Joined: 3/28/2009
Posts: 28
I recently decided to pick up TASing on a whim, and Boktai 2 was one of the first games I considered as it's a personal favorite. The main problem with TASing Boktai is that it's very puzzle and stealth heavy, so there isn't any reason to fight monsters except for during Traps and farming items. Compounded on that, I've not yet found any sequence breaking opportunities or means of making rooms/dungeons faster other than perfecting paths. It makes me think most people probably wouldn't find it a very entertaining run, unless perhaps you like the game as I do. Maybe not though, it's an awful lot like MGS in some respects, and the TAS of that that's being worked on is reasonably entertaining. There's also a whole lot of talking sequences that seem to be unavoidable (though some of them I've found are), which can also be a deterrent to watching. There isn't a lot that I've found in the way of speed saving so far; staircases and ramps are coded in such a way that you can walk onto them several pixels above their lowest point, you can speed up fights by standing parallel to a monster and facing a certain way to hit behind them, which in some cases means you don't have to move to far in one direction to optimize damage done, and . One thing I've had in mind since I started that could eventually lead to a problem is that the solar patch version I'm using has the solar bar set to L, which could possibly get in the way later in the game when I acquire dash. The alternative patch uses select, I believe, and that would mean only being able to turn the solar gauge up when changing weapons/magic. I've hit a roadblock with the WIP on Durathror and need to go back in the run and find a monster that I can manipulate into dropping a stronger weapon and possibly level up once or twice, so I'm in the process of cutting the run up room by room and optimizing each one as I re-record from the old file. I should have a newer WIP sometime soon, I've so far saved about 2000 frames on my old one. It's too bad there wasn't already a topic for this game, or even the first one. The most I could find about either was 5 minutes of a TAS on youtube and an unoptimized example run of Boktai1 somewhere on these forums. :/
Editor, Emulator Coder, Expert player (2158)
Joined: 5/22/2007
Posts: 1134
Location: Glitchvania
I'm not familliar with this game. The only thing comes into my mind is that you may plan for items in advance so that you can get most of them from enemies along the way rather than farm from the same type of the enemies that you just sneaked past, which you must have already been aware of since it's evident.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days <adelikat> no doubt <adelikat> klmz, they still do
Joined: 3/28/2009
Posts: 28
klmz wrote:
I'm not familliar with this game. The only thing comes into my mind is that you may plan for items in advance so that you can get most of them from enemies along the way rather than farm from the same type of the enemies that you just sneaked past, which you must have already been aware of since it's evident.
Farming items can be minimized because the only ones I can see ever really needing would be weapons, specifically rare drops. If I can manipulate a rare drop hammer or spear it would probably save the time forging (which I might end up having to do at some point anyway). I planned on farming an enemy along the way, the problem is finding one that isn't in a room with a bunch of others. I might experiment with a couple traps and see if I can get drops off of the monsters in them without losing too much time. So far I haven't been able to tell if drops are determined by X-Y co-ordinates, frame number or what. I think i've used fruit like 3 times so far, maybe, and I'm pretty sure they were the ones you start with, so i think rare weapon drops will be all I'll need to worry about. I'm hoping I can find the calculations in the memory that decides how much str attributes and damage on your weapon needs to be done to do more than 1 damage as that would help my planning. In the meantime I'll probably just go to each boss and up my damage one point at a time to see. :/ Also, having the solar gauge up makes fruit (and probably other replenishing items) rot, and if you eat rotten items you get a 5 health gain, but it also makes the screen pixelate. I did it once in the run, but I might just optimize the damage I took earlier and my solar gauge use because it was actually kind of annoying to watch the game and not be able to see anything.
Editor, Emulator Coder, Expert player (2158)
Joined: 5/22/2007
Posts: 1134
Location: Glitchvania
Hmm, my previous post looks a little messed up... however you seem to understand what I tried to express. Well, I'm just curious about which patch you are using at the moment. I've only seen a Boktai 1 patch using L for the solar gauge.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days <adelikat> no doubt <adelikat> klmz, they still do
Joined: 3/28/2009
Posts: 28
klmz wrote:
Hmm, my previous post looks a little messed up... however you seem to understand what I tried to express. Well, I'm just curious about which patch you are using at the moment. I've only seen a Boktai 1 patch using L for the solar gauge.
No Frills has an L button patch for the E version, and I'm pretty sure there's a select one somewhere (though I could be mistaken). I'm using a pre-patched version of the U rom, though.
