I was looking through the marketplace and saw a map for an absurd price, I read that the map creator used World Machine to assist in creating a realistic landscape. With more research I found this tutorial.
As you can see this was the final output, I created a texture with a series of different images sourced online which linked into one texture material as you can see in "terrain". This was the tutorial I used.
Once I had the map settled I begun experimenting with 'spawn actor class' a way to spawn objects for example projectiles in realtime. This is the result of me playing around and experimenting for a bit. I also did some experimentation with putting an enemy inside the map but this resulted in a very clunky enemy because of the over simplicity and also complications with the nav mesh bounds volume. Tutorial here.
When I learnt more about projectiles and conducted testing I found out that when collision was enabled on projectiles they could actually be ridden by the pawn. From this an early game play idea was to have skill that could be ridden by the player.
With more experimentation I moved into finding a use for the projectiles I had created earlier. Obviously I wanted to fire them at players however they would normally head on a straight path which was quite boring. I liked the concept of a homing projectile so I decided to look into that and see if it was viable for my game. Turned out that the actors couldn't really be casted as a static mesh so that failed. The tutorial for homing projectiles.
My first experimentation with using the HUD from this I created an MP bar and restrictions on some skills. I tested this with my projectile to have my mana drain whenever the projectile was fired. Tutorial here.
Implementing an HP bar was basically the same as the MP bar but just had damage applied to it through the 3rd person character blueprint.
With a basic knowledge on how to use UE4 I begun to look at game objectives. One such objective was to capture bases. I tested these by spawning an emitter at the trigger boxes location. These had a few bugs with them which I fixed later on such as reacting to everything that passed through them.
With an objective I begun developing a map more appropriate for what I wanted. My first ideas were to have the map change depending on which trigger box the player was in so the player could tactical cut their enemy off. This idea was ditched due to its over complicatedness in terms of map design and balance.
Above is my experimentation with changing the textures of a player in real time. I figured that if a player was using a skill but couldn't see or hear anything, the skill might as well not exist, half the fun of games are the interesting effects that come a long side them.
This was my first playable game, I put this under review with my peers who quickly shot it down due to its numerous problems such as glitchy and immobile AI (due to problems with too many sharing a behavioural tree), and a broken character.
From there on out I reinvented the game based off advice given to me and what direction I wanted to head into. This included changing the map arrangement to a series of rocky bridges between pillars. Fixing a countless amount of bugs so I could get my goblin as close to a boss that I could make it. I also tried to maximise on different elements such as the game being stealth based so changing the map to have rocky plates for hiding behind and the goblin to react in different ways.