Final Fantasy 1 - Segmented (Run Submitted!)

2009-11-12 11:57:50 pm

Actually, as far as inputs go, you can use whatever you have available. SDA just doesn't accept the VHS format.

Great to see you back here, WarMech. That's one notch off my queue. =D

mikwuyma

2009-11-13 08:06:13 am

My feelings on The Demon Rush

Good to see you back, Warmech.

Quote from Carcinogen:

Actually, as far as inputs go, you can use whatever you have available. SDA just doesn't accept the VHS format.

This. I'm not sure how you record an RF signal on a DVD recorder, but I'm sure it's possible.

BTW, if you end up recording with VHS, I can convert it to DVD for you.

Carcinogen

2009-11-13 09:12:13 am

I've had two DVD recorders to date that have RF Input. They do exist. <_<

Winkwonle

2009-11-13 09:44:25 pm

Thanks for the welcoming words ! :-)

Yeah, I've told people that I was coming with a single segment on the forum, but nobody replied. Anyway, it's done and I hope it comes out on the official speedrun list soon. I don't think it'll bring much valuable info for your segmented run, however, since you'll beat me by more than an hour !

On with the discussion...

Do you manipulate your KYZOKU and ZomBULL battles so that you have to move around the least possible on the world map ? Since you're using the Power trick, you might as well move around for exactly one step less than required, hit reset, and get a free one-step encounter for your next segment. Also you may tent in water, there is no point getting out of water for that.

On that note of counting steps before resetting, I've tried a little experiment with the emulator. We know the number of squares that you have to travel before the first encounter is fixed, provided you hit Power beforehand. I'm not sure, but it seems that it's also fixed for the second encounter, and so on. I've counted 24 for the first, then 44, 55, 50, 31, 48, 73, ... If you really want to do some mad optimization on your time, you could try to max that number out while travelling on the world map. Unfortunately, when I tried in on the NES, it didn't seem to work more than 25% of the time. Also the Power leveling trick doesn't work all the time on my machine, I don't know why.

Why do you farm ZomBULLs *before* grabbing the TNT and looting the Dwarf cave ? Sure you cover a little bit less of distance on the world map, but you get the House you needed, and much better equipment with the extra 9175 gold, at the cost of an extra trip to Elfland. That extra trip could even be to Melmond instead. Hell, once you're there, do you really need that farming ?

You said at some points that you couldn't run away from Frost D's and SORCERERs, I think. Well, even if you didn't say so, for reference, I copied the list of unrunnable encounters here:

Battle 1C (2-4 Wizards / 3-7 Wizards)
Battle 21 (1 Earth / 2-4 Earths)
Battle 22 (0-1 Cerebus, 1-2 WzOgres / 1-3 Cerebus, 0-2 WzOgres)
Battle 27 (1-2 Fires / 3-4 Fires)
Battle 2E (1 FrGiant, 0-2 FrWolves / 2 FrGiants, 2-6 FrWolves)
Battle 32 (1-3 ZomBulls / 1-4 ZomBulls, 0-2 Trolls)
Battle 46 (1 Phantom / 2-5 Ghosts)
Battle 49 (1-3 Waters / 3-6 Waters)
Battle 4B (1-2 Zombie Ds / 2-4 Zombie Ds)
Battle 4E (1 Blue D / 2-3 Blue Ds)
Battle 57 (1-2 Worms / 3-4 Worms)
Battle 59 (1 Gas D / 2-4 Gas Ds)
Battle 69 (1 Eye / 2-3 Eyes)
Battle 73 (1 Lich)
Battle 74 (1 Kary)
Battle 75 (1 Kraken)
Battle 76 (1 Tiamat)
Battle 77 (1 Tiamat [2])
Battle 78 (1 Kraken [2])
Battle 79 (1 Kary [2])
Battle 7A (1 Lich [2])
Battle 7B (1 Chaos)
Battle 7C (1 Vampire)
Battle 7D (1 Astos)
Battle 7E (9 Pirates / 1-2 WzSahags, 8 R.Sahags)
Battle 7F (1 Garland / 1-2 IronGols)

In the game, there are seven Houses, three of which you pick up: Marsh cave (yes), Dwarf cave (yes), Gurgu volcano, Castle of Ordeals (yes), some uninteresting Cardia island, Mirage tower, and Sky palace. You're planning to use them: on ZomBULL farming, before the Ice cave (to replenish from Ordeals), and before entering Onrac (to save time from the trip to the Inn). If you'd spend two during the Zombull farming instead, like I suggest, you'd have to pick or buy an extra one. That can be done at the point where you buy Soft potions from Elfland, or when you're getting the Canoe at Crescentlake, or even replaced by a special extra trip to elfland on your way back from Ordeals, where you can buy FAST as well (and never go back for CUR2). You'll notice that Elfland is on the NE path from Ordeals to the Ice cave anyway. (a good thing we're measuring all trips in terms of the Manhattan distance, right)

On the subject of the LIFE spell, I'm not sure I was understood right. I was suggesting that you make use of LIFE in the TOFR to be able to keep very lucky attempts in which one of your characters happen to die. Sure my suggestion is meaningless if your patience in redoing the segments is unlimited. But you're not aiming at beating the TAS, right ? Now I'm not a pro at segmented runs, but I've done my share of probability courses, and here's what it tells me...

You can view your power to redo segments as a luck bank. The more you're allowing strategies that require luck, the more you're able to reach TAS-like times, but also the more you'll have to redo the segments. Ideally, you would gauge the exact amount of luck you need on the exact patience that you have as a player. If you buy the LIFE spell, you're putting extra luck in your bank; if you renounce to unrunnable encounters, you're drawing luck from it, etc. If you think you can do the TOFR without ever losing a character and without exceeding your patience ressources, then you should consider doing it at lower level. It's just a question of balancing your luck.

I'm sorry for the FiFiWmWm suggestion. You're right, it's much too weak during farming. I don't know what I was thinking about. Also, when I was talking about a black mage, I was talking about replacing a Fighter, not the Red mage. That would still be weak during farming, though. But since you're thinking about dropping a character for the endgame...

I suppose you're thinking of keeping the guy up to right before the EYE farming, are you ? If so, will you consider using a different team ? Perhaps two Red mages, or Red/Black, or Red/White, or even White/Black ? You may still use an under-levelled mage as a spell bank for the rest of the game. You may even use the (my ?) petrification trick on a weaker Black mage to save FAST charges for CHAOS.

