New York, USAHey, as far as I remember there is no consistent manip, but sometimes the boss just sinks slow (ideal pattern, the movement up to boss might matter) It's good to come in lower than the boss and get really close and spam its feet, and hope it's good pattern, worth the risk to take a hit when the boss is lower, but when you have to follow too high its sometimes best let it go up and come back down again, then repeat earlier. not as bad as 2nd boss. which is way more complicated/random, unless you find something.
Getting the early platform cycles (Wolf level and Asylum) and the cemetery level are the hardest parts of the game and also last level. One thing you'll notice, at least for me, is that sword slashing enemies can be inconsistent, due to enemy overlap which means one enemy doesn't take damage, and also I suspect the game does not deal damage on some frames. plus i suck at fast slashes/mashing. Jumping in the cemetery made the level much more consistent and can be managed with little practice, plus throughout the game i want to avoid unnecessary damage because it screws up where i can tank and death abuse/warp later in the run.
In the cemetery, after the checkpoint (which i death abuse at to get my health back) depending on the two patterns you can get with the upcoming knights, you have to either jump on a grave stone and jump over or kill one and jump thru. Basically you can take damage from birds but avoid it from the knights due to knockback. The next flat section and the downhill section are easier and consistent and the last flat section before the boss is the hardest. You can also bail out and just kill the enemies if your jump timing is off at that ending section.
Hey. Unfortunately here are no tutorials for the Normal category afaik. I just copied existing runs and tried to improve upon any thing I could by trial and error. The Wolf Boss freeze I found by accident when I took damage as the boss turns around. You can set it up best in the beginning by walking to the right and then turn around at a specific spot and walk back towards the wall at the left and then turn back into the boss when it reaches you. I had maybe a 2 out of 3 success rate with it. You can do a similar thing with Renfield as well but the Wolf is much harder when you don't get the freeze.
Hi. We are open to any new category suggestions. In general, a proof of concept run is needed to demonstrate the particulars to that run. In this game's case, since it isn't linear there would be a lot of items/traps that I'm not even aware of all of them myself. Leracos and Numbers know much more about these things than I do, so can be much more helpful in defining the specifics to the run.
A lot of discussion about the game is done over at our discord which I linked below.
This is one of the level 6's I was getting frequently by early character selection. The level layout orientation might change but there is a door and key right next to where you start. Also level 1 always would start me in the upper left corner on this particular seed.
I was messing around a bit with OpenEmu using the Gambette core and the seeding works entirely different on Gameboy from the NES.
After some trial and error, input timing in the title screen/music part doesn't seem to affect the level seeding. The seeding seems to be the timing of when you select the character. After the title screen there is a second of delay you have to wait before selecting the character in that screen. Unfortunately you can't buffer an input to get the character select on the 1st frame available for easy confirmation. By mashing or very precise timing I was able to get some of the same seeds over again. By precise timing I mean frame perfect not like the large window of frames in the NES version.
I guess testing an actual GB would be important to see if it will behave the same way as emulator.
So I actually did bother trying to beat all levels (1 loop) in this game and it takes a while. I confirmed that after level 102 the game will loop so level 103 will be identical to level 6. Level 1-5 are always fixed and only in the beginning.
Since I have some completed runs I was wondering smoulder if you could add an All Levels category. Goal is to simply to beat all levels before they start repeating again. Run ends on exiting in Level 102. I also include levels 2-5 to give it a more 100% completion, I guess an alternative would be to skip to level 6 and just do a 1 loop but why not do them all.
There are also levels which are not numbered, Secret rooms occur every so often which can be avoided (they are huge time wasters) Treasure rooms seem to randomly occur every once in a while.
Congrats to Chardcore for the Level 10 sub 1:30 run and using the same seed as I did. This confirms that Mesen behaves the same way as the console in the timing to set the seeds. Nestopia is also accurate as well. I found out that Bizhawk and Fceux each have entirely different seeds than the console. I couldn't even find the Level 10 and Kill a dragon seeds that I used on Console/Nestopia.
