Posts for Meerkov


Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
To be fair, almost any run of LTTP should get accepted considering the game is fun to watch, and the state of the current "best" submission is really non-optimized.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
Oh I forgot to post a link to my encode. Added it now. https://www.youtube.com/watch?v=-lrvo4Xct78 No I didn't use any script, the rerecord count you see is all manual.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
The previous Genesis run used "Amateur" level. Mine aims for highest difficulty level too, which is "Detective". That's why mine is slower. That costs me about 11 frames. My frame count is 1672 compared to the previous run which was 1667. Therefore, my run is actually 6 frames faster during gameplay.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
"Game Over" actually means you won. If you press "A" one further time, it goes back to the SEGA screen. Does that answer your question? Maybe you're saying the bk2 is misformatted, it shouldn't have blank inputs at the end.
Post subject: What are the rules?
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
What sort of run are we even talking about? Trying to define which glitches are "minor" or "major" seems silly. For example, YBA gives arbitrary code execution (see the canceled Acmlm submission on this site, even). Time was ~4 minutes and isn't even optimized (and was canceled because it just skips to the end instead of doing anything particularly cool). Any% w/Major Glitch is more or less "solved". Time is <3 minutes. Showing off glitches is fun, but how do you make a run which doesn't devolve into "Skip to the end"? 100% with Major Glitches is easily done in code execution. Maybe you just ban YBA? (But this is no guarantee against arbitrary code, the game is really really really buggy, especially the japanese version). Maybe allow YBA, but just ban using it for code execution? (How do you define code execution?). Maybe you also have to ban Wrong Warps/door way glitches/wall clipping teleports.... Is the Major Glitch category simply too broken to have any meaningful run? [Edit: I started a 100% major glitch run. My time for EP->Zelda (clocking the second the sermon ends) was about 8:05. My estimates say that if you do Zelda->EP (clock the moment you leave after getting Boots) you would have a time of 7:20. In either case, you would EG as soon as possible to continue the quest. One issue I was debating was whether I should stop by to pick up the Cane of Somaria while running between EP (with boots) to save Zelda. It turns out it's on the way. As I mentioned, waiting to save zelda costs 45 seconds, so that run is a dead end anyway. However, I noticed that current RTA 100% speed runners abuse YBA to do things like: Kill all bosses while in escape glitch mode, and only return to the dungeon to pick up their diamonds. If YBA lets you do crazy things like this, then I had to take a step back and really figure out if it made any sense to start with Zelda at all, and not just YBA and do crazy things. But then why not just YBA and 100% the game with code execution... My conclusion was that (And you might disagree), the game is simply too glitchy to make even "no glitch" categories have any comprehensive understandable rules (which wall clips/teleports are allowed, and which aren't? Is it ok if you teleport 6 pixels, but not 60?..... The whole thing is incomprehensible. Edit2: And if you glitch between areas (escape mode, wrong warps, clips, teleports...etc) you can skip all the keys too. Despite what Tompa said on this thread 3 years ago, boss chests do open up even without the right key, and everything unlocks if you are technically in the "Cave" dungeon. But you can also use doorway glitches to confuse the game into thinking you're somewhere else, and thus open chests that aren't even on the same screen as you. Do you even need to go to the Big Chest screens at all, at that point? Playing with the game and glitches is fun, but the game is way more broken than anyone knew 8 years ago. It's a big problem. There is no way I can continue to work on a run without knowing the rules (Though I'll probably give up because it seems no rules can reign in this pile of garbage code).
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
For the record I don't intend to spend another second looking at Jaws. I didn't realize we reject videos that are improvements to the current record on the site, even when they have positive audience response...
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
Can you please post a video/input file of this? I'm really curious about it.
Yep I updated the submission so it now has a video and movie file associated with the bug (movie file is worth a thousand words).
Is your run any faster than this
Aw man you're killing me. I didn't realize this guy already had a TAS of this game ( watched his RT run while researching this game). Yeah his does look faster than mine. Not sure what the policy is here. Clearly I have room for improvement still, but my run is faster than the site record. I can try to reach out to them and see if I can get a copy of the movie file...
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
Hi Everyone, thanks for the encouragement! Let me reply to a few comments:
Somehow I had a feeling that when you were attacking old adelikat runs, you might get around to this.
:) Yeah I took a look at this run, realized right away I could trim about 9 frames off the end of the movie by timing the boat ram better, and decided it was reasonable to try for some savings.
Has anyone ever tried just killing Jaws on level 1?
Short answer: yep it doesn't work. Level 1 attack does 8, where as I am doing 64. If you only do 12.5% of his life per battle, you won't outpace his healing (25% of his bar after every fight). If you can RNG manipulate 13 shells from the first fight of 24 enemies (more than half the enemies!) then you could power up to 32 damage, which is enough to outpace the healing, and could maybe let you deal some partial damage while you build up enough points to earn the Sub. I couldn't figure out a way, though.