Quick review of character features
---

White mage:
- HARM at level 1, kills undeads for free up to the Ice cave, cheaper than FIR2 charges.
- MUTE at level 2, free win against ASTOS lets you manipulate critical hits instead of luck against his spells; less effective than ICE though.
- HEL2 at level 12, very good healing device (remember to use in battle only, where it behaves like HEL3 due to a bug), saves lots of HEAL potions.
- EXIT at level 15, substitutes for WARP in waterfall and 2xWARP in Sea Shrine (but you have to go back in the village).

Black mage:
- ICE at level 2, kills ASTOS pretty fast.
- FIR2/LIT2 at level 5, two slots at level 6.
- FAST at level 8, three slots at level 11, four slots at level 14
- WARP at level 12, two slots at level 13
- Also has access to BANE, which is ok

Red mage:
- Fights well (= ninja)
- CURE, ICE, MUTE available at same level as White/Black mage
- FIR2/LIT2 at level 6
- FAST at level 10, three slots at level 13, four slots at level 16
- WARP at level 15, two slots at level 16

That's all I can think of for now.

Thanks for the welcoming words ! :-)

Yeah, I've told people that I was coming with a single segment on the forum, but nobody replied. Anyway, it's done and I hope it comes out on the official speedrun list soon. I don't think it'll bring much valuable info for your segmented run, however, since you'll beat me by more than an hour !

On with the discussion...

Do you manipulate your KYZOKU and ZomBULL battles so that you have to move around the least possible on the world map ? Since you're using the Power trick, you might as well move around for exactly one step less than required, hit reset, and get a free one-step encounter for your next segment. Also you may tent in water, there is no point getting out of water for that.

On that note of counting steps before resetting, I've tried a little experiment with the emulator. We know the number of squares that you have to travel before the first encounter is fixed, provided you hit Power beforehand. I'm not sure, but it seems that it's also fixed for the second encounter, and so on. I've counted 24 for the first, then 44, 55, 50, 31, 48, 73, ... If you really want to do some mad optimization on your time, you could try to max that number out while travelling on the world map. Unfortunately, when I tried in on the NES, it didn't seem to work more than 25% of the time. Also the Power leveling trick doesn't work all the time on my machine, I don't know why.

Why do you farm ZomBULLs *before* grabbing the TNT and looting the Dwarf cave ? Sure you cover a little bit less of distance on the world map, but you get the House you needed, and much better equipment with the extra 9175 gold, at the cost of an extra trip to Elfland. That extra trip could even be to Melmond instead. Hell, once you're there, do you really need that farming ?

You said at some points that you couldn't run away from Frost D's and SORCERERs, I think. Well, even if you didn't say so, for reference, I copied the list of unrunnable encounters here:

Battle 1C (2-4 Wizards / 3-7 Wizards)
Battle 21 (1 Earth / 2-4 Earths)
Battle 22 (0-1 Cerebus, 1-2 WzOgres / 1-3 Cerebus, 0-2 WzOgres)
Battle 27 (1-2 Fires / 3-4 Fires)
Battle 2E (1 FrGiant, 0-2 FrWolves / 2 FrGiants, 2-6 FrWolves)
Battle 32 (1-3 ZomBulls / 1-4 ZomBulls, 0-2 Trolls)
Battle 46 (1 Phantom / 2-5 Ghosts)
Battle 49 (1-3 Waters / 3-6 Waters)
Battle 4B (1-2 Zombie Ds / 2-4 Zombie Ds)
Battle 4E (1 Blue D / 2-3 Blue Ds)
Battle 57 (1-2 Worms / 3-4 Worms)
Battle 59 (1 Gas D / 2-4 Gas Ds)
Battle 69 (1 Eye / 2-3 Eyes)
Battle 73 (1 Lich)
Battle 74 (1 Kary)
Battle 75 (1 Kraken)
Battle 76 (1 Tiamat)
Battle 77 (1 Tiamat [2])
Battle 78 (1 Kraken [2])
Battle 79 (1 Kary [2])
Battle 7A (1 Lich [2])
Battle 7B (1 Chaos)
Battle 7C (1 Vampire) 
Battle 7D (1 Astos)
Battle 7E (9 Pirates / 1-2 WzSahags, 8 R.Sahags)
Battle 7F (1 Garland / 1-2 IronGols)

In the game, there are seven Houses, three of which you pick up: Marsh cave (yes), Dwarf cave (yes), Gurgu volcano, Castle of Ordeals (yes), some uninteresting Cardia island, Mirage tower, and Sky palace. You're planning to use them: on ZomBULL farming, before the Ice cave (to replenish from Ordeals), and before entering Onrac (to save time from the trip to the Inn). If you'd spend two during the Zombull farming instead, like I suggest, you'd have to pick or buy an extra one. That can be done at the point where you buy Soft potions from Elfland, or when you're getting the Canoe at Crescentlake, or even replaced by a special extra trip to elfland on your way back from Ordeals, where you can buy FAST as well (and never go back for CUR2). You'll notice that Elfland is on the NE path from Ordeals to the Ice cave anyway. (a good thing we're measuring all trips in terms of the Manhattan distance, right)

On the subject of the LIFE spell, I'm not sure I was understood right. I was suggesting that you make use of LIFE in the TOFR to be able to keep very lucky attempts in which one of your characters happen to die. Sure my suggestion is meaningless if your patience in redoing the segments is unlimited. But you're not aiming at beating the TAS, right ? Now I'm not a pro at segmented runs, but I've done my share of probability courses, and here's what it tells me...

You can view your power to redo segments as a luck bank. The more you're allowing strategies that require luck, the more you're able to reach TAS-like times, but also the more you'll have to redo the segments. Ideally, you would gauge the exact amount of luck you need on the exact patience that you have as a player. If you buy the LIFE spell, you're putting extra luck in your bank; if you renounce to unrunnable encounters, you're drawing luck from it, etc. If you think you can do the TOFR without ever losing a character and without exceeding your patience ressources, then you should consider doing it at lower level. It's just a question of balancing your luck.

I'm sorry for the FiFiWmWm suggestion. You're right, it's much too weak during farming. I don't know what I was thinking about. Also, when I was talking about a black mage, I was talking about replacing a Fighter, not the Red mage. That would still be weak during farming, though. But since you're thinking about dropping a character for the endgame...

