Case study: The level design of Frog Out!
Updated: Nov 20, 2019
In the Frog Out! -project I worked as a level designer and artist. My responsibility was to design and implement levels that explored the interesting aspects of the game’s main mechanic: eating and shooting the other player frogs around. To me this mechanic felt simple and intuitive. It also made many different game modes possible without adding anything extra for the player to worry about. Need a football game mode? The frogs are the footballs. How about projectiles to break down walls in the arena? The frogs are the bullets. The possibilities were endless. The only thing against us was time.
About levels
Due to the mentioned time constraints we settled down with three different levels: pond, factory and street. In the pond level the goal was to break the walls of the lily pad down by shooting fellow frogs against them, and once broken down, throw your friend off the lily pad to be the last frog standing. This level was the first one that came to us, and the concept had been proven successful in many other party games. We wanted a level that would use the idea of breaking objects in the level using other players as ammunition, at it seemed to fit in this level naturally.
The factory level is all about territory control. Periodically a hydraulic press-like stage hazard slams down on the stage, and all the frogs left underneath are taken out of the game until the next round. Before the press comes down, a flashing indicator on the floor shows the players what area they should avoid. The goal then is to keep the safe area under control and throw the opposing player under the press. Believe it or not, this level got started after joking about making a battle royale-mode for the game. In that mode the players are fighting for the control of a small safe area, much like the factory level in “Frog Out!”.
The Street level is almost self-explanatory: what is a frog-themed game without a reference to the infamous Frogger? Instead if one frog trying to cross the road, this time up to four frogs would be fighting to be the first among their friends. A race to a goal with the mechanics of the game sounded immediately fun me: imagine a mode where everyone is desperately trying to be the first, meanwhile being constantly messed with, eaten and tossed backwards, under obstacles and so on. It felt like the level with the most potential. We decided to make the level flow horizontally to keep the frog size on the screen constant and improve the readability of the level. Level had two busy roads with cars on the either side of the level, and a classic river with moving logs going to both directions. The cars worked as stage hazards, that the players needed to avoid, or could throw each other under them. To give the players more tactical choices, I split the level into two lanes that connected back together close to the end goal.
Mess on the Street
Whether it was beginners’ luck or great design by accident, the first two levels, Pond and Factory, ended up working great from the get-go. Both were immediately understandable, fun to play, and created exiting gameplay moments. In hindsight, I would’ve tweaked the Pond level a little by rotating the lily pad by a certain amount between each round, or by making every wall of the level breakable. This would’ve helped make the level feel less repetitive in the long run.
Despite all its potential, the Street level didn’t end up working nearly as well as the others. There were a multitude of reasons to this. As the level is much bigger than the others, the frogs appear smaller on the screen, decreasing readability. This wouldn’t have been problem if the game were a single player experience, but the whole game mode relies on seeing well what the other frogs are doing, and because the game is a local multiplayer game, the players tend to be sitting farther away from the screen than usual.
The game is at its best, when the players are interacting with each other in various ways. Following this logic, the levels should encourage this as much as possible. Unintentionally, the street level seemed to do the exact opposite. The level was big, so by nature there was a lot more distance between the player characters. As mentioned previously, the level was split into lanes, to give each round of the game more variety. Instead, all this decision did was further separate the player from each other. The split level might have worked better with more players, but as we had the 4-player limit, this decision didn’t work well for us.
Even when the players interacted (usually within the first couple of seconds of each round), things didn’t go perfectly. The speed of the hazardous cars was adjusted to make the crossing the road difficult, in the spirit of Frogger. This encouraged the players to focus on surviving themselves, rather than making it more difficult to others. Even when a player managed to eat someone, it was very difficult to throw them in front of a car. Mostly they ended up throwing them quickly in a random direction, being afraid of being squashed themselves.
Unlike in the other levels, if a player dies, instead of being out of the game, they would respawn where they started. I thought this would help make each round an exciting fight to the end. Even if you died, you could still have a chance at winning, or at preventing someone else from winning. This didn’t end up working in practice. Almost every round, one player would survive the initial chaos by a chance and amass a big lead. The other players, either respawned or left behind, had no chance to prevent the player in lead from winning.
