Half Life 2: Conclusion


  • Hammer world editor

  • Created in 2 months, half-time

  • Built with assets from Half-Life 2 and it's episodes

  • Created while attending The Game Assembly


During the fall last year we were tasked with creating an entire HL2-campaign together with the others in the whole class. We were given our own mission in the story progression, and we're quite free with what we were doing in that level. It only had to somewhat follow the story, and we couldn't use weapons or AI that wasn't previously introduced to the player. I was supposed to introduce the Headcrab Canisters.

As the level is in a combine-controlled area I was not able to use the normal zombies, but thankfully I could use the combine-version. This gives the level a mix of normal combine soldiers trying to fight of the headcrabs and the other combine soldiers that have been turned. 

The level is the end of the campaign, so it also has an ending. In th ending the player gets teleported to a cellar where an advisor is resting in a pod. The player destroys the pod using energy spheres to kill the advisor who is controlling the area, but the advisor escapes. Then the game fades to black, and the credits roll.

Headcrab canister in the orignal HL2 campaign

Headcrab canister in the orignal HL2 campaign


A different start

The game started out very different to the end result. This was because of two reasons, one being the way that the engine and the editor worked. The planned gameplay and ending wouldn't work very well, due to the size required of the map. The second reason to why the map changed a lot was due to the level before me being cancelled, forcing me to redo the beginning. I also had less tools to use, that where originally introduced in that level.

Halfway through the project the players path was quite messy, and it confused a lot of the play-testers on where to go. I had a lot of different elevations, which was also very confusing. Turns out most players only look right and left, and often ignore up and down. I then did an overhaul to the map, and simplified the level a lot. I used the areas where people thought they were supposed to go as new paths, so it all felt very natural.

I also went over a lot of the bigger buildings, and made them flow more with the rest of the level. I removed bad corners and small pillars etc.


In the final version of the level the path is very clean, and the players rarely have issues finding their way through the level. The objective, to open the gate, is also presented a lot better in my opinion, with it being raised from the ground and lit up with search-lights.


Here is the final top down map of the level, which was iterated and worked on a lot during the project. I chose to not show combine-spawners in the map, due to the amount of them, and how cluttered it would be. The apartment bit is disconnect from the rest of the map, to show that it is on another floor.  It could have been connected, but I think that it wouldn't be clear that it was above then.


Working with Bsp

Old version of the map, showing off the dev textures and bsp-brushes I worked with

Old version of the map, showing off the dev textures and bsp-brushes I worked with


Building interiors were hard at first, but I got better at it after a while. Half-Life 2 follows a very true-to-life way of making buildings, and this was a bonus for me as I could often use real life references for my buildings.

It was also not hard to create believable spaces, because of the many assets and textures HL2 has, that comes with the modding tools. I could quickly and easily create rooms and windows/doors, making iteration times faster.

I enjoyed working in hammer with the BSP tools a lot, because of the quick iteration times it gave me. I was able to completely rework my level very quickly, and it gave me a way to do fast iterations during playtesting. I was able to change the level and make a new build during the playtest sessions sometimes, and that helped a lot. 

It really felt that the editor and the tools were shaped around the BSP workflow, and I could after just a few weeks fluently work without having to do weird commands and look up how to do things. This resulted in a very nice workflow, and it was very easy to get into the flow of building.



This was a good project for me to work on because it allowed me to focus on iterations and just level design, instead of having to do game design and complicated scripting as well. I really enjoyed the process of working with BSPs, and because of the simple shapes that the BSP is built with I could quickly iterate. It improved my skills at working quickly with blockouts, and I think I grew a lot in that area. The simple-designed AI of Half-Life 2 was also easy to work with, and allowed me to quickly create battle areas to test out different ideas.

Various Development screenshots