Duration: 0:54

1,863 views

49

This is about 14 months old now, but I'm only just now learning basic video editing skills. Thanks to Dimon12321 for uploading it to YouTube previously!

Still probably the most frustrating TAS I've done. Co-op massively exposes the volatility of Dooms RNG. All inputs for all players were entered manually, and when one player shoots or moves, RNG changes, changing monster behaviour and damage caused by everything else. Therefore, if I built a decent portion of movement for one or more of the players, there was a high chance that when I moved another player, the progress I already had was now desynced and invalid.

This also meant that I had a lot of flexibility in being able to perform actions with various players to have more control over RNG manipulation (for example, shots across 4 players could be timed and controlled to affect RNG so that the barrel hiding the shotgun in the first room could be destroyed immediately AND the shotgunner in the east room could be killed in two shots), but this of course massively increases the complexity and difficulty of the building process.

So I had to build this EXTREMELY slowly and carefully. A couple of frames for one player, switch to another, do a couple of frames, etc. I had to constantly extensively check that things were actually going to plan, and often I realised my plan had to change because efficiently routing this was really hard, so scrapping progress was much more heartbreaking than it is usually. So even though there are four players, it was FAR more than 4x as long and difficult to do. It's not perfect, but I'm pretty happy with it!

If you're wondering, extra items appear during multiplayer. So no, you're not going crazy, you haven't actually been failing to notice a rocket launcher in e1m1 for over 25 years. :-)

__________
Original txt file:

Comments: Well, this was hard.

I admit I wasn't quite prepared for the frustration this would cause, and gave up on the idea of optimising movement for each player after about a day of working on it, having to re-do little bits and pieces all over the place was time consuming enough as it was. I'd imagine 7 seconds isn't possible, so having gotten 8 I'm not too fussed about squeezing out every last tic. There's a few bits of movement that look particularly ugly in XDRE, and some movement decisions that mightn't seem entirely logical, but happened to be what worked to produce the luck I needed so they stayed. Maybe I'm being a little harsh because having said all that I'm still pretty happy with this, I had no real idea of what could be achieved when I started but hoped to get 9.xx so 8 is good.

I'm willing to bet that there's better routing ideas. I'm also willing to bet that I'm not gonna try this again. This was a demo I'm glad to have built but it's also given me the knowledge that I'm not interested in building co-op demos, it's just too frustrating for me.

Player 1 - Kills one of the shotgunners in the west room, runs outside, grabs the rocket launcher, kills two enemies each with two rockets, notably two of them being the imps on the raised platform. Also shoots enemies with the pistol, ended up staying in for luck manipulation.

Player 2 - Grabs a shotgun at the start, kills the southernmost shotgunner in the western room, does some cleanup following Player 3, grabs the two eastern secrets.

Player 3 - Does a hasty nyoom to the exit. I didn't do the x.000000 trick at the end door because it wouldn't have made a difference to the final time since I needed to wait for Player 4 to kill the imp anyway. More importantly, it was 4:30 in the morning and I was going to fall asleep waiting for brute force.

Player 4 - Kills some shotgunners in the west room, opens the hole in the wall, kills the shotgunners in the secret tunnel, kills some enemies in the second last room, kills the end imp.