penalizing the whole team feels unfair, so we opted for a more personal consequence. If you trip the alarms you are named and shamed in the game feed, and your character will be highlighted in red so the other team can see your position, and the guards will swarm you, rather than your teammates. Your team can remain hidden and continue in stealth if they choose not to jump to your defense and get involved in open combat.

That said, open combat, sprinting, sliding, throwing grenades, and firing explosive arrows is just so much fun and key to embracing the fantasy of a heist. Just imagine how anticlimactic the movie Heat would have been without the shootout during the getaway scene. So, we absolutely embrace this in the final act. The battle for extraction should be tense and dramatic. But if you managed to get the treasure chest loaded without the guards seeing you, then you can make that stealthy escape. This is rare, but absolutely possible with the current AI escalation rules. 

We know there are a lot of players who desire a more hardcore approach with things like limited lives, limited UI, fail conditions based on alarms tripped, and more. This is something we’re very keen to deliver (we had fun with these types of rules during development), but we need to build our player base first, so accessible gameplay and clear mission objectives were our priority for launch. 

With many ways to fight that include archery, stealth attacks, and melee weapons, can you delve into how you developed the game's combat system? 

Willans: Our very first designs involved all characters having both a ranged attack and melee. So John,  the Brawler, was intended to throw his hammer over a distance, and Tooke, the Mystic, launched his mace from the end of his flail (chain). This sounded great in theory, but when we started breaking the design down further, we decided that recovering a thrown weapon wasn’t much fun, and they could easily be lost in combat leaving you with just your bare hands. So, these discussions helped us define the two main fighting styles of ranged and melee and gave us our first combat rule: no character should be without their weapon. This forced us to think creatively about how Robin’s bow could be used for a melee attack, and also led to the creation of the wrist-mounted crossbow with the knife attachment for Marianne. 

In our early playtests, John was actually seen as the weakest character because he was so vulnerable to headshots. All that rage running towards you was easily defeated with a well-timed shot to the head. So we introduced blocking and parrying as a means of defense, and also created more differences in the damage you could deal with either a fully charged, or quick fire shot from the bow. We followed this and iterated through the entirety of development. Characters needed to feel OP for the power fantasy, and then we tuned up the other outlaws to match.  

Assassinations were essential to delivering on the promise of stealth, and their development journey is still far from over. Make them too easy to perform, and all you see are players backstab fishing, looking like a deadly conga line. Make them too hard, and players feel the system is unfair or buggy because they are used to the incredibly user-friendly functionality seen in single-player games like Assassin’s Creed. Finding that sweet spot for competitive play is tricky, but now that we have proper data to track, we can tune things with more accuracy. One of the first things we’re addressing is the ability to stop an assassination in progress by striking the attacker mid animation.
The game has been praised for being accessible despite featuring many moving parts. To what do you attribute this?

Willans: We always thought our biggest challenge was to communicate the three stages of the heist, so we definitely doubled down on the player-facing information and how frequently it updates. Having a character like John, the Brawler, enables anyone to just jump in and have fun without needing to hone the skills required to accurately hit a shot with an arrow. I also think that the team objectives nudge players into cooperation. We have so many moments where team play is essential for success, and I hope that most experienced players are guiding and assisting their gang of thieves as it is in their own interests to do so.  

Many of the seniors on the team are former Ubisoft devs, myself included, and accessibility is at the heart of everything they do. We try to love the player wherever possible, but with our team size, it was definitely a challenge to polish everything to the goals we set for ourselves. It’s great to see that the most important elements are being understood so easily. 

With fantastic lighting and shadows coupled with beautiful architecture, Hood: Outlaws & Legends looks great. Can you talk about how you nailed the game's polished look? 

: A very talented and passionate art team is the short answer. From the minute we knew we were “going medieval,” they were out and about all over the northeast photographing castles, sunsets, and storm clouds. We have a wealth of history on our doorstep, and this really helped drive the realistic look to the buildings and textures. I saw a lot of 8K rock images in 2019, a lot! 

While Hood is not meant to be historically accurate, I’m pretty sure it’s close. The team really captured the feel, mood, and tones we associate with that period from film and literature. Game of Thrones was a huge influence, but we tried to push things further with some maps to give them more identity. Being a stealth game, the danger was always that we ended up with all night maps, or 50 shades of grey, but if you look at maps like Marshland or Outpost, you can see how the contrast between bright light and shadows is just as effective for stealthy maneuvers.  
Image courtesy of Sumo Newcastle
While Hood is a multiplayer-only game, the studio has stated it aims to paint a narrative with its environmental storytelling. Can you share how the team went about achieving this? 