The levels after 5 seem random but can be manipulated to a specific order. To get a consistent level seed you have to input (start, A or B) at a specific window of time in the title screen (You can used a musical cue for timing). In my notes I found a total of 39 different seeds and they seem to be in about a half second window of time intervals. The last seed (I guess the default seed) you will get after the title screen/music plays out.
I can't confirm that the seeds we use for each category are the best but based on playing all the levels they likely are.
lucandor158 brings up a good point I did not think about for different emulators. I only tested the moonkicks on keyboard earlier using Bizhawk and the L+R inputs at same time was not disabled. When I just tested with Nestopia v1.4 the L+R inputs was disabled by default and I could not find anywhere to change this.
Also, I feel bad and I hate to be the guy trying to discredit someone and like Slack said lucandor158 should be congratulated for what was accomplished is such a short time. And to be fair the rules were not clarified and I’ve seen this before where people's hard work was negated because of rule changes/clarity after the fact. At the same time all runs on the leaderboard shouldn’t have inputs that’s impossible to perform on console/original hardware.
So I tested out moonkicks on keyboard and it looks like Ouijawii beat me with the technical explanation to the keyboard advantages. But what immediately struck me was that when doing the moonkick with keyboard I can have the impossible (left+right) inputs held at the same time. I never have to let go of (left) when doing the (up+right). Having (left+right) held at the same time on a NES d-pad controller is impossible unless the controller is broken, that is why it is commonly banned on all NES games.
This is from the general NES rules:
Turbo/L+R Turbo is not allowed as it provides an unfair advantage. There may be a few exceptions regarding JRPGs, but there is an overarching rule that turbo is not allowed. Left+Right or Up+Down is not allowed.
https://docs.google.com/document/d/1apP69VDIzUx-iVO_sj5YjTxBAE2ukHPa6kl33AOW9hA/edit
To clarify I run on emulator using an original NES controller with USB adapter. While I don't use an original console I never wanted to run any games on keyboard. Though I'm now curious to test out the moonkicks on the keyboard.
Interesting and difficult stuff to consider. As an emulator runner, I guess my 1st question is there any other NES speedruns that ban emulator and/or keyboards users? I think there are a few obscure games that ban emulator because of RNG manips. I believe most games will just make the rules stricter like live streams, cam controller, input display, show resets, show no movie, etc. I rather have the rules worked out more as the game becomes more popular/competitive than to go for a simple solution like no use of emulator or separate leaderboards.
As far as using a keyboard, my opinion is that it can have an advantage for moonkicks in this game. And whether the advantage is slight, it is still unfair for original hardware users since the game is getting so precise now. This also would apply to custom controllers which could easily set up a way to unfairly improve moonkick consistency.
My thoughts for competitive emulator users is no use of keyboard and to provide a controller cam to show that the controller isn’t rigged in some unfair way. The emu/controller cam rules should only apply for top times. I’m not sure what the competitive cut would be since even my leaderboard time doesn’t even have moonkicks in it either.
I have a suspicion the grenade/weapon drops are based on the frame they are hit on a global timer. I have tried mashing through the intro and hitting at a precise time the first enemy at left which can drop the red gun icon. Some of my save states for practice seem to also have consistent drops for an enemy hit right at the beginning and when I re-do the save state the drop consistency will change. Unfortunately this means getting drops will still come down to luck and of course not leaving stars on the screen when killing enemies and projectiles. The 3-4 Boss might work in this way, in that the frame a turret is destroyed on, will count as a credit of damage or not. Also another anomaly I've come across is when I kill an enemy and another enemy dies as well even though it was somewhere else on screen. This happens in 1-2 and 4-2 with the enemy that comes in from the bottom left dying sometimes when I kill an enemy towards the center. They still give me credit for the kill which is good because a lot of times I'll get screwed because of despawns, dud grenade throws, and phantom hit boxes.
Thanks. I should have looked further into the timing stuff but I felt warmed up and just went into attempts. The only good thing now is i have a free 4 seconds to save for both level 3-2 and 5-3.