I suppose you're thinking of keeping the guy up to right before the EYE farming, are you ? If so, will you consider using a different team ? Perhaps two Red mages, or Red/Black, or Red/White, or even White/Black ? You may still use an under-levelled mage as a spell bank for the rest of the game. You may even use the (my ?) petrification trick on a weaker Black mage to save FAST charges for CHAOS.

Quick review of character features
---

White mage:
- HARM at level 1, kills undeads for free up to the Ice cave, cheaper than FIR2 charges.
- MUTE at level 2, free win against ASTOS lets you manipulate critical hits instead of luck against his spells; less effective than ICE though.
- HEL2 at level 12, very good healing device (remember to use in battle only, where it behaves like HEL3 due to a bug), saves lots of HEAL potions.
- EXIT at level 15, substitutes for WARP in waterfall and 2xWARP in Sea Shrine (but you have to go back in the village).

Black mage:
- ICE at level 2, kills ASTOS pretty fast.
- FIR2/LIT2 at level 5, two slots at level 6.
- FAST at level 8, three slots at level 11, four slots at level 14
- WARP at level 12, two slots at level 13
- Also has access to BANE, which is ok

Red mage:
- Fights well (= ninja)
- CURE, ICE, MUTE available at same level as White/Black mage
- FIR2/LIT2 at level 6
- FAST at level 10, three slots at level 13, four slots at level 16
- WARP at level 15, two slots at level 16

That's all I can think of for now.

Winkwonle

2009-11-13 10:10:35 pm

Oh, sorry for that, EXIT is at level 16 for the White mage.

WarMech

2009-11-15 05:28:31 pm

Thanks for the support! I'd thought you guys might have forgotten about me... Sad

To tell you the truth, I'd been lurking these forums every month or two since I took my break and actually downloaded your WIP single segment run. Obviously, it was just a test, but I thought the principles in it were very solid: Good group choice, good knowledge of power cycles, and good balance between security and speed. I wish I'd had time to comment on it, but know that I was one of the only people to download and watch the whole thing! Grin

Yes, I will manipulate the ZomBull/Kyzokus for a one-step encounter. I didn't in my test runs because they were only tests, but this will save 1-2 seconds per segment in the actual run.

That's a cool bit of info about the encounters and was what I was starting to get at in #2 and 3 in my last post. I will be adding segments to do the maximum number of steps before an encounter in the later segments when I have enough tents to support it. Earlier on, I will track exactly what part of the power cycle will generate an ideal set of encounters in that segment. That knowledge on the number of steps before the first encounter will be critical, although it might be modified for different terrians (water, river, woods) since I think that effects it.

I like the suggestion about going to the dwarf cave first. We may need to do a little timing but that really sounds like it could be a time saver in the long run. I was trying to minimize the distance walked at first, but I think getting the house there would more than make up for it.

I see where you are coming from now with the Life and it makes sense. The question is now whether the time gained from having an extra live if I need it is worth the time gained from a Fire 3 to burn some enemies in TOFR. You suggestion has much more merit if I decide to go with 2 fighters rather than 2 at the end.

Your reasons are very logical for a single segment, but here is a summary of my counter-argument for the white and black mages:

White:
- Harm doesn't buy me all that much. If I have bad muck with undeads, I reset. Smiley

- Mute is not worth it against Astos; I kill him in less than a minute. Again, bad luck = reset.
- Heal 2 is tempting, but I am not at level 12 when I reach Melmond the first time, and I just don't think it is worth it late game since the only time I would even use it would be TOFR.
- Warp > Exit because of the Sea Shrine issue you mentioned.

Black:
- Both the Red and Black mages have the same number of level 2 spells at level 4-5 when I fight Astos.
- Getting Fire 2 a level sooner was the biggest temptation for using Black over Red. The issue that I ran into was that the extra time gained from being able to fight ZomBulls earlier (about 3 minutes), is lost in the lack of attack/defense/healing power during every other segment.
- The only advantage to having Fast at level 8 would be against Lich, and Lich dies so fast too that the net time saved would only be maybe 20-30 seconds.
- Getting Warp early would not save any time because I am leveling from level 10-16, so that makes no difference. I never need more than 2 warps at any point in the run.
- Red has Bane too, which I use once against Tiamat.

I'm so glad there is still interest in this! Keep the suggestions coming!

-WarMech

Thanks for the support!  I'd thought you guys might have forgotten about me... :(

To tell you the truth, I'd been lurking these forums every month or two since I took my break and actually downloaded your WIP single segment run.  Obviously, it was just a test, but I thought the principles in it were very solid: Good group choice, good knowledge of power cycles, and good balance between security and speed.  I wish I'd had time to comment on it, but know that I was one of the only people to download and watch the whole thing!   ;D

Yes, I will manipulate the ZomBull/Kyzokus for a one-step encounter.  I didn't in my test runs because they were only tests, but this will save 1-2 seconds per segment in the actual run.

That's a cool bit of info about the encounters and was what I was starting to get at in #2 and 3 in my last post.  I will be adding segments to do the maximum number of steps before an encounter in the later segments when I have enough tents to support it.  Earlier on, I will track exactly what part of the power cycle will generate an ideal set of encounters in that segment.  That knowledge on the number of steps before the first encounter will be critical, although it might be modified for different terrians (water, river, woods) since I think that effects it.

I like the suggestion about going to the dwarf cave first.  We may need to do a little timing but that really sounds like it could be a time saver in the long run.  I was trying to minimize the distance walked at first, but I think getting the house there would more than make up for it.

I see where you are coming from now with the Life and it makes sense.  The question is now whether the time gained from having an extra live if I need it is worth the time gained from a Fire 3 to burn some enemies in TOFR.  You suggestion has much more merit if I decide to go with 2 fighters rather than 2 at the end.

Your reasons are very logical for a single segment, but here is a summary of my counter-argument for the white and black mages:

White:
- Harm doesn't buy me all that much.  If I have bad muck with undeads, I reset.   :)
- Mute is not worth it against Astos; I kill him in less than a minute.  Again, bad luck = reset.
- Heal 2 is tempting, but I am not at level 12 when I reach Melmond the first time, and I just don't think it is worth it late game since the only time I would even use it would be TOFR.
- Warp > Exit because of the Sea Shrine issue you mentioned.