Willans: Each map has its own story to tell. The state is the ruling authority, and they have seized this power by force and manipulation. As players explore the maps, they will see where the old world meets the new. In Marshland, the deliberately flooded homes and old places of worship crumble beneath the towering state keep and its new church. In Outpost, the once dense forest has been felled to create a wooden fortress to protect the state’s garrison. In Citadel, players will see the cold heart of the state in all its oppressive glory as gold-trimmed buildings dominate the skyline and towering statues gaze down on the peasants below.

We also have some really nice info text to accompany each map, and a number of collectible trinkets that are map specific. These give some context to the plight of the people within each region.  
The journey has only just begun, and we’re committed to building Hood into the game we all envisioned it could be."
- Game Director Andrew Willans
What has been the most challenging development aspect of delving into this new genre?

Willans: Huge ambitions and limited resources. We knew we had a solid design for the PVPVE heist. Even in the very early prototype, you could see the potential and feel the fun in every match. Scaling up to deliver that vision at the quality benchmark we wanted was always going to be a challenge. On paper, it was a huge task: maps that are dense enough to be highly replayable and feature enough variable objectives, a character roster that appeals to both FPS and melee combat players, an AI system that could respond and provide a challenge for up to eight players at the same time, stealth mechanics that provided advantages for those that embraced them, but didn’t limit the fun, and often loud chaotic PVP encounters. 

To deliver best-in-class on all these features would be a struggle even for the 600-person teams of the bigger studios; we had 60. So I’m incredibly proud of what we achieved at launch, and will continue to build on over the course of the year. During our Early Access weekend, we were glued to streams on Twitch. Most of our players saw past the rough edges and were having a great time. They were cheering at clutch saves on the winch, fist pumping when they wiped out the opposing team, and screaming with surprise when they pulled off insane shots from halfway across the map. They sounded exactly like we did when we played. They got it.  

Doing something new is always scary (and risky), but hopefully our growing community will stick with us as we improve, polish, and build on some pretty solid foundations for fun. 
Image courtesy of Sumo Newcastle
What made Unreal Engine a good fit for the game?

Willans: Most of our studio expertise has been in Unreal Engine. The last project we did (EVE: Valkyrie) was in UE4 and so it was the logical and best choice for Hood. It's great for rapid prototyping, Blueprinting, and the Marketplace is so well stocked that we could use temporary assets to sanity check all kinds of things before going into full production. I think the word that comes to mind for most of the developers is “easy.” It's what we know, the results are great, and the tools are amazing for workflow. 

Was it challenging to incorporate real-time ray tracing and 60 FPS performance on next-gen consoles?

Willans: Not really. Largely thanks to the support and feature set from UE4. Obviously, I don’t want to downplay the hard work of our coders in the optimization department, but the biggest challenge was making sure we provided enough opportunities in the maps to showcase it. Marshland is a great example due to the huge body of water that makes up 50% of the map, but it's equally impressive to see the ray tracing in action on the muddy puddles of Outpost. 

Hitting 60 FPS on next-gen was a goal from day one. We always wanted as much parity between consoles and PCs as possible. This is a competitive game, and it's essential for a lot of gamers (myself included) so that cross-play feels fair. 
What excites the team most about developing games for next-gen?

Willans: Framerate, resolution, and ray tracing are the obvious answers, but the extra power means that we are less bound by technical restraints, so we have fewer scenarios where a game design cannot be achieved because we cannot support 30 AI at the same time, for example. I don’t think we’ve fully embraced it yet. We had to deliver a great experience for PlayStation 4 and Xbox One, so we still had one foot in the past, and for good reason.  

Maybe that’s a question I could answer with more authority on our next project. 

How do you plan to evolve Hood: Outlaws & Legends after launch? 

Willans: There’s a year's worth of content already planned that includes some incredible new cosmetics via the Battlepass, and we are in full production on a number of new features. Much of this is still under wraps, but I can say the fifth character, Eidaa, is going to really shake things up when she arrives. Our stunning new mountain map will be arriving soon, and we also plan to deliver a new PVE mode (by popular request) that rewards XP and game currency for those looking for a more co-op experience.  

We are also digging into player feedback as we design new features, game modes, incorporate bug fixes, tune, and balance to ensure that we address the right things at the right time for our community. The journey has only just begun, and we’re committed to building Hood into the game we all envisioned it could be. 

Want to learn more about Hood: Outlaws & Legends? Check out the game's official website, Twitter, Facebook page, and Epic Games Store page

    Get Unreal Engine today!

    Get the world’s most open and advanced creation tool. 
    With every feature and full source code access included, Unreal Engine comes fully loaded out of the box.