Honestly, this didn't do much for me. There's a fairly strong middle section that utilises the flashlight for some fun and tense scenarios, but the opening and closing sections are quite dull. Alyx is a fun character, but her survivability feels a little ridiculous. Sure, Gordan shouldn't have got through the original Half-Life, but as the player I've seen that his success was only possible with the loading of MANY quicksaves, and thus the unlikeliness of his success is properly conveyed and justified. Alyx's invincibility renders her a non-factor, and as such it's hard to buy this as a joint adventure.

Unlike the original Half-Life, I knew what qualities to look for in Half-Life 2. This was the showcase for Valve's Source engine, which would form the basis for their prolific and acclaimed output for the next decade, and one whose advanced physics and malleability allowed for extensive use in the modding scene, for which the accessible Garry's Mod was a catalyst. As a teenager I may have played few levels of Left 4 Dead at a friend's once, but otherwise my only prior interaction with the Source engine was through Portal and Portal 2, both of which I adore. Otherwise I was console-based, only getting into PC gaming in adulthood, and even then not owning anything more powerful than the relatively cheap gaming laptop I currently use.

Finally playing Half-life 2, it's so clear why it was such a huge deal at the time. The physics engine allows for each object to serve a mechanical purpose, and this is used to great effect for puzzle solving and exploration. The level-design and incorporation of physics are frequently inspired, and consistently elevate Half-Life 2 even in its weaker moments. Using buoyant containers to lift ramps, or holding a car door to shield yourself from gunfire, are examples of developer-expected uses of these mechanics which prod the imagination towards a world of less conventional, less "intended" possibilities. The physics' looseness suggests that solutions to the game's challenges may be less prescriptive than they appear on the surface.

I think what ultimately limits the game is its format, the linear first-person shooter. Don't get me wrong, the physics wonderfully compliment the level-design, but the straight path of the game's progression keeps truly ingenious manipulation from yielding substantial gameplay feedback. An immersive simulation style of game made with craft on display here in Half-Life 2 could have truly shown off the potential of this engine.

Taking the game for what it is, I think it's a great first-person shooter, albeit one with serious issues. Keeping everything in-engine from Gordan's perspective makes a welcome return, and in some moments dramatically outclasses the first game. An early section where Gordan must escape a house raid is exceptional in its pacing, building of tension, and drawing the player's attention to the narrative beats of the set-piece. I could only imagine how this sequence would play out in a modern Naughty Dog title, wrestling control of the player's movement and camera towards every important detail - I'd likely begrudgingly accept that, to achieve this kind of moment, sacrificing control might be a necessary evil. But here's Half-Life 2, every bit Naughty Dog's equal in pacing out such a set piece, but many times more effective by keeping it in the player's control. When in control, there's real tension, because I could have exercised that control differently. In contrast, if my movement is halted before my attention is forcefully drawn towards a door getting kicked in, I know that I have time to escape the armed men pouring through, as I'm reacting in what I know to be the optimal moment; I'm seeing the designers' intentions as I play, rather than just being present in the scenario.

However, I think perhaps the main issue with Half-Life 2 is that it doesn't know when to stop. Fresh ideas are in abundance here, but there's just too many of them, and the pacing suffers. Take the vehicle sections - the stops along the way are always interesting, but the structure overstays its welcome. This problem isn't so bad until later levels where the core idea that's being iterated on isn't that enjoyable in the first place.

Larger battles with many combatants feature more prominently towards the end, and are the worst example of a bad idea being iterated on endlessly. They'd often put me in that unfortunate place of trying to understand what the designer intends of me; Half-Life is strongest when it is player-driven, but in these larger battles I'm often left pondering a couple of questions - "Are these enemies continuously respawning because I need to move forward and just haven't figured out how? Or can I not see the way forward because after I kill enough a loose rocket or something will explode some rubble, clearing a path?" The answer was almost always the latter, but the ambiguity meant that I was never quite sure, and the causation between my actions and the circumstances that allowed for progression were atrociously contextualised, and thus the player-driven design is lost. It's here that Half-Life 2's priorities backfire, as when this design contrivance occurs in, say, Bayonetta, magic red barriers appear out of nowhere and it's extremely clear what needs to be done to make them go away. Half-Life 2 can't have magic red barriers, so it probably shouldn't have ever adopted this style of encounter design.

Still, Half-Life 2 is superb a vast majority of the time. If it had trimmed each level down to its best two-thirds, and maybe one-third in the case of some later levels, then I might regard it as the masterpiece its often hailed as. As it is, Half-Life 2 still stands tall among first-person shooters.