Black:
- Both the Red and Black mages have the same number of level 2 spells at level 4-5 when I fight Astos.
- Getting Fire 2 a level sooner was the biggest temptation for using Black over Red.  The issue that I ran into was that the extra time gained from being able to fight ZomBulls earlier (about 3 minutes), is lost in the lack of attack/defense/healing power during every other segment.
- The only advantage to having Fast at level 8 would be against Lich, and Lich dies so fast too that the net time saved would only be maybe 20-30 seconds.
- Getting Warp early would not save any time because I am leveling from level 10-16, so that makes no difference.  I never need more than 2 warps at any point in the run.
- Red has Bane too, which I use once against Tiamat.

I'm so glad there is still interest in this!  Keep the suggestions coming!

-WarMech

DoubleThink

2009-11-15 06:57:55 pm

There's a word for that

I know the third Fighter isn't much use against Chaos (outside of crits), but how does he stack up in all the other battles leading up to that?

snorlax

2009-11-15 11:53:13 pm

zzzzzzzzzz

I really thought that the number of steps between encounters was much more random than a lot of other things in the game. That's what Ben Siron's FAQ led me to believe, and I tend to trust him.

Winkwonle

2009-11-16 06:04:31 am

I may have spoken a little fast about the number of steps between encounters. I really don't know for sure. I was more suggesting to investigate on the matter than giving actual useful statistics.

Just a little note again on the White mage.

Suppose you're killing the white mage off before farming EYEs, just to stick to the essential FiFiRm, with the unbelievable power and the ability to cast FAST and WARP. At the part where you fight packs of five KYZOKUs and then single ZomBULLs, you could manipulate your way into packs of four GEISTs instead, which you can kill with a single use of HARM, in a single turn, without losing one tempo. These give 468 G and EXP (total). With that, you wouldn't be required to buy Short swords right away. You can even kill these GEISTs before beating Bikke's pirates, since there is a little area right down of Pravoka where they are found.

Let's say hypothetically that this is a worthy suggestion. Then what you want to do is buy HARM and a couple of TENTs as part of your starting gear (or your extended starting gear, that is counting the gold from Garland).

You can burn yourself a free GHOUL inside Garland's as your first battle, then complete Level 2 with Garland, and save at coneria. Then manipulate first-strike encounters, or non-first-strike encounters where you're the first to go and successfully run away, down to the GEIST region just five squares or so below Pravoka, and tent. Fight three rows of 4 GEISTs, tent between each, then go to Pravoka, buy ICE, defeat Bikke, go downtown and buy TENTs and/or PUREs and/or HEALs, but mostly TENTs. Then Inn on your way back, reset and leave Pravoka for good. If you're not satisfied with your ressources as they are then, you can fight GEISTs again at Elfland, some five squares to the west-south-west.

As I said, your white mage will remain well busy until class change, just saving you HEAL potions. You can have him/her die inside the Ice cave, maybe, to avoid sucking up EXP from the others. You can also revive him/her before dungeons and put him/her in the front row to absorb hits. I really don't think a dead character should be so much of a nuisance anyway.

By the way, what I was calling the petrification trick is, like you saw in my earlier test run, allowing a weak character to contribute to a tough boss battle without having to maintain him on the way to the boss. You can get the guy petrified in a couple of places. In a single segment, the entrance of the Ice cave looked best, but in a multi-segment, really, there are petrifying monsters everywhere.

I'm saying that in view of your argument against the Black mage. You pointed out that having a black spell caster is essential, because FAST cuts the boss battles in half, and because WARP is good to have around. Also, you're advocating that the black mage is too weak to maintain inside dungeons. Well, I say, let him be the witness of the fighter's strength, in silence, as a statue, and then when the boss comes, suck up his FAST charges. If you're having only two Fighters travelling alone, you're in fact SAVING time, because they take the least damage. And they can use all the Ribbons they want.

And having FIR2, or LIT2, at Level 5 is great. You don't need those stinking Short swords, when you can just burn the WIZARDs down with a LIT2. You also don't need Silver swords for LICH if you have two magic users around. I know the Short swords were for ZomBULLs, but again, please consider GEISTs instead.

Talking about it really makes me want to try segmented running myself. My FF cartridge is old, though. I got my NES around 1989, and my FF game only a couple of years after. It is not rare that my saved games just disappear after I leave them idling for a few months. I understand that you got that problem during your last running attempt, and you... had your cartridge's batteries replaced ? How does that work ?

In the meantime, you got my entire support. When someone finally completes a segmented run for that game, they'll be doing a great service for humanity !

I may have spoken a little fast about the number of steps between encounters. I really don't know for sure. I was more suggesting to investigate on the matter than giving actual useful statistics.

Just a little note again on the White mage.

Suppose you're killing the white mage off before farming EYEs, just to stick to the essential FiFiRm, with the unbelievable power and the ability to cast FAST and WARP. At the part where you fight packs of five KYZOKUs and then single ZomBULLs, you could manipulate your way into packs of four GEISTs instead, which you can kill with a single use of HARM, in a single turn, without losing one tempo. These give 468 G and EXP (total). With that, you wouldn't be required to buy Short swords right away. You can even kill these GEISTs before beating Bikke's pirates, since there is a little area right down of Pravoka where they are found.

Let's say hypothetically that this is a worthy suggestion. Then what you want to do is buy HARM and a couple of TENTs as part of your starting gear (or your extended starting gear, that is counting the gold from Garland).

You can burn yourself a free GHOUL inside Garland's as your first battle, then complete Level 2 with Garland, and save at coneria. Then manipulate first-strike encounters, or non-first-strike encounters where you're the first to go and successfully run away, down to the GEIST region just five squares or so below Pravoka, and tent. Fight three rows of 4 GEISTs, tent between each, then go to Pravoka, buy ICE, defeat Bikke, go downtown and buy TENTs and/or PUREs and/or HEALs, but mostly TENTs. Then Inn on your way back, reset and leave Pravoka for good. If you're not satisfied with your ressources as they are then, you can fight GEISTs again at Elfland, some five squares to the west-south-west.

As I said, your white mage will remain well busy until class change, just saving you HEAL potions. You can have him/her die inside the Ice cave, maybe, to avoid sucking up EXP from the others. You can also revive him/her before dungeons and put him/her in the front row to absorb hits. I really don't think a dead character should be so much of a nuisance anyway.