Does getting slightly less than perfect speed up the bonus round?
Unfortunately it's sort of like Gradius. If you don't kill the enemies, they will dance around for a few seconds before exiting the screen. You always get 30 enemies to kill so it would take longer to not kill them.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
NOTE: The actual run of this game is from 2005
Whoops, fixed the submission text. Thanks for pointing that out Challenger! And nope, I have no plans to do a 100% run of Zombies Ate My Neighbors, but thanks for the appreciation!
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
Glad everyone is liking the submission. Happy to be up for Moon status. Since I spent quite a bit of time playing this game I thought I'd offer my thoughts on the 1P v 2P discussion. I read the previous submission's forum posts, and I had a few thoughts on the subject. I'll give a few arguments for having only one any% category, here's why: This game doesn't really change much with 2 players. It reminds me of games where Hard Mode just gives enemies 2x the HP. In this game, you would need to collect 2x the items. 2P doesn't allow any new tricks. In Sonic 2, Tails can carry Sonic, which can really change the strategy. In this, a 1P and 2P run isn't very different. In terms of the guideline that 2P runs can often have more "action": This run is a full 8 minutes faster than the 2P run. I think that is very competitive when it comes to action on screen.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
@lord Yoshi, hah well actually a little bit:). I'd fix it except I'd need to redo the rng for the entire game. It's not completely for nothing. I do end up with 4 more bazooka ammo as a result of that path, which is not a great trade, but does help since I actually do run low at one point during the run.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
@XYZ correct. You can't manipulate the victims location. The way it works is you have between 1-10 victims on each level, equal to the number saved on the previous level (+1 if you get 40k points). Each victim has a number from 1 to 10 as well. So I you have 2 victims left alive, the next level spawns victim #1 and #2, which are always in the same locations, regardless of which the id number of the victims you save. This is the issue with using the boss skip glitch on the level 48. The #1 victim is always locked behind the boss door. The victim#2 is superduper far away from #1. You game over if you have 0 victims left, so you always have to spawn victim#1 which is behind the boss door.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
@Allbeert I'm glad you liked it! Sure, I'm happy to provide the script. I copy/pasted the values for Shoes and Skeleton keys directly into the script so I could highlight them in my HUD. If you use this script as is, you'll get the HUD I was using. To generate the drop table, I was using gatherRNG(), but it's not hooked up anymore. I provide no warranty, use at your own peril! http://pastebin.com/x6UfXhPw
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
First of thanks to the early voters, I'm glad this has a fairly positive response so far! @Spikestuff That's alright, but don't let the write up confuse you! I'd like to help explain. I have a habit of being a bit verbose, so I really tried to remove unnecessary over explanation from the post. Grassini is right. I think you are talking about the -304 frames from Lv 15, 21, 29, and 30, the only 4 levels I was slower than Allbeert on. Great observation. The explanation I didn't think would be very exciting to read, but Grassini is spot on: it's just me picking up items and manipulating RNG: Here's the breakdown: Level 15, I picked up 1 more item than Allbeert (-24 frames) Level 21, I picked up 1 more item than Allbeert (-22 frames) Level 29, I picked up 7 more items than Allbeert (-195 frames) Level 30, I picked up 3 more items than Allbeert (-63 frames) Total: 12 more items, -304 frames. Average of 25.3 frames per item. When doing this run, I pick up 70+ items, mostly from manipulating chests. I thought it would be too repetitive so I only included the more unique path changes, and difficult decisions. My whole strategy was to pick up waaaay more items. But it's a good point that these are the only 4 levels I lost frames on. So, why is spending 25.3 frames to pick up an item, worth it? I actually did a bit of math: The shortest level is now level 2, at 199 frames of input (it takes a bit longer for the door to open up, but the real hurry is to pick up the survivor). If we were moving at 2x speed, we would save 100 frames minimum. Every key I picked up on these levels could have been a shoe (I just would have picked up the key on a different level), so we can assume that these 12 items are all worth at least "1 shoe" worth of savings each. We know a shoe is a minimum savings of 100 frames, so spending 25.3 frames to save 100+ frames later is then obviously worth it. It's a bit mundane, but this was the guideline I used when planning my route, by seeing which chests I needed to go out of my way <100 frames for. I fully admit, it's possible I didn't select the optimal ~70 chests ( I mention it as a possible area of improvement), but I'm fairly confident in the route. I'll update the submission to mention this.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