Joined: 3/28/2009
Posts: 28
It would seem the most damage I'll be able to do after manipulating a Mace off a Golem and using the 9 stat points I have at lvl 11 is 13 to the first 2 forms of Durathor, 16 if I use darkness. However, I remembered that you can use wolf on her final form which inflicts 23 damage per bite. That should speed up the end of the battle significantly, but I see this damage/str cap being a problem down the road as it isn't possible to get better weapons without forging or spawning a crimson monster, which I have no idea how to go about doing lol. Either solution might prove to provide to little entertainment to make the run viable... such a shame. I hope I find something to prove otherwise. If someone wanted to watch the less-optimized WIP, let me know what you think about the entertainment factor. I've figured it's probably low, I just wanna know for sure. Here's a link, be sure to use US version. http://www.mediafire.com/?nz4jdbzyt05 http://dehacked.2y.net/microstorage.php/info/82452022/Boktai%202%20-%20Solar%20Boy%20Django%20TAS.vbm
Editor, Emulator Coder, Expert player (2158)
Joined: 5/22/2007
Posts: 1134
Location: Glitchvania
The first ~9 minutes full of conversations look very boring, while it looks better and better as more actions are introduced. I personally like puzzle games, so the heavily puzzle-solving looks comfortable to me. Lags seem to be a big problem. Is there any way that you can reduce lags? It seems that when the icon on the lower-left blinks, more lags occure. If you will make a run for publication, be sure that you start recording it from Power-on instead of a snapshot.
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days <adelikat> no doubt <adelikat> klmz, they still do
Joined: 3/28/2009
Posts: 28
Yeah, the talking seems to be a recurring problem. I had forgotten how much dialogue there was until I started the TAS. Maybe eventually I'll find a way to skip some of it. As to lag, I'm not really sure what's up with it. I think it's an emulation problem and I've not yet been able to fix it. Sometimes it's there and other times it isnt. I originally thought it was my computer, but I guess that's not the case. I don't know why it starts from the snapshot. I had recorded from power on originally. How far into the movie did you get? I'm having desynch problems with it.
Editor, Emulator Coder, Expert player (2158)
Joined: 5/22/2007
Posts: 1134
Location: Glitchvania
I didn't have any desync problems, weird. One more thing comes into mind: Isn't this game using a battery-powered Real-Time Clock? Your wip didn't go with RTC emulation enabled, though. EDIT: There's one infamous possible cause why you have desynch problems: VBA v19.x has a bug that sometimes it mysteriously erases the Lag Reduction flag from your new movie files. Be sure to replay any GBA movie that you just started to record and see if Options->Emulator->Lag Reduction (GBA) is checked. If not, you may check it and start over, or hex-edit the byte at 17h in your lagged movie file to 30h (RTC disabled) or 34h (RTC enabled). This bug has been fixed in v21-interim. See also: VBM format
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days <adelikat> no doubt <adelikat> klmz, they still do
Joined: 3/28/2009
Posts: 28
klmz wrote:
I didn't have any desync problems, weird. One more thing comes into mind: Isn't this game using a battery-powered Real-Time Clock? Your wip didn't go with RTC emulation enabled, though. EDIT: There's one infamous possible cause why you have desynch problems: VBA v19.x has a bug that sometimes it mysteriously erases the Lag Reduction flag from your new movie files. Be sure to replay any GBA movie that you just started to record and see if Options->Emulator->Lag Reduction (GBA) is checked. If not, you may check it and start over, or hex-edit the byte at 17h in your lagged movie file to 30h (RTC disabled) or 34h (RTC enabled). This bug has been fixed in v21-interim. See also: VBM format
I think it was one of the setting I had messed with, I just ran one of my backups instead and it was fine. I saw so many things I did wrong, lol. I should've rewatched my progress more often. What does the RTC do, exactly? I messed around with it on and off once but I didn't see any difference. In any case, the only thing I can see using the clock for is changing monster spawns and money duplication, and changing the time for former requires a lot of planning.