By the way, what I was calling the petrification trick is, like you saw in my earlier test run, allowing a weak character to contribute to a tough boss battle without having to maintain him on the way to the boss. You can get the guy petrified in a couple of places. In a single segment, the entrance of the Ice cave looked best, but in a multi-segment, really, there are petrifying monsters everywhere.

I'm saying that in view of your argument against the Black mage. You pointed out that having a black spell caster is essential, because FAST cuts the boss battles in half, and because WARP is good to have around. Also, you're advocating that the black mage is too weak to maintain inside dungeons. Well, I say, let him be the witness of the fighter's strength, in silence, as a statue, and then when the boss comes, suck up his FAST charges. If you're having only two Fighters travelling alone, you're in fact SAVING time, because they take the least damage. And they can use all the Ribbons they want.

And having FIR2, or LIT2, at Level 5 is great. You don't need those stinking Short swords, when you can just burn the WIZARDs down with a LIT2. You also don't need Silver swords for LICH if you have two magic users around. I know the Short swords were for ZomBULLs, but again, please consider GEISTs instead.

Talking about it really makes me want to try segmented running myself. My FF cartridge is old, though. I got my NES around 1989, and my FF game only a couple of years after. It is not rare that my saved games just disappear after I leave them idling for a few months. I understand that you got that problem during your last running attempt, and you... had your cartridge's batteries replaced ? How does that work ?

In the meantime, you got my entire support. When someone finally completes a segmented run for that game, they'll be doing a great service for humanity !

Winkwonle

2009-11-16 08:21:19 am

By the way, now that you're manipulating world map encounters with TENTs, it's definitely best to go to Ordeals by airship, that is, do the Ice cave before Ordeals (provided you do Ordeals at all, but you seemed pretty convinced that it was worth it).

Winkwonle

2009-11-16 01:31:46 pm

In case you're interested, I've just done a test this morning for the GEIST thing. Maybe it's a little hard to do, finally. I mean, they're the second battle on the power cycle, not the first, so it's a little longer to obtain. Also, having the white mage go first is 1 chance in 8. Then having HARM kill all four GEIST is also something like 1 chance in 20, from what I've observed, so all in all it's just much too frustrating to manipulate. On the other hand, if you allow them to hit you for a total of less damage than a TENT heals, then it's really nothing to manipulate at all. So I guess if you think this option is worthy, it's up to you to decide where your patience ends.

Also, I tested putting all the stress on the black mage at the beginning, buying him that juicy LIT2 and roasting the WIZARDs at Level 5, then Astos. It's really a piece of cake. I even got 4 WIZARDS, and it was no problem at all. As far as I'm concerned, I'm pretty sure it's better this way than by putting your ressources in the fighters. Of course, that would be if you're using a black mage, which you're not. I'm just saying.

Finally, I tested not doing Ordeals, just for fun. I was a little shocked to get Defense and not equip it. Normally, I don't care about the Ice sword because of Defense, but now I was without an Ice sword (found in Ordeals) and without a good replacement. Fire sword is OK, I guess, but it takes until Level 13 or something before dealing 3-hits, instead of 11 for Ice sword and 8 for Defense. Lacking armor made no noticeable difference, though. I thought it was interesting to know. Normally I grab the Opal armor, but now I'm thinking I may not do anymore.

I understand that you've done a lot of tests on different aspects of the game, different strategies, teams, and all. Do you know exactly how much time WARP saves in total ? And is there just the Waterfall and the top levels of the Sea Shrine that you use it for, or do you also save some time in villages, like after talking to Unne for example ? And do you grab the Aegis shield and/or the Heal helmet on the first floor of the Mirage tower, then WARP out ? I'd really like to know what the faculty of WARPing represents. And if possible, also, what having only one WARP charge changes... I think it's just something like 10 seconds of walking on a no-encounter floor... And finally, I'd be interested to know what EXIT saves, if you don't have WARP. If you've done these tests already, I'd be pleased if could you share that information. Cheesy

It's just that people have been saying that FiFiFiRm is best, but I'm really not convinced of that yet. Currently the thought of petrifying a Level-8 black mage is haunting me.

Anyway, please say if I'm bringing back too old matters, and if you would really rather that I not question your whole plan. Your plan is perfect, in a way, because you only fight battles that you are forced to or that you have planned for farming purposes. Whatever other plan I suggest can only save the much time that you're spending farming, thus. And I know that's not so much. But it's still something like 40 minutes that we can work with. It's unfortunate that we don't have the time to do tight runs for all strategies that we think of.

By the way, I'm so glad you took the time to check my test run 8)
(In which, incidentally, I use FiFiFiRm, because at the time I thought it was best.)
I looked at it again today, just for fun, and I smiled at the kind of naive way I kill regular monsters and I never farm EXP anywhere. It's true the monsters mostly come from a planned-ahead power cycle, but I may have overestimated the income from that.

In case you're interested, I've just done a test this morning for the GEIST thing. Maybe it's a little hard to do, finally. I mean, they're the second battle on the power cycle, not the first, so it's a little longer to obtain. Also, having the white mage go first is 1 chance in 8. Then having HARM kill all four GEIST is also something like 1 chance in 20, from what I've observed, so all in all it's just much too frustrating to manipulate. On the other hand, if you allow them to hit you for a total of less damage than a TENT heals, then it's really nothing to manipulate at all. So I guess if you think this option is worthy, it's up to you to decide where your patience ends.

Also, I tested putting all the stress on the black mage at the beginning, buying him that juicy LIT2 and roasting the WIZARDs at Level 5, then Astos. It's really a piece of cake. I even got 4 WIZARDS, and it was no problem at all. As far as I'm concerned, I'm pretty sure it's better this way than by putting your ressources in the fighters. Of course, that would be if you're using a black mage, which you're not. I'm just saying.

Finally, I tested not doing Ordeals, just for fun. I was a little shocked to get Defense and not equip it. Normally, I don't care about the Ice sword because of Defense, but now I was without an Ice sword (found in Ordeals) and without a good replacement. Fire sword is OK, I guess, but it takes until Level 13 or something before dealing 3-hits, instead of 11 for Ice sword and 8 for Defense. Lacking armor made no noticeable difference, though. I thought it was interesting to know. Normally I grab the Opal armor, but now I'm thinking I may not do anymore.