Oh boy, how embarrassing. Yeah the times are swapped after level 5, but the right two columns(difference between runs) is correct. I'll double check all the data later tonight. Also, yes! I started a 2 player run, got half way finished, and realized 1p is much faster because you need to pick up half as many speed shoes; a big improvement. Even using two players, MechChicken's run has areas to improve, but the need for shoes forced me to start over as a 1p run. This is a very highly optimized run, with the exception of a few things, which I call out in the areas to improve. Level 1, I did make a mistake (doh!), but still was faster than the previous run. On level 3, I actualy did try both ways, using a shoe on 3 and 4. 3 was not much faster. I'll make that more clear I tried both. I did also a lot of different things, like I tried earning a bonus victim on level 47 to use the boss skip glitch mentioned in the forums here. It turned out to be about 100-200 frames slower than the 2P run. Then I had to go back, decrease my score, avoid taking unneccessary hits ( that allowed me to die quickly for the glitch), and then fight the whole boss, resulting in a reductions of 200 frames, a 400 frame swing, without even considering the damage and bonus victim message time. It's not easy to tell, but let me tell you, that spider boss is nasty!! We're basically underneath the boss using a melee weapon while it shoots AoE attacks and exploding bats orbit me. Phew it was a lot of work, basically going back 20 levels and redoing all the rng! While there are areas to improve, we're talking about maybe a dozen seconds, over the whole run. In most areas I fight for every frame possible, hugging every corner perfectly, and trying dozens dozens (like seriously took me hours) of variations when fighting the spiders to get the result I present to you today :)
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
Wow what are the odds? I've been working on a ZAMN for over a month (in secret whoops). Good news though. This record won't last for too long. I'm on level 48, and should be done soon. New tricks not used in this run too. I hope everyone isn't ZAMNed out by the time I submit. (edit: I will have to update my notes to compare my run against this new run, but I'm fairly certain for the record that there can never be a 2P run as fast as the 1P run. Stay tuned.)
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
@WarHippy: Actually I tried that, but falling is actually faster. It's actually a pretty neat trick. It turns out jumping from the windowsill to the shutter is really quick, because by falling, we actually are within the acceptable limits to continue moving without playing a landing animation. So I'm not sure if there were that many frames to save by this method (maybe a few)... but! In addition, due to having basically perfect pixel/subpixel combination, we can boost from the shutters directly to the signpost booster below. In every other attempt I made, such as jumping like you suggest, I couldn't get those extra few pixels or subpixels that I needed to trigger the boost, and would be forced to play a landing animation and subsequently jump to reach the spinning signpost. I tried a bunch of various timings, but falling was definitely the fastest as a result.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
@Mugg Playing on hard is what is better, not the lack of skipping. Some of the longer levels are boring (dance minigame) but others can be interesting (fish level). This is of course, the problem with trying to compare an easy playthrough to a hard playthrough. @Archanfel 3 seconds? :( I think it's partially because I couldn't jump onto the stingray near the end of the level (timestamp 11:06 you can see I try to jump on it, but on hardmode again, I jump a few pixels lower than on easy). As for the fireworks, Mmm yeah, I was having a hard time there. Once for example, I had both the 1st and 6th explode almost at the same time, and I couldn't pick both up before one would disappear. Tricky spot, but I did manipulate the ones I could reach in time. Likely room for improvement like you suggest.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
@arandomgameTASer there's definitely tradeoffs. Some of the levels are longer and boring (e.g. the dance minigame), but others have additional content (playing through the full level 1 and 7 and 8). The easy mode skipped major portions of those levels. It's hard to see, but I do actually use some new tricks that weren't present in the original run, such as clipping through certain objects, using the boosters in lv1 more efficiently by turning around before hitting them, and new strategy for fighting the moths. Some positives, some negatives to be sure.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
Well done finding a way to encode the video (I just figured it was a lost cause). Also I commend your efforts to sneak out those extra hundred frames compared to my run.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
Doh... I'm uploading the correct video now as well.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
Spikestuff, I certainly believe there is room for improvement. I'd estimate ~1/32 spins results in a "10". However, there is some other RNG which seems to decide how long the spinner takes to spin. For this run, I manually searched through a few hundred possible spins over the run. This means I occasionally got more than one "10" and chose the spin with the lower number of frames. Scripting or some understanding of the RNG would probably make finding good spins a lot easier than what I was doing.
Active player, Experienced Forum User, Published Author (397)
Joined: 10/4/2015
Posts: 98
@Invariel: Well I didn't try it but let's see: The path I took, I need to skip 1 level, and finish 1 level (actually they take about the same time). I need to spin 3 spinners. In your path, it would be 5 spins a quiz and a level. The spins actually are kind of long, so even though it would be 1 fewer level, I think the 2 spins + quiz load time would not work (those 3 would be greater than 3 level loads I'm fairly sure). I haven't tried though. @Dwedit I've played this on other emulators and the levels don't glitch out, but the game was not an official Nintendo Approved etc game. As a result I suspect they program things in some non-standard ways that not all emulators understand (bizhawk in this case). The time spent spinning does seem RNG dependent, but for such a simple game, I'm just not sure it's worth the effort for those extra couple frames.