Editor, Emulator Coder, Expert player (2158)
Joined: 5/22/2007
Posts: 1134
Location: Glitchvania
RTC emulation will automatically synchronize the in-game clock with your PC's computer, and you don't need to manually set up the time. If you start the movie (from power-on) at night, you will see it's also at night in the game world. If I recall correctly, some monsters only spawn at certain time periods. So using RTC may give you more controls as well as the advantage to skip setting the in-game clock. However I haven't recorded any movies enabling RTC myself. So I am not certain if movies re-recorded with RTC enabled play correctly. This demands tests. Oh, there's a published movie with RTC: Pokémon Sapphire
<klmz> it reminds me of that people used to keep quoting adelikat's IRC statements in the old good days <adelikat> no doubt <adelikat> klmz, they still do
Joined: 3/28/2009
Posts: 28
I'll mess around with time and see which monster spawns are affected, it could help with getting better rare drops depending on which spawns change (I only know a couple off the top of my head). I need to find a rom without the Rising Sun intro thing as well. I think I saw a prepatched/cracked U version without the intro somewhere... If not I'll just patch an E version.
Joined: 3/28/2009
Posts: 28
Restarted my WIP and saved a handful of frames (some weird combination of a+b+left+right on rapid saved me like 14 frames just in the intro alone. no idea why). I think I'm about 3000 frames faster currently, might be less. There was some bad routing in my first WIP i've improved, and I think I can cut a couple rooms out completely. For planning stats, I think i only have to pump AGL, depending on when it caps out. 2 STRG = 1 Damage, I don't see putting point into it being very useful. I put points into STRG 2 times during the old WIP, but I may go back and change that. Every 4 AGIL = an 8 frame speed increase, so I'll likely only need to spend stat points per 2nd and 3rd levelups. I am not sure about upgrading SPRT, it would be dependent on points left from upgrading AGIL or STRG. I shouldn't need mana enough to bother spending points there since I'll have max sunlight for most of the run and plenty of fruit (i think). The thing about weapons is that getting a more powerful weapon increases damage more than stat points do. I think you can only buy stronger weapons as you forge, but I could be wrong. So it depends on how long rare drop weapons will last (and if I can get them) whether or not I ever have to forge. Which brings me to to the problem that's halted all progress with the run so far... I can't figure out how to manipulate drops. My initial testing made it looks like it was coordinate-based in relation to where you are from the monster, but there also seemed to be a certain window from when your last hit lands (something like 15 frames) where you either have to stop moving completely, or find the specific area where the drop will spawn. I wasn't able to manipulate rare drops at all. I'm going to try and pinpoint exactly how the RNG works but I'm still a new at watching/finding memory addresses. If I can't find anything useful I'll give up on that and see how I do just buying items. Hopefully I'll get lucky with finding the RNG, though. EDIT:: Speed caps at 42 agility per 8 stats. Unless Strength does something significant later on I might not use stats past there
Pokota
He/Him
Joined: 2/5/2014
Posts: 779
Mucking around with a baseline route for Boktai 1, and I ran into the issue that Delusion Forest can't reliably be navigated at night. Also, can't run max sunlight all the time, the game still gives overheats after spending so much time at max; maybe spend dungeons at 1 and only use max for purifications and finals. Here's what's available so far. Uses Prof. 9's Solar Sensor Patch; this resulted in the following hashes: SHA1:0D3A978EA8FB02A356B98DE51E7AB07BD1AAECC0 MD5:50411EC2A648DA081DABEE3162FC7DEF Notes: Went through Catacombs to pick up the Beatmania frame to make Monsters-on-weight-switch puzzles more direct (thus far there's only 2, both in Bloodrust Mansion, but BM trivializes the first one and the second one you need stunning to do anyway) Picked up the Lance frame in Crumbling Mine; there's probably a faster route to Firetop Mountain (and probably better weapons to use, with how the Solar Piles work during Purification) but I like this type of "Gun". Because Garmr plugs up the hole that the sunlight comes down through, it's considerably faster to do Muspell first; even without Permafrost forcing you to take both paths on a round trip, it takes longer to do Garmr without the Flame Lens than it does to do Muspell without the Frost Lens (to my knowledge, there aren't any useful Frost Lens-only paths in Firetop Mountain)
Adventures in Lua When did I get a vest?
Joined: 6/27/2008
Posts: 49
Pokota wrote:
Uses Prof. 9's Solar Sensor Patch; this resulted in the following hashes: SHA1:0D3A978EA8FB02A356B98DE51E7AB07BD1AAECC0 MD5:50411EC2A648DA081DABEE3162FC7DEF
As solar sensor has been offically supported by mgba (https://github.com/mgba-emu/mgba/blob/master/CHANGES), I think it's better to use Bizhawk with mgba core instead of a 3rd party patch. Any way, as a boktai fun, I'm looking forward to your run!
