Like any form of competition, speedrunning generates arguments over authenticity. Does a speedrun count if it relies on a bunny-hopping mod, in-game glitches and different runners tackling different parts of the game in short segments? I'm not sure I care either way. No matter the methods, Half-Life 1 being completed in 20 minutes and 41 seconds is an accomplishment of endurance, skill and effort. More importantly it's a beautifully entertaining video, full of ingenuity and grace and physical comedy. The new record time is embedded below. You must watch it.
As thrilling as the video is to watch, the details of how it was made are even more fascinating. The description under the YouTube video mentions that "after almost four years of painstaking planning, theorycrafting and execution, we have arrived at our final time, smashing all of our wildest expectations."
Theorycrafting? It seems this kind of speedrunning is deeply methodical. This Google Spreadsheet breaks down the timestamps and the gains in time made during each segment of the run. If you click the tabs along the bottom you'll find further information, including stats (only 28 enemies are killed during the entire run), weapon charts (the crossbow can confer as much velocity to the player as the RPG), and a step-by-step breakdown of the plan for each level in the game.
There is undoubtedly a difference between this kind of speedrun, and the solo, continguous efforts you can watch live on a site like . But I don't think that matters. Segmented speedruns have been around for a long time, are a thing unto themselves, and still take absurd skill and effort. Half-Life 21, as this new run is called, contains 317 segments, "249 of which are shorter than 5 seconds." That's impressive. It's nine seconds quicker than the last previous record made using the same techniques. That's impressive!
If anyone has contact details for quadrazid, or any of the makers of the speedrun, email me.