Anyone interested in a run of this game? I don't see any listing for it yet. Or doesn't anybody know about this game? Or would it be a boring go-straight-to-boss-and-slaughter-him kind of thingy? Because you could do a get-all-character run or something. Any thoughts?
Life is like a box of chocolates-when you want peanut butter, all you get is coconut.
This thread hasn't been posted to in almost two years, so I could probably make a new thread, but I have the feeling that most people would just post a link back to this one, so I've decided to post here.
I've always enjoyed this game, and it saddens me that there is yet to be a completed run. I figured it could be fun to give it a shot (that, and I really, really need a break from pokémon).
Before I get to my work so far, I'd like to make a few general comments. After reading through the thread, the general consensus seems to be that a Princess unit would be the best way to go. I believe contrary. Having a Princess as the group leader for the main combat unit has the advantage that every unit member gets an extra attack every battle. Usually, with proper manipulation, this is unnecessary. A simple 2 x Wizard/Valkyrie, Hero group (5 targeted attacks) can take out most bosses without problem, as long as no shots miss. In the few situations where this isn't the case, as single tarot card can be used to tip the balance. If one were to have a Princess as the unit leader, manipulation would be a lot easier, because you can afford to miss one or two shots, but would universally be slower. Plus, manipulating a Crown is no easy task (I'll get to that later). That, and every map you'd need to deploy this second unit, as the opinion leader must be the leader of its group (not a big deal, but still slightly slower).
It's also been said many times that Liches should be used. Although they are very powerful, in a speed-oriented TAS this is simply infeasible. The reason, of course, is the necessary (low) alignment. The general idea is to run directly to the map boss without stopping to fight any units. Many times, this is not possible, because by the time the necessary detour around the enemy units has been made, the home base has already been seized. The solution is to plot a course through the advancing army, fighting a single unit, crippling them, but not defeating them (if you defeat a unit, and new unit appears at the enemy base. It is sometimes possible, however, to defeat them directly in front of the enemy base, so that you can encounter the boss before a new unit spawns). Other exceptions occur early on, when a few units need to be fought to gain experience. The result is that the main combat unit will always be lower in level than their opponents. This raises alignment. The only way to possibly lower alignment enough would be to either fight nothing but holy units (not an option), or waste time fighting low level units early on, which does not fit the goal of 'aims for fastest time'.
Phil wrote:
The randomness is pretty complicated in this game. Not as easy as DW. All variables can change the randomness. Number of frames, position X,Y of your deployed units, what kind of units are deployed, what Tarot cards you currently have etc... Even position X,Y of your cursor :P
I'm not sure which game Phil was playing, because this is not accurate. Or even close to accurate. As far as I can figure, the PRNG for this game works in the following manner. Before every map (or rather, upon returning to the main map screen), the PRNG is seeded. Every time a random value is accessed, the value is advanced to the next value in a calculated, determinate sequence. If one performs the same actions in the same order, regardless of how much time passed, which coordinate you sent your units to, if you deployed extra units, if you used items (other than Jokers!), etc., the results will always be the same. You will always pull the same Tarot Card when liberating, battles will be exactly the same (same damage, same combat order, same hit/miss), buried items will always be the same, enemy units will always drop the same items. This makes manipulation kind of a pain, because you need to manipulate the entire map at once, and some outcomes just aren't possible. I've only found two spots where one can successfully manipulate the initial seed. One is when you select your gender (useful for manipulating initial tarot cards and first map), and the other is at the save screen. Note that it doesn't matter whether you save or not. Well, not really anyway. Not saving is about 5 frames faster. However, after these initial 5 frames, saving or not saving both produce the same seed.
My work so far. Included are two runs, an unassisted proof-of-concept run, and a partially completed TAS test run. BE WARNED: the proof-of-concept run is exactly that, a proof of concept. I dabble around a lot, but the goal was to determine if the game could be beaten with a single unit in a sensible manner. My main unit progression looks like this:
"Samurai Master" Opinion Leader unit (3 Slash / 1 Iainuki)
2xWizard/Hero/Gryphon
2xWizard/Hero/Wyrm (stolen from Gilbert)
Wizard/Valkyrie (Norn)/Hero/Wyrm
General (Debonair)/Muse (Norn)/Hero/Wyvern
I take the time to recruit Gilbert to my army. Seems like a long process just to obtain a Wyrm, but I think I've got the method down pretty smoothly. These are the steps that need to be taken:
