Alwa's Awakening, an indie game released in 2017, just recently released a complete NES version that adds more content to the original game. What I find interesting is that despite not seeming to have some of the glitches found in the original release had and having more content to play through to reach the ending, the current NES record as of this post is around 4 minutes faster than the original release. I guess the reduced amount of load times, faster screen transitions, and reduced screen size is what makes for the time disparity. The main bread and butter tricks of block jumping and bubble hopping are still around so a TAS of this game might be interesting enough to explore.
itch.io page for buying Alwa's Awakening NES.The current world record of the PC version in 38:03.The run of the NES version that I found in 33:51.
Joined: 4/7/2015
Posts: 331
Location: Porto Alegre, RS, Brazil
Oh definitely! I've played the beta when they were porting it and immediately thought its TAS would rock! I could definitely help making a Lua script and an atlas encode for it, would be beautiful.
Games are basically math with a visual representation of this math, that's why I make the scripts, to re-see games as math.
My things:
YouTube, GitHub, Pastebin, Twitter
Did a brief bit of TASing up to getting the map and I found a few bits of tech.
Jumping adjusts Zoe's horizontal position in whatever direction she's jumping in very slightly. If given a straight path across a room, every jump will clear the room sooner by one frame.
Soft resetting via the Reset button on the NES rather than going to the pause menu and quitting back to the title screen saves 22 frames, but for SRAM to be updated the soft reset has to be done 6 frames after pressing Start to enter the pause menu.
Resetting after getting the magic staff saves about 37 frames compared to climbing back down.
Grabbing onto ladders and releasing from them when moving horizontally gives a small boost to horizontal distance traveled.
Made a work in progress TAS up to reaching the Shrine of the Sea Monk.
Link to videoUser movie #637925661270209087
Block jumping is done by getting on a block in mid-air by summoning it at the right time. That timing for the most part seems to be summoning the block 3 frames after releasing A.
EDIT: Turns out that bit of movement that I thought was unoptimized ended up being faster without a ladder boost by 12 frames, thanks to the block spell cooldown getting in the way.
Here's an encode from the start of the Shrine of the Sea Monk to picking up the Bubble spell upgrade. I think a fair amount of this is pretty cool.
Link to videoUser movie #637926753949437295
Something worth noting is that the cooldown for the bubble spell is 24 frame longer than the cooldown for the block spell.
Oh cool, an out of bounds trick was found that only needs Block 1 (to get bubble) and Bubble 1 (to perform the zip at map location 3, 11).
Here's a video timestamped to said out of bounds glitch.
I think I get the steps of how this happens.
A bubble is created near the ceiling at the screen border but not too close to the top.
Zoe jumps on the bubble when the bubble is at a high enough point such that the bubble still goes up but Zoe is able to stand on where she would normally stop the bubble.
At a certain point pressing a direction will cause the game to eject Zoe out of the wall, but does so in a way that she gets screen wrapped to the left side.
Once that's done it's a matter of navigating the out of bounds areas for about 2 minutes which is apparently full of ladders. I might investigate this after finishing up the run I have that doesn't use this glitch. Thanks to Rayquaza911 for messaging me about this glitch.
Here's my progress so far without the newly discovered out of bounds and zip glitches.
Link to videoUser movie #637927613920656566
Void Tower was interesting to optimize with figuring out the best way to manage spell cooldowns.
To clarify on this, it's a screen wrap akin to the one in Zelda 1, that can only be done on the right side of the screen and you can only wrap to the left screen with it. If you happen to land in a wall, you can zip in whatever direction. If there is no room to your left you will be brought out of bounds where every tile is a ladder.
To do the wrap, you use the bubble on the very edge of the screen, you know it's working when the bubble travels through Zoe. It can be done on any screen, you don't need a ceiling, in fact it's easier without ceilings above you.
Thank you for this info clarifying the zip glitch process. Having that info available in any public form helps a lot!
Meanwhile, I've cleared all the old content sans the final boss of the original 2017 version. From here on out it's the exclusive NES content before this is finished up. Sadly I couldn't get the one cycle for that Altar of Echoes room. There's also concerns about the damage route I have right now but I don't think it's gonna be more than a few seconds at best.
Link to videoUser movie #637928477120291987
While working on the Any % Glitches TAS, I have noticed that the animation cycles for walking and climbing are quite important and sometimes push me a single pixel forward, especially when getting off of or jumping off of ladders. There also seems to be some weirdness when going down a ladder from the top for a single frame and then holding left or right. Sometimes you fall down the ladder, but sometimes you continue walking on top of the ladder. Both are useful for specific spots.
Also, my TAS only relies on RNG for the last boss, but I got confirmation from the developer that RNG is at $A2 (figured out from there that it's just a 24 bit linear feedback shift) and it updates when you kill an enemy, every frame of the title screen (including the company logos at power on), and every frame you are in a save room. I figured out on my own that the RNG resets to 0x010000 on console reset. I don't know if killing a boss modifies the RNG or if being in a boss room modifies it at all.
My TAS gets the rod before yours does and I'm curious exactly what mechanisms are causing that. I think it's based on how I jump off of the ladders and possibly the animation frame I get to the top of the ladders on. I'll try to get something up soon so things can be compared so we can try to figure out that mechanism to make both of our TASes faster and possibly more consistent.
I only really caught it at the end of the TAS I finished but I think the ladder dismount frames being 1 or 2 can potentially be changed by how long A is pressed for jumps prior. I think there's some sort of weird acceleration factor or air cycle going on with that.
So I'm in room 07,13 at the moment and getting drastically different results on the second ladder depending on what happens with the first ladder, so I think the ladder animation frames/counter are tied together. If that's the case, then it's likely the case between rooms as well.