I understand that you've done a lot of tests on different aspects of the game, different strategies, teams, and all. Do you know exactly how much time WARP saves in total ? And is there just the Waterfall and the top levels of the Sea Shrine that you use it for, or do you also save some time in villages, like after talking to Unne for example ? And do you grab the Aegis shield and/or the Heal helmet on the first floor of the Mirage tower, then WARP out ? I'd really like to know what the faculty of WARPing represents. And if possible, also, what having only one WARP charge changes... I think it's just something like 10 seconds of walking on a no-encounter floor... And finally, I'd be interested to know what EXIT saves, if you don't have WARP. If you've done these tests already, I'd be pleased if could you share that information. :D

It's just that people have been saying that FiFiFiRm is best, but I'm really not convinced of that yet. Currently the thought of petrifying a Level-8 black mage is haunting me.

Anyway, please say if I'm bringing back too old matters, and if you would really rather that I not question your whole plan. Your plan is perfect, in a way, because you only fight battles that you are forced to or that you have planned for farming purposes. Whatever other plan I suggest can only save the much time that you're spending farming, thus. And I know that's not so much. But it's still something like 40 minutes that we can work with. It's unfortunate that we don't have the time to do tight runs for all strategies that we think of.

By the way, I'm so glad you took the time to check my test run 8)
(In which, incidentally, I use FiFiFiRm, because at the time I thought it was best.)
I looked at it again today, just for fun, and I smiled at the kind of naive way I kill regular monsters and I never farm EXP anywhere. It's true the monsters mostly come from a planned-ahead power cycle, but I may have overestimated the income from that.

WarMech

2009-11-18 04:36:11 pm

So Winkwonle, I never thought I'd see the day where I heard myself say this, but I'm actually giving serious consideration to dropping the third fighter for the white mage and killing him/her off by the time I do the EYE farming. This would, of course, mean completely re-writing most parts of the run.

I'm 99% sure that 4 Geists are more efficient the Kyzokus with the use of Harm. It wouldn't even be necessary to kill them off in one hit on the first hit (although that would be ideal) so long as I could get them dead in one round. I spend 4 rounds and 1:30 to kill 5 Kyzokus and get 300 XP 600 Gold. I would spend 1 round and < :30 to kill 4 Geists and get 468 XP and 468 G. You're probably right; I may not even need to deal with short swords and solo ZomBulls... I could power straight to level 5 and get Harm2.

This also re-opens the idea of using a black mage. Since I'm not fighting Kyzokus and ZomBulls early, the fact that he's as weak as the US economy is irrelevant. Dual black and white mages at level 5 would absolutely own! Even after the ZomBull leveling, they would still be useful for Lich, and I could just kill the white mage off by the time I get the Floater like you said.

That third fighter really is useless. He is probably only 25% as effective as the strongest fighter at best, even in the bosses leading up to Chaos. Not worth it. 2 fighters and 1 black/red mage would be sufficient.

Oooooo... I'm talking myself into it.

OK, here's what Warp does:

Waterfall - Saves 2-3 minutes.
Sea Shrine 2x - Saves 4 minutes and potentially some unavoidable encounters
Melmond (2nd time) - Saves :15
Lefein - Saves :20
TOFR - Saves 2 minutes on a VERY difficult floor after I get the Masamune

This is about 9 minutes which almost complete makes up for the trip to Ordeals. The items that you get on the way, the weapons/armor that you can use as a result of Ordeals, and the magic for the fighters is just icing on the cake and leads me to believe Ordeals is worth it for any of these groups.

I think one way or the other, this is going to require a test. I may run two tests; one with FFFR under the changes we've made since my last test and one with FFWR or FFWB based on which of these we decide is best. It's going to be close because some of my original concerns with white or black mages still exist (slower EYE farming for example) but it's worth anther test!

To clarify, I actually borrowed my cousin's old cartridge. It was not played nearly as much as mine, but I'm always going to be a little nervous.

-WarMech

So Winkwonle, I never thought I'd see the day where I heard myself say this, but I'm actually giving serious consideration to dropping the third fighter for the white mage and killing him/her off by the time I do the EYE farming.  This would, of course, mean completely re-writing most parts of the run.

I'm 99% sure that 4 Geists are more efficient the Kyzokus with the use of Harm.  It wouldn't even be necessary to kill them off in one hit on the first hit (although that would be ideal) so long as I could get them dead in one round.  I spend 4 rounds and 1:30 to kill 5 Kyzokus and get 300 XP 600 Gold.  I would spend 1 round and < :30 to kill 4 Geists and get 468 XP and 468 G.  You're probably right; I may not even need to deal with short swords and solo ZomBulls... I could power straight to level 5 and get Harm2.

This also re-opens the idea of using a black mage.  Since I'm not fighting Kyzokus and ZomBulls early, the fact that he's as weak as the US economy is irrelevant.  Dual black and white mages at level 5 would absolutely own!  Even after the ZomBull leveling, they would still be useful for Lich, and I could just kill the white mage off by the time I get the Floater like you said.

That third fighter really is useless.  He is probably only 25% as effective as the strongest fighter at best, even in the bosses leading up to Chaos.  Not worth it.  2 fighters and 1 black/red mage would be sufficient.

Oooooo... I'm talking myself into it.

OK, here's what Warp does:

Waterfall - Saves 2-3 minutes.
Sea Shrine 2x - Saves 4 minutes and potentially some unavoidable encounters
Melmond (2nd time) - Saves :15
Lefein - Saves :20
TOFR - Saves 2 minutes on a VERY difficult floor after I get the Masamune

This is about 9 minutes which almost complete makes up for the trip to Ordeals.  The items that you get on the way, the weapons/armor that you can use as a result of Ordeals, and the magic for the fighters is just icing on the cake and leads me to believe Ordeals is worth it for any of these groups.

I think one way or the other, this is going to require a test.  I may run two tests; one with FFFR under the changes we've made since my last test and one with FFWR or FFWB based on which of these we decide is best.  It's going to be close because some of my original concerns with white or black mages still exist (slower EYE farming for example) but it's worth anther test!

To clarify, I actually borrowed my cousin's old cartridge.  It was not played nearly as much as mine, but I'm always going to be a little nervous.

-WarMech

ShinerCCC

2009-11-18 06:57:36 pm

Yes, a worthless avatar riding my posts.

Now I know some people are thinking "NOOOOO the run is changing and now it's never going to be finished bawww"

But I just want to say, my Faxanadu run was drastically improved by last-minute changes even though they weren't perfected.