Pokota
He/Him
Joined: 2/5/2014
Posts: 779
That... actually changes quite a bit of thing. New Hashes: Boktai - The Sun Is in Your Hand (USA) SHA1:7164326283DF46A3941EC7B6CECA889CBC40E660 MD5:FF75C62AB690410CC8FCA24204D783E9 I'll have to fiddle with the calibration; the game is programmed to overheat after so much SOL is collected in one sitting. Since more sunlight is better than less it'll need to be studied. In RAM. (dreading this) Also, I'm making a baseline route rather than a dedicated TAS; any TAS of the same category should either beat my time or else have a very good reason for why it doesn't. I've got a day off on Friday; I'll probably set up a full baseline then. Need to science with the RTC so that it's night up until entering Delusion Forest - having sunlight at night makes Bloodrust Mansion and Fog Castle a fair bit faster as you're not dealing with near as many enemies. I want to say it also screws with the Count's AI since he's supposed to be smart enough to not dash straight into light coming in from the skylights (why he even has those in his mansion...) E: It doesn't screw with his AI, he did it to me again just now with the game seeing daylight hours. So having max sunlight for the first Count fight is a Good Thing™ since the stronger the sunlight, the more suicidal he gets past the halfway mark.
Adventures in Lua When did I get a vest?
endrift
Any
Emulator Coder
Joined: 12/14/2014
Posts: 161
While debugging the solar sensor code, I managed to find the offset of the value itself, which may help cut out some of the early studying time. It's 0x3004508 in the US version, dunno if it's different in different regions though.
Pokota
He/Him
Joined: 2/5/2014
Posts: 779
So I tried setting the in-game clock for 5:15 AM on this coming June 20th (to try and see if the Full Moon notification would be worth the Full Moon perk of getting free healing bugs in the moonlight) but Biz seems fixated on making it 8 PM on this coming June 19th (I set the date in the RTC settings to be 6/20/2016, so it's at the very least missing leap day). Can someone confirm whether this happens to them as well? Here's what was going to be the route test.
Adventures in Lua When did I get a vest?
Joined: 4/10/2017
Posts: 8
This seems to have died so I started doing my own science tests. I found out that for me on the east coast of the US TASing at night the best way for me to use the solar sense in Bizhawk is to use a european copy and pretend I'm in perth (12 hours ahead right now) for Boktai 2 I plan on making mild route changes to the current WR run as a baseline for now and go from there It may end up being faster to use the Prof.9 patch since that is accepted by the boktai community as valid and only takes 1/3rd of a second to get to max sunlight regardless of time because I can already see potential desync issues when encoding
Player (62)
Joined: 10/18/2011
Posts: 47
Resurrecting this thread as I'm currently working on a TAS of Boktai 2, and progress has been much faster than I anticipated thanks to the help of the very talented minds over at the Taiyoh Network Discord server. Boktai 2 is separated into several dungeons that must be done in sequential order: Cathedral, Remains, Dark City, Aqueduct, Spiral Tower, Undead Zone. No current means exist to sequence-break these or skip any story events, and even if you could, most of your loadout obtained through story events is required for Undead Zone. Previous attempts from others at this game didn't appear to get very far, I'm currently on the interim leg between Dark City and Aqueduct, so just over halfway through the final movie file. The biggest impediment to progress has been the EXP routing, and RNG manipulation; I had to redo Remains twice due to simple arithmetic errors in my EXP routing, even whilst working on a spreadsheet using formulas. As far as working with RNG goes in this game, the golden memory value is 46B8 found in IWRAM. which will henceforth be referred to as RNG2. RNG2 is what controls all of the random events that we care about, such as weapons dropped by enemies, forge results and other area-specific random calls. It is a value that defaults to 833 after booting the ROM, then progresses until 1023, after which it overflows to 0 and cycles again. Thus, RNG2 is always an integer between 0-1023, and is progressed by different events. Some of these are automatic, and some of these can be caused by the player. Of note is the drop table, which was published some time ago by the previous RTA WR holder, Coa, and shows which RNG2 values are required when a random drop check is called to get either a Common item or a Rare item. In virtually all cases, we are interested in Rare items, which can drop on 12 of the possible 1024 RNG2 values, approximately a 1.2% chance. 