There were also lessons learned about player focus: all they do is follow their character. Because the action in the other two levels always happens close to the center of the screen, the players can follow quite easily what was going on. This wasn’t the case with the Street level: because frogs are moving horizontally across the screen, the attention of the players needed to be spread to a larger area. Almost each time somebody crossed the finish line and the round reset, half of the players were confused about what happened.
All these problems combined created a level that was unclear and chaotic in the worst ways. The players seemed to be confused about what they were supposed to be doing. Clearly, something needed to be done about how the level worked. I needed to encourage the players to play the level in the way we wanted: in the most fun way. Luckily, there was enough time in the project for me to rework the level. Could I finally experience the magic of iteration firsthand?
Reworking the street level
I started with adjusting the size of the level. Taking out the water element from the middle was a painful thing to do at first, I felt like one of the essential elements of the level. I had to come to a realization: there simply wasn’t enough room. I experimented replacing one of the remaining roads with the river, but it didn’t work in either place. At the left it would create a bottleneck, and at the end it would make stopping the leading player too difficult.
Now that the level was smaller, the static camera could be brought closer. The size of the player characters on the screen were now uniform with the other levels, and it made the action more readable to the players. As the distance to the goal was much shorter, one player could no longer a massive lead.
Next thing that I did was removing the multiple lane design of the level. Even with the smaller size, the multiple lanes that the players could take still spread them too thin. Having a single path through the level really pushed the players to interact more, resulting in a better experience. Another thing that needed doing was majorly slowing the car obstacles down to be more in line with the design goals of the game. With the cars being slower, the players could focus on throwing their opponents under them rather than getting constantly crushed trying to cross the road.
Now the level worked much better, there was still one major problem remaining: the player with a lead would always win the game without exception, unless they walked under a car themselves. The goal of the level also still seemed to be unclear for some of the testers. These problems ended up requiring more work that the rest combined and needed a new feature to be programmed.
The solution to these problems was adding traffic stoppers to the level (the cylinder ones that rise and lower in the ground). The top of the cylinders have arrow signals that change color from red to green, in order to make clears to the players whether they are lowered or not. The cylinders appear in three choke points on the level, one right at the start, one in the middle, and one at the very end right in front of the goal. The obstacles work on timer, and until it has run out, the obstacles stay up, preventing the players from passing. The purpose of these obstacles was to prevent the players from escaping beyond reach, as a catch-up mechanic for the players left behind, and to get the players close to each other in clear intervals. The first stopper opens the same second the round starts, and only its purpose is to show the players how they work and set the goal for them, even if it’s their first time playing. I set the timers so that even if the players die right in front of the final barricade, they still have time to get back before it opens, giving them a fighting chance.
After this final addition the level worked as it was intended, with good amount of player interaction, readability and fairness. At this point it felt like it was meeting the high-quality standards set by the other two levels. This was apparent in game testing as well: instead of moving back to playing the other levels after a few rounds like previously, after the changes the street level kept the attention of the players considerably longer.
About team play
One thing that we didn’t have time to implement was team play, which in my opinion is a huge missed opportunity. In the current game every player is against each other with no co-operation. It takes a lot of depth away from the gameplay.
Think about it this way: when the game is in free-for-all mode, the players only are motivated to help themselves or to prevent others from winning. Helping others or self-sacrifice aren’t viable strategies. However, if the player has one friendly frog on their side and they are competing as a team, their tongue, previously offensive-only tool, gains supportive abilities as well. Here’s some examples: Is your teammate in mortal danger and hasn’t got enough time to jump to safety? You can fling them away from danger by eating them. Is an enemy frog in the lead, and the two of you have no way to catch up to them by jumping? You can launch your friend to victory, as spitting is way faster than walking. Are two of you playing against a skilled player, about to be crushed under a hydraulic press? One of you can eat the opponent, and purposefully get squished as the other escapes to victory.
As you can see, the strategic depth of the game would have multiplied, resulting in more funny emergent moments during the gameplay. After all, isn’t that what Frog Out! is all about?