1. Liberate Latingur (to be told of Canopus).
2. Liberate Bah'Wahl (to talk to Canopus).
3. Revisit Latingur (to be told of Canopus' sister, Yulia).
4. Liberate the hidden Roshiafallen Temple (to recieve a 'Wing' from Yulia).
5. Revisit Bah'Wahl (to give 'Wing' to Canopus to recruit him).
6. Defeat Gilbert (Canopus does NOT need to be in the unit).
I start by sending the main unit to Latingur, and a second High Sky unit to Bah'Wahl. Shortly after I send a third Water unit to Bah'Wahl. After liberating both, I send the second unit to the hidden temple, and revisit Latingur with the main unit, fighting an experience encounter while I'm there. I then buy 13 Heals, 1 Revive, and 15 Jokers. I'm basing these numbers on my proof-of-concept run, although I think I have too many Heals and not enough Jokers. I sit on the town regaining a little health for a while, before finishing off the weakened unit and defeating Gilbert, just after the other two units reach their destinations.
At the time, I had thought that this was the best, fastest way to obtain a Wyrm early on. However, it is not. The best way would be to manipulate a 'Chime' drop on the second or third map, and the returning to the first map to catch one (at level 8, three levels higher than those from Gilbert, no less). Random drops only occur after completely defeating a unit, but I do so anyway just before Gilbert. As the Gilbert battle is more or less trivial, this would be the only thing that would need to be manipulated for this map, so it shouldn't be so difficult.
The rest of the run defeats each map in what is, in my opinion, minimal fashion (although in the proof-of-concept run, I fight more unnecessary battles). The unfinished TAS ends just after Kalbian Sea (although the last map is not frame optimized). This early on, I'm avoiding using tarot cards like the plague. The reason is that the very last battle *requires* a tarot card between every attack. Even then, I had to fight him in two segments, and needed a Revive in between, which is why I bought one. I intend to finish the test run, if for no other reason to better plan the number of Heals/Jokers needed. The course I've chosen looks like this:
01. Castle of Warren
02. Sharom
03. Sharom District
04. Jannenia Lake
05. Slums of Zenobia
06. Island Avalon
07. Diaspola
08. Kalbian Peninsula
09. The Tundra
10. Shangrila
11. Fort Allamoot
12. Dalmuhd Desert
13. Fort Shulamana
14. City of Xanadu
15. Zeteginea
16. Temple Shalina
Although, for 12 and 13, Ryhan Sea and Shrine of Kulyn could also be used. I'm not sure which would be faster, but it could be tested.
One thing you may notice while watching these is that I switch tactics a lot in the middle of battles, although it costs ~30 frames every time I do so. Most of the time when I do it, it looks completely unnecessary. Why then, you ask? Not only does the selected tactic affect which opponent is targeted, something I discovered is that it also has a side-effect on attack and defense levels. Here's a brief summary:
Leader
-Effect: Targets the leader (if possible. If not, Best is used).
-Side-Effect: None.
Best
-Effect: Targets the strongest unit of one of the available enemy classes (apparently randomly chosen. Can be manipulated, but does not affect the main PRNG. Most likely based upon frame count).
-Side-effect: None.
Strong
-Effect: Targets the strongest overall unit.
-Side-effect: Lowered attack, increased defense.
Weak
-Effect: Targets the weakest overall unit.
-Side-effect: Increased attack, lowered defense.
The optimal battle strategy usually consists of always having Strong selected when the enemy attacks, and always having Weak selected when you attack (unless you need to weaken a specific target). A lot of times, this allows one to survive in situations without using a healing card, or to defeat the enemy more quickly.
I think that's about it. Comments and suggestions are welcome :)
Joined: 3/29/2005
Posts: 229
Location: The boonies.
primorial#soup wrote:
I'm not sure which game Phil was playing, because this is not accurate. Or even close to accurate. As far as I can figure, the PRNG for this game works in the following manner. Before every map (or rather, upon returning to the main map screen), the PRNG is seeded. Every time a random value is accessed, the value is advanced to the next value in a calculated, determinate sequence. If one performs the same actions in the same order, regardless of how much time passed, which coordinate you sent your units to, if you deployed extra units, if you used items (other than Jokers!), etc., the results will always be the same. You will always pull the same Tarot Card when liberating, battles will be exactly the same (same damage, same combat order, same hit/miss), buried items will always be the same, enemy units will always drop the same items.
That seems to contradict the fact that when I'm playing at normal speed, I can manipulate the Tarot card I draw when liberating by simply having the unit wait outside the city for a second or two before liberating.
And... ok, so after a little testing, I'm sure Tarot draws when liberating towns can be manipulated by simply taking more or less time to liberate them. (I tried stopping before entering and moving to a different section of the town to liberate it; both changed the draw.) Though that may have been because I tested during the early parts of a mission, while the computer is still deploying its units and giving them orders, so there may have been calls to the PRNG during that time that might not have otherwise been made. Still, manipulation is possible outside of changing the seed - it may simply be difficult or inefficient.