The bosses are interesting. 1-4 Boss - At first I thought it had some RNG for the damage credit but I can consistently quick kill it now with tight timing and precision. 2-4 Boss - If I have enough grenades (at least 35) it can be killed without dying consistently, but requires some dodging, and a good angle to unload the grenades quickly. I use a musical cue to get the timing on when to start spamming grenades. Before I used to throw grenades a little later and constantly got the dreaded die with one credit left of damage. 3-4 Boss - I still treat this Boss as pure luck and just hope for the best. If I'm low on grenades I'll try to throw less. 6 Cycle is the best I ever had which is like 5 or 6 seconds faster than what I got in my PB. On avg. I get 8, 9 or 10 cycle which is a reset except for the 8 cycle which is barely passable. 4-4 Boss - I admittedly really need to work on this one still but I think it can be consistently killed without dying like the 1-4 Boss. 5-4 Boss - Yes this boss has a little RNG. I got what I usually get, sometimes I can get 1 turret less (which i mistakenly thought i had for a second) or 1 more. Only a few second variance so not terrible like 3-4. I stand at the far right and without delay throw 3 grenades at the two turrets to the right to prevent them from shooting at me. If done quick enough you don't have to worry about getting hit or dodging bullets. I also have to be careful not to drift to the left at all.
The biggest setback I'm facing now besides 3-4 Boss is the grenade count. I usually don't always have enough grenades to do all the strats I practice for. This run was an exception because I always had enough.
I assume you're talking about the 2nd part of the shooter section of the game, last before the death star trench auto-scroller. While the 1st part is arguably much harder to do optimally for the speedrun, the 2nd part gave me a lot of trouble at first just to even survive it without dying. One problem that I was doing was trying to follow the TIE fighters with my cursor and they seem to always get out of reach. A trick you can do is keep your cursor towards the center briefly which can bait the TIE fighters towards the center then do quick jerks towards them to pick them off. There are 2 phases with the horizontal 1st phase somewhat easy. The 2nd phase where they come from all directions will take some more practice.
I mentioned in my comments that where you end up at the end of the level doesn't effect the times. Well that is wrong and I figured out why by reviewing some videos.
I'll break down two different 2-2 levels with one in which I finish mid-left and another that finished far-left edge.
Time from final enemy hit to character animation reaching center: Mid-left = 62 frames Far-left = 114 frames
Time from character reaching center to when music stops (when I split): Mid-left = 216 frames Far-left = 164 frames This means both are now at 278 frames total (hence why I thought they are the same) but then afterwards the mid-left will transition into level 2-3 faster.
Time from music ending to the end of the level (black screen): Mid-left = 106 Far -left = 158 That's 52 frames quicker. If I was at the center then is would be even faster.
TLDR: Get to the center at the end of the levels if possible in order to not lose time. It is up to a 2 second time loss finishing from far edge vs the center. This only applies in levels if all scenery is removed. Light scenery takes 1.5 seconds to fall giving you time to get back to the center.
I think the categories you have are fine for now. I was checking some other games that also don't have an ending. The Pinball speedrun page for example has a first to 100k points and another for rescue pauline which is comparable to complete room 10 and kill a dragon.
Once I have time maybe later next week I'll give these categories a go.
I finally got around to making a tutorial for this game. I covered a slightly different route "Bomb Room Route" than the one i use in my current Any% record. Its what i used when starting the game out because it provides for leeway in inventory management and doesn't require some more advanced tricks.
It is also useful for beating the game as well since the game kinda forces you to speedrun it in order to have a chance to get through some of the rooms.
Cool to see a page for this.
The problem I had with this game is what to set as a category. Beating the game is impossible and setting a limit at 100 or 120? (whatever number it starts repeating again) is really long and i would never bother trying that.
Setting a shorter limit seems reasonable but it also feels arbitrary what room number to end at.
The only category i could come up with is -first dragon kill- (the mini bosses scattered through the game) At least there some sort of specific goal and accomplishment that isn't arbitrary. The only problem is the randomness of when you'll come across the dragon which is anywhere from around levels 9-14 usually.