7 of these are obtained over the course of the run so far, with a planned collection of two copies of a specific tarot card, "The Tower", which is on a separate drop table at a rate of 40/1024, or 3.91%. The most critical element of the route by far has been forging a Katana at the earliest possible opportunity. Some posts were previously made about getting access to better weapons, and which weapons would be best to take into the boss of Dark City, Red Durathror. Something that these posts omitted is that weapons in this game are level-locked; it would do no good to magically get the best weapon in the game to drop, because you're not a high-enough level to equip it at any point in an RTA run, let alone a TAS. By the point that forging becomes relevant, you are generally Lvl 12. The best weapon you can reasonably take into the next dungeon (Dark City) is the Lvl 13 sword, Long Sword. Hammers and Spears are also usable in this game, but Swords are overwhelmingly the best weapon choice due to their attack speed. However, the TAS route hinges on crafting a Katana, which is a Lvl 17 weapon, then taking the most efficient EXP grind possible in order to be able to equip it for the boss fights in Dark City. This sounds like a sub-optimal solution, but the timeloss of taking your next-best options into Durathror is much worse, to the tune of a few minutes. Getting the Katana craft to work at all was a near-impossibility, we knew it should have been theoretically possible, but it had never actually been proven in RTA play. First of all, I needed eligible craft materials. Forging in this game follows different rules depending on which types of weapons you use as materials; same-type weapons will always result in a weapon of that type, whilst different-type weapons will result in the remaining type. In order words, Sword+Sword will always create a Sword, whilst Hammer+Spear will also always create a Sword. However, where these situations differ is the level modifier applied to the resulting weapon. When using same-type weapons, the game will tell you explicitly which weapon you will get as a result of the forge, because a high level modifier is applied to the resulting weapon; the result of a forge is (Weapon A Level + Weapon B Level) / 2 + r, where r = 6 when materials are of the same type. You will always get the highest-level weapon that does not exceed the final level value. As a practical example: Short Sword (Lvl5) + Broadsword (Lvl9) = Long Sword (Lvl13) (5 + 9) / 2 + 6 = 13 This is simple enough for same-type weapons, but when different-type weapons are used, r will become a random number between -6 to +6. In the above scenario, the +6 bonus is necessary to create the Long Sword, and any lower value results in a Broadsword, as it is the next-highest weapon after the Lvl13 Long Sword. This is a problem for the TAS route, because we don't have access to high-level swords as crafting materials! The only high-level weapons that can be manipulated during Remains are Maces, the Lvl 9 Hammer. We must do two crafts to get the Katana, and they're both different-type crafts, which turns our automatic +6 weapon level bonus into a random -6 to +6 bonus. Only +6 works, so both crafts are now a 1/13 chance to result in the correct weapon. Thus, the craft route becomes: Short Sword (Lvl5) + Mace (Lvl9) = Lance (Lvl13 Spear) Lance (Lvl13) + Mace (Lvl9) = Katana (Lvl17 Sword) But we're not done yet, the Lvl17 Sword is the Zweihander, not the Katana. At least, that's the weapon you'll get 99 times out of 100. The Katana is an "R-rank" weapon, essentially a Rare craft. An additional check is performed when initiating the craft to see if the weapon will be eligible to progress to "R-rank", which for all intents and purposes means hitting a short frame window; in the TAS I have a 1% chance of getting an R-rank craft, which is defined as starting the craft in a 4-frame window, on a framerule of 400. If the craft is successfully started at the correct time, RNG2 must also be at the correct value when the forge finishes in order for the resulting weapon to be the Katana instead of the Zweihander... And I have next to no control over how RNG2 advances during the forges. Additionally, I'm trying to apply an extra variable during the Katana forge by giving the weapon its "SP effect", which is granting extra attack power based on your Agility stat. Thankfully, this variable is skipped in practical terms by simply doing enough strikes in the forge to grant a 100% chance for SP effect acquisition, which also allowed me to finish on the correct RNG2 value to get the Katana SP. After gaining this weapon, it's basically "go-mode", just get to the end of the game as fast as possible. No weapons later in the game are worth taking the time to collect and use, as the Katana remains a competitive damage option up to and including the final boss, and due to its SP effect it punches even further above its weight (Essentially gives you +6 Strength) and never suffers durability loss that forged weapons normally suffer over time, which causes their damage to degrade over time and causes undesirable lag whenever this degradation occurs. My current WIP file should be available for download through my uploaded files, and should sync on the newest Bizhawk release, 2.9.2. I'm using the unmodified JP 1.0 ROM, so don't attempt to play the movie on a patched ROM or it will not sync.