I think I have a copy of FF1 for NES here, so if I feel like it maybe I'll try to learn this stuff too. I have Dawn of Souls too but that doesn't count for anything.

WarMech

2009-11-18 07:40:58 pm

If by "some people" you mean me, then you're absolutely right.

BTW, I'm really looking forward to the new Faxanadu run; I thought the last one was pretty kick-ass so I really can't imagine how you are taking off so much time... Cheesy

Winkwonle

2009-11-21 03:46:26 pm

I'm glad to see that you're really considering my suggestions. On the other hand, I hope that this new uncertainty about the fastest team won't make you abandon the project.

I'd think FFWB would be fastest for a multi-segment, after what we've been saying. The white magic of the Red is cool but it isn't worth so much. Pretty much all dungeons can be done with only one white mage's spell supply, and good armor; no heal potion.

Also, while I was single-segment running, with my current test team being FFRB, I noticed that entering the ToFR at Level 12-13 with everyone is sufficient. Since I'm allowing myself to eventually get to Level 18 with both fighters, entering at Level 15-16 is also not a waste of time. When I was considering where to farm EXP to attain that level, I had three main choices: EYEs, a little more random encounters on the way to CHAOS, and... AGAMAs on the way to KARY.

I'm not sure what it's worth for other teams, and for the extremely low levels that you're imposing on your characters for your multi-segment, but I counted that each AGAMA takes 18 seconds to kill (counting the averaged time for level-ups, potential though absent healing pauses, and surprise battles). To get to Level 15 with three characters, it takes me around 10 minutes. That is faster than the 24 minutes that you plan spending in the Ice cave.

In case you're interested, I've been working on some tools to make my life easier when running. Attached is a poster that I generated and used for planning random encounters. I'm not a graphics designer or anything, so it might not be very aesthetic, but if it may help you then I'm glad to share. Also of course if you happen to know an artist who could make a cooler version (for free), then please provide me with said version...

I'm glad to see that you're really considering my suggestions. On the other hand, I hope that this new uncertainty about the fastest team won't make you abandon the project.

I'd think FFWB would be fastest for a multi-segment, after what we've been saying. The white magic of the Red is cool but it isn't worth so much. Pretty much all dungeons can be done with only one white mage's spell supply, and good armor; no heal potion.

Also, while I was single-segment running, with my current test team being FFRB, I noticed that entering the ToFR at Level 12-13 with everyone is sufficient. Since I'm allowing myself to eventually get to Level 18 with both fighters, entering at Level 15-16 is also not a waste of time. When I was considering where to farm EXP to attain that level, I had three main choices: EYEs, a little more random encounters on the way to CHAOS, and... AGAMAs on the way to KARY.

I'm not sure what it's worth for other teams, and for the extremely low levels that you're imposing on your characters for your multi-segment, but I counted that each AGAMA takes 18 seconds to kill (counting the averaged time for level-ups, potential though absent healing pauses, and surprise battles). To get to Level 15 with three characters, it takes me around 10 minutes. That is faster than the 24 minutes that you plan spending in the Ice cave.

In case you're interested, I've been working on some tools to make my life easier when running. Attached is a poster that I generated and used for planning random encounters. I'm not a graphics designer or anything, so it might not be very aesthetic, but if it may help you then I'm glad to share. Also of course if you happen to know an artist who could make a cooler version (for free), then please provide me with said version...

Attachment:

encounters.pdf

Winkwonle

2009-11-29 02:22:19 am

I have done further tests concerning the number of steps between random encounters. It turns out the rules are simple, and very NOT random. I'm pretty sure that I now have the thing reverse engineered.

Here is how it goes.

The game increments an internal counter at each step you take except on squares where no encounter is possible: automatic battles, stairs, town ground, harbor, spiked squares, etc. It does NOT matter where you are when you take those steps, the counter is incremented by the same amount. That does NOT correspond to what Ben Siron said according to Pat Buns' FAQ. Note that it also doesn't increment when you embark/disembark either the ship or the canoe. (You can cross rivers as much as you want without encounters.)

The number of steps between battles always follows this pattern:

Ship battles 15, 56, 124, 61*
Dungeon battles 15, 24-32, 29-44-40-11, 46-15*
Land battles 15, 24, 32, 29, 17-27, 40, 11, 27-19, 15*
Sky Palace 5F battles 4-11, 7-17, 10-17-5, 22-7, 17, 18-1-4-1-3, 3-24-2-11, 11, 27, 8-11, 10-5*

*When the end of the pattern is reached, there is not an encounter. Instead, the pattern has a 75% chance of getting flipped. For example, after the 124-steps encounter in water terrain, there is not another encounter 61 steps later. Instead, that number is kept aside and merged with the next (possibly flipped) one, either 61 or 15, giving a total of respectively 122 or 76 steps, then 124 or 56, then 56 or 124, and so on.

The rule for flipping the pattern is the following: RRLRLLRL, repeat.

That means the pattern will be read from left to right first (the "right" way), from left to right again, then from right to left, then left to right, etc. If you prefer to think of it in terms of actual flipping, the pattern gets flipped every time except the first of each group of four: don't flip, flip, flip, flip, repeat.

If this wasn't clear enough, here is a quick compilation of steps between ship battles:

15*, 56, 124, 76, 56, 124, 122, 124, 56, 30, 56, 124, 122, 124, 56, 76, 124, 56, 30, 56, 124, 122, 124, 56,
30, 56, 124, 76, 56, 124, 122, 124, 56, 30, 56, 124, 122, 124, 56, 76, 124, 56, 30, 56, 124, 122, 124, 56,
30, 56, ...

You can check it easily by using a TENT in water, and start counting from a fresh POWER. *Note that the square you start on also counts, so it just takes 14 actual movements before the first encounter.

When you go from one zone to another, for example when you board the ship, the step counter remains the same but the pattern is switched. As you may have noticed, the four patterns insert into one another, that is, there is always a dungeon battle when there would be a ship battle, always a land battle when there would be a dungeon battle, and always a sky palace 5F battle when there would be a land battle.

Here is an interesting consequence of this system. If you're on land, and have your ship right beside you, you can avoid most of the land battles by boarding your ship just before taking the "random encounter step", move one square in water, then get back on land. You will not be able to avoid the few encounters that are also scheduled for water terrain, though.