And... ok, so after a little testing, I'm sure Tarot draws when liberating towns can be manipulated by simply taking more or less time to liberate them. (I tried stopping before entering and moving to a different section of the town to liberate it; both changed the draw.) Though that may have been because I tested during the early parts of a mission, while the computer is still deploying its units and giving them orders, so there may have been calls to the PRNG during that time that might not have otherwise been made. Still, manipulation is possible outside of changing the seed - it may simply be difficult or inefficient.
Hmm. You are right. There seems to be a strong corrilation with when the unit is initially clicked upon, and how far they marched to reach their destination. It seems to change every 60 frames or so (within a 60 frame timespan, the results are always the same). On the border of this 'break', one can manipulate a different card by having the unit march a little farther into the city, or a little less. Whether multiple moves commands are issued, or if items are used, seems to have no effect, as long as one remains within this timespan. Note though, that this has nothing to do with the frame count, but rather the 'in-game time' (which is why items and such can be used, multiple units released, multiple move commands issued, etc., since these all pause the in-game timer). I tried waiting a very long time (presumably until after all enemy units had be released), but this effect was still to be observed. It seems that the game makes regular calls to the PRNG, possibly to control enemy unit behavior. I appreciate the observation, as this could certainly help with manipulation.
EDIT: Upon giving this a little thought, I think this could be very useful. When sending a unit out on their detour, send them a little farther than necessary. Then let them march 60 frames or so further, or 120, etc. before sending them towards the boss. This could increase the number of possible boss battles by a factor of three or four. 'Wasting' 120 frames in this manner seems like a shame, but the potential gains from a well manipulated battle (attack order, for instance) could be much more than this.
Joined: 3/29/2005
Posts: 229
Location: The boonies.
Yeah, I just wish I could offer criticism/advice regarding the planning/route of the game. =P I loved the game, but I was never able to play it past the first floating temple thingie, for whatever reason.
I agree that the Lich wouldn't be worth shooting for, since there aren't any feasible ways to manipulate the ALI of your units in a speedrun. Though manipulating a Crown just to have an early hit-all Holy attack for the Leader's team might be valuable, say for Pogrom Forest or something.
-shrug- Just a thought.
Also, don't you need the Herostar to get Norn and (therefore) Debonair to join you? Or am I just imagining that?
I agree that the Lich wouldn't be worth shooting for, since there aren't any feasible ways to manipulate the ALI of your units in a speedrun. Though manipulating a Crown just to have an early hit-all Holy attack for the Leader's team might be valuable, say for Pogrom Forest or something.
-shrug- Just a thought.
Also, don't you need the Herostar to get Norn and (therefore) Debonair to join you? Or am I just imagining that?
I chose to skip Pogrom Forest for the (much) shorter Lake Jannenia. And besides, the only princess one could have at this point would only be level 3.
Also, the Herostar is not required to obtain Norn (if you have the patience, watch the unassisted run). The only requirement that I'm aware of is a reputation meter at least half full.
I'm not sure which game Phil was playing, because this is not accurate. Or even close to accurate. As far as I can figure, the PRNG for this game works in the following manner. Before every map (or rather, upon returning to the main map screen), the PRNG is seeded. Every time a random value is accessed, the value is advanced to the next value in a calculated, determinate sequence. If one performs the same actions in the same order, regardless of how much time passed, which coordinate you sent your units to, if you deployed extra units, if you used items (other than Jokers!), etc., the results will always be the same. You will always pull the same Tarot Card when liberating, battles will be exactly the same (same damage, same combat order, same hit/miss), buried items will always be the same, enemy units will always drop the same items.
It's rather strange that remaining of your novel and next posts are in contradiction...
Lol, I didn't know that you could both lose and win at fort Allamoot. I always killed them.
It's rather strange that remaining of your novel and next posts are in contradiction...
For what it's worth, the very first timespan is much larger than the rest (something around 300 frames or so), presumably because the enemy units aren't released immediately. When I did my preliminary testing, I must not have waited long enough. It's still very different than you describe though.
Phil wrote:
Lol, I didn't know that you could both lose and win at fort Allamoot. I always killed them.
I was rather surprised as well. I remember thinking to myself, "Crap, I lost. Wait, why are there fireworks?" It may look a little strange to lose and still complete the map, but if the game let's you do it, why not? ;)
It's rather strange that remaining of your novel and next posts are in contradiction...
