Ideation & Prototype
The game started with a very unique concept for the market at the time, consisting with a few key elements:
-
Huge Mechs VS Monsters battles.
-
Real time combat, but slower and more methodical almost turn based feeling.
-
First person perspective from withing the cockpit with diegetic controls UI.
Inspired by one of my favorites Anime Neon Genesis Evangelion, AAA games like Metroid Prime and Nier Automata, and hyper casual games like Retro Battle and Mecha Run.

Mechangelion Idea Doc
Video from our very first test
at the start it was hard to find a studio willing to work on the project, as it was very unique and harder to develop compered to the typical hyper casual prototype, making it more risky.
eventually we landed on TsyaTsya a small studio of 2 brothers from Ukraine, which fell in love with the idea and gave it an amazing execution.
the initial prototype consisted of the main gameplay and 3 unique monsters (each is a level) looping.
The first test gave us moderate results of 31% D1 retention and 0.7$ CPI,
With a short iteration focusing on the polish of the game like:
-
Better feedback and juice on hit
-
Armor that gradually break on hit
-
Dinosaur teeth break on hit
-
Better looking city environment
The results where a massive improvement in KPIs with 45% D1 retention and 0.35$ CPI, reaching the initial testing goal and letting us proceed into advanced iterations.

Adding Depth
Very early on my goal was to add more depth to the combat system, the aim was to keep the game simplicity while adding depth to increase the skill ceiling keeping players engaged longer.
We did that with a few things:
Parry -> Objective: adding risk reward
-
Player can hold a shield before an attack but pressing it just as the attack hit will result in a parry negating all damage and an enemy stun animation.
-
this adds a risk reward layer, by waiting to the last second of the attack the player risk getting hit but get a huge return, this also require the player to learn all of the enemy different attack animations to properly parry.
Shield energy & break -> Objective: discouraging passive play
-
holding shield consume its energy, when depleted it enter a recharges state until it is 50% full.
-
on hit energy reduce, if it reach 0 while getting hit the shield will break and player will receive full damage
-
By adding a Shield Health layer, we turned the "Block" button from a "panic button" into a resource management tool, discouraging turtling and encouraging aggressive play.
Different weapons -> Objective: adding loadout strategy and different play styles
-
we added weapons with different behaviors, i.e.:
-
minigun that shoot bullets while holding it until the bullets runs out.
-
charging plasma gun that the player can hold longer for more damage.
-
sword than can hit multiple targets by slicing everything in its path.
-
-
adding variety and unique loadout strategy.
Enemies -> Objective: diverse enemies to learn and try different strategies on
-
adding different types of enemies like the horde enemy that consist of multiple small enemies and bosses that have weak points that you have to hit.
Attack Types -> Objective: clear visual language for player pattern recognition
-
enemies attacks are split into 3 main types:
-
Normal: 1 hit, medium damage, can be shielded, dodged or parried
-
Multi hit: number off hit depending on the attack, each doing small damage, player need to hold shield to block the whole attack.
-
Strong: 1 hit, has a red VFX indicator, cant be shielded or parried, has to be dodges
-
All those added more depth to the game while the basic controls stayed simple, letting a variety of players to enjoy it, and keeping the more "hard core" players playing for much longer.
Progression - Planet System
In terms of progression from the start it was clear that just going with the standard linear level progression would not work, every level is a small boss fist and creating hundreds of those would be too time consuming.
that is when I came up with the circular planets system, at the start the idea was as follows:
-
each planet has a few types of monsters
-
monsters spawn on the planets in groups, when player complete a group a new random group is spawned with gradual increase in health and damage
-
player earn medals by defeating monsters
-
when player earn enough medals a boss will spawn and by defeating it they will unlock the next planet
-
player can return and play on previous planets after completing them
this system solved our main problem by creating a non linear progression where players can repeat encounters while working towards a bigger goal
Planet Transition
this system proved very successful in terms of game metrics and player feedback but we found out it had 1 major flaw which was: it was too non linear
-
difficulty of monsters were dynamic
-
stages at which monsters appears was random as well
-
it was impossible to read and understand level data as it was all over the place
so the next thing we needed to figure out is: how to make a system which allow us to reuse content without making it feel repetitive while being able to have full control over the level experience and data.

Circular Planets

Linear Planets
After a lot of of discussions with Yotam (my head of game design at the time) we landed on a compromise of a linear planet system:
-
the framing is similar groups spawn in group that seems random
-
but in reality all monsters types and states are all pre determined and linear
-
player cant revisit planets after a planet is done
-
player still has a choice which enemy to fight first within a group but it is much more limited and less dynamic
The advantages of this system:
-
player experience still feels dynamic and non linear
-
player still works toward a bigger goal while repeating content
-
we have full control on the player experience without random variables
-
data is easy to follow and understand
This hybrid system allowed us to scale the game's content almost indefinitely using a pool of monsters while maintaining a hand-crafted difficulty curve for the player.
Playtest - Problems & Solutions
Throughout development we conducted multiple playtests, in this section I will detail the problems that we found and how we solved them:
-
Problem: Players dont realize they can hold the shield and just press it, and dont understand what to do with multi hit attacks.
-
Cause: Shield stays up for a about 0.5s after pressing enough to block normal attacks.
-
Solution: Make the shield go down immediately on release requiring the player to hold and giving them immediate feedback.
-
Results: Players now understand very early how to hold and block even longer attacks.
-
Problem: Players dont realize they can aim the attacks
-
Cause: until the boss level aiming is not required and very easy to pass without it.
-
Solution: Added a "core boosted" level where one part of the enemy is highlighted, and breaking it gives the player extra currency.
-
Results: Players now try to aim earlier at the highlighted part to get the reward and learn the mechanic naturally, lower churn on boss levels.
Player learning to aim at the core boosted level
-
Problem: horde level is overwhelming and feel very hard and chaotic
-
Cause: every enemy in the horde attack randomly separately creating a non stop attack wave without time to counter.
-
Solution: sync the attacks so the go in short wave giving the player time to attack back between wave.
-
Results: Horde level is much more manageable, players feel less overwhelmed, less churn on horde levels.
Old horde level playtest
-
Problem: Players try to block strong attacks with the shield.
-
Cause: Players dont always notice the red indicator, and even if they do they are not sure what to do about it.
-
Solution: create the separate tutorial level prompting the player to block normal attacks and dodge strong attacks.
-
Results: Player now have a better understanding what to do with each attacks.
Rapid Fire Section
Monetization:
-
Goals: Design a more casual feeling monetization that push player toward IAPs.
-
Solution: Loot boxes based on equipment with infinitely scaling merging, together with infinitely increasing difficulty, to create a constant need for the player to get more loot to progress.
-
Results: The highest IAP rate of any game in the company at the time of around 20%.

Mechangelion Shop
Raid Boss:
-
Goals: Improve long term retention, create more of a "co-op" feature where player work toward the same goal
-
Solution: Fake raid boss as an event that triggers on weekend, with a huge health bar that decrease over the weekend as players fight it.
-
Results: improved D7, D14 retentions, players play more on weekends.

Fake PvP:
-
Goals: create competitiveness, showcase different loadouts, push player to get different skins and weapons.
-
Solution: Fake PvP system with different AI behaviors, loadouts and skins, together with a fake match making screen leaderboard and usernames.
-
Results: Very few player figured out the PvP is fake, skin shop is very popular among player and overall increase in player engagement.