There are still some details that I'd like to confirm. First, I'm not sure that Sky Palace 5F is the only exception to the slack "dungeon" pattern. Second, I'm not 100% sure of the exact number of steps in Sky Palace 5F since I did my tests with a weak party that kept dying. But I'm sure about the steps for outside, on ship, and in regular dungeons.

Regarding this run, knowing this allows to synchronize TENTs with the 40, the 27-40 or the 24-32-29 sequence, and also to obtain more easily that EXP-farming segments start with an immediate encounter. Of course, it also helps planning dungeons ahead of time with an exact plan regarding what to encounter where.

(By the way, if someone uses this piece of information in their FAQ, please mention me somewhere.)

---
Question for the moderators: do single segments have to start by hitting POWER, or can they start from RESET ? Personally I'd say it has to start with POWER, but if it doesn't then I'd be a fool not to profit from it. Indeed, that would mean I'd be able to do massive encounter table manipulation and base my whole run upon it.

I have done further tests concerning the number of steps between random encounters. It turns out the rules are simple, and very NOT random. I'm pretty sure that I now have the thing reverse engineered.

Here is how it goes.

The game increments an internal counter at each step you take except on squares where no encounter is possible: automatic battles, stairs, town ground, harbor, spiked squares, etc. It does NOT matter where you are when you take those steps, the counter is incremented by the same amount. That does NOT correspond to what Ben Siron said according to Pat Buns' FAQ. Note that it also doesn't increment when you embark/disembark either the ship or the canoe. (You can cross rivers as much as you want without encounters.)

The number of steps between battles always follows this pattern:

Ship battles			15, 56, 124, 61*
Dungeon battles		15, 24-32, 29-44-40-11, 46-15*
Land battles		15, 24, 32, 29, 17-27, 40, 11, 27-19, 15*
Sky Palace 5F battles	4-11, 7-17, 10-17-5, 22-7, 17, 18-1-4-1-3, 3-24-2-11, 11, 27, 8-11, 10-5*

*When the end of the pattern is reached, there is not an encounter. Instead, the pattern has a 75% chance of getting flipped. For example, after the 124-steps encounter in water terrain, there is not another encounter 61 steps later. Instead, that number is kept aside and merged with the next (possibly flipped) one, either 61 or 15, giving a total of respectively 122 or 76 steps, then 124 or 56, then 56 or 124, and so on.

The rule for flipping the pattern is the following: RRLRLLRL, repeat.

That means the pattern will be read from left to right first (the "right" way), from left to right again, then from right to left, then left to right, etc. If you prefer to think of it in terms of actual flipping, the pattern gets flipped every time except the first of each group of four: don't flip, flip, flip, flip, repeat.

If this wasn't clear enough, here is a quick compilation of steps between ship battles:

15*, 56, 124, 76, 56, 124, 122, 124, 56, 30, 56, 124, 122, 124, 56, 76, 124, 56, 30, 56, 124, 122, 124, 56,
30, 56, 124, 76, 56, 124, 122, 124, 56, 30, 56, 124, 122, 124, 56, 76, 124, 56, 30, 56, 124, 122, 124, 56,
30, 56, ...

You can check it easily by using a TENT in water, and start counting from a fresh POWER. *Note that the square you start on also counts, so it just takes 14 actual movements before the first encounter.

When you go from one zone to another, for example when you board the ship, the step counter remains the same but the pattern is switched. As you may have noticed, the four patterns insert into one another, that is, there is always a dungeon battle when there would be a ship battle, always a land battle when there would be a dungeon battle, and always a sky palace 5F battle when there would be a land battle.

Here is an interesting consequence of this system. If you're on land, and have your ship right beside you, you can avoid most of the land battles by boarding your ship just before taking the "random encounter step", move one square in water, then get back on land. You will not be able to avoid the few encounters that are also scheduled for water terrain, though.

There are still some details that I'd like to confirm. First, I'm not sure that Sky Palace 5F is the only exception to the slack "dungeon" pattern. Second, I'm not 100% sure of the exact number of steps in Sky Palace 5F since I did my tests with a weak party that kept dying. But I'm sure about the steps for outside, on ship, and in regular dungeons.

Regarding this run, knowing this allows to synchronize TENTs with the 40, the 27-40 or the 24-32-29 sequence, and also to obtain more easily that EXP-farming segments start with an immediate encounter. Of course, it also helps planning dungeons ahead of time with an exact plan regarding what to encounter where.

(By the way, if someone uses this piece of information in their FAQ, please mention me somewhere.)

---
Question for the moderators: do single segments have to start by hitting POWER, or can they start from RESET ? Personally I'd say it has to start with POWER, but if it doesn't then I'd be a fool not to profit from it. Indeed, that would mean I'd be able to do massive encounter table manipulation and base my whole run upon it.

Carcinogen

2009-11-29 06:52:07 am

Quote:

Question for the moderators: do single segments have to start by hitting POWER, or can they start from RESET ? Personally I'd say it has to start with POWER, but if it doesn't then I'd be a fool not to profit from it. Indeed, that would mean I'd be able to do massive encounter table manipulation and base my whole run upon it.

Hi, Speed Demos Archive. Grin

ShinerCCC

2009-11-30 07:32:27 pm

Yes, a worthless avatar riding my posts.

Amazing step counter data, Winkwonle.

I don't think single segment has to start from a hard reset, I've heard of Majora's Mask runs making a new game, playing far enough to learn the Bomber Code, then restarting that file and the same code ends up working.

Essentia

2009-12-02 12:44:23 am

Everything's better with Magitek

A single-segment run starts when you choose to start a New Game; it doesn't matter how long the console has been on or how many times it's been reset before that.

Winkwonle

2009-12-02 04:52:23 am

Ok, thank you for the information ! That opens up a lot of possibilities for optimization, for the better or for the worse. I can just imagine regular people watching a run of their old favorite game and not understanding how it got so easy all of a sudden. Smiley

Paulygon

2009-12-07 05:59:11 am

I like to watch

Hmm, I think most, of not all, of that battle-steps information has been available in the "FF Formulas" sticky topic on GameFAQs for some time now. :-} There's lots of other interesting mechanics info in there, as well.

However, even the steps info there isn't totally complete, from what I remember. Not all dungeon floors are the same. The game uses a table of thresholds for the different dungeon floors (the first of which correspond with overworld terrain types, I think; see FF Hackster for the corresponding floor order). I don't have the table handy, so the exact values may be off slightly. IIRC, a large portion