For what it's worth, the very first timespan is much larger than the rest (something around 300 frames or so), presumably because the enemy units aren't released immediately. When I did my preliminary testing, I must not have waited long enough. It's still very different than you describe though.
I never claimed it to be 100% accurate. Also, I am not talking about ONLY randomness in battles. If that's what you meant?(It looks like that)
I don't mean to argue. My main point was that if you perform the same actions in the same order, even with a relatively large time variance (like I say, ~300 frames at the beginning of the map, which isn't what I originally indicated, I admit that I was wrong) the results will be the same. Liberating towns in different order or fighting battles in different order will obviously change things at lot.
At any rate, from your previous post I infer that you've seen the two runs. I'm very interested to hear any feedback you might have. Considering that you had been working on this game yourself, I think this could give you a perspective that other people, myself included, might not have.
Liberating towns in different order or fighting battles in different order will obviously change things at lot.
Well, kinda similar to what I said. :P
My opinion is that getting a lich isn't impossible( It doesn't waste so much time) since I can get it in my run(Well I don't have it since I must finish a stage before).
Also, whether which choice you made, the final bosses battle will determine what is the best choice which isn't seen in your test run. I am afraid that you may lose lot of time because of weak characters. Well, it's hard to know. I thought of many strategies without a lich and theoritically are similars. One was like you except getting Canopus and Gilbert. I still don't understand that choice since you could use a griffin that is in the reserve army.
Also, with a lich, I could answer No for Norn(Or take another route...) ;)
I am waiting to see your run that beats the final boss so, it will be easier to make an idea. But at first glance, I think I overestimate those bosses. Also, in your test run, characters did fight a lot and gained many levels where in a timeattacks it won't happen. Anyway, I don't think I will complete my run. Because those first 2-3 levels were done without frame advance since it didn't exist and more importantly, I did collect to many treasures. I did get many valuable items to be sold to get money to buy power-ups from Merchant which, now, doesn't looks useful. Anyway, I am working on some other games and I prefer letting you continue on this.
You'll have to forgive me if this has already been mentioned, but I wasn't able to read all of the above. I realize that you said your TAS was still unoptimized but I also noticed that when you bought items, you increased the amount in increments of 1. If you want to buy 15, you can push up 5 times in the 1's place, then move to the 10's place and press up once. Just a little something I noticed.
Again, sorry if this has already been mentioned. I'm glad someone finally started working on it again. Being a huge fan, I'll enjoy it thoroughly.
New WIP with progress through Shangrila. Boss battles continue to be no problem even without tarot cards. I had a thought while doing The Tundra, and that was that later on, I could probably use neutral characters for necessary leveling, as they'll be high enough level to cause my characters to level up, but won't do so much damage as a normal unit would.
I hate to be yet another person who just bumps the topic, but I believe one of the reasons runs like this die is because people simply stop posting in the topic. Perhaps bumping it (again) will help get it back on the road of completion.
So what's the latest update? This is one of my favorite games ever, so I'd love to see it completed.
EDIT: I've decided that instead of just a bump, I should also give advice from what I saw in the run as well.
I realize that the WIP is probably just a test run, but there are a couple things I noticed you could've done. Firstly, in the options menu, you can speed up text as well as battle speed. Battle speed refers to movement on the stage, not actual battles. You can also turn off animations from the same menu, which may same time since you should be in this menu anyway to turn up the text and battle speed anyway.
And like I mentioned in another post, you can buy items in increments of tens.
Also, is it possible to manipulate hidden items or item drops from battles? Judging from your unassisted WIP, it would appear so, but I wanted to make sure. Actually, forget hidden items because it would probably be much faster to manipulate drops from battles than going to certain points on the map to pick up an item. Either way, an Undead Staff and Undead Ring should be acquired so that when you get Saradin in your group, you can turn him into a Lich. Though when I think about it now, with proper manipulation, you may not even need a Lich. He may be just fine as a Sorcerer.
Will the run be going for the best ending or just any ending? Just curious.
Oh, and do you think it would be faster to replace Debonair with someone else like Rauny later on? That Wyrm/Wyvern is obviously a permanent member of the group, but he should be fine on being the tank (that is, taking all the physical hits.) Besides, Rauny already comes as a Muse and a powerful one at that who can hit everyone.
I think that's about it. If I think of something else, I'll post it.
Well, I'm a lurker from a long time ago and registered just to post here. I would really like to see a run completed, too!
Perhaps we could have more smaller runs of the worst endings, according to this page:
http://www.houseofwacks.org/ogre/faq.html
FIRESEAL would be a pretty neat run, with just one stage and free units.
Maybe the speed is greater from the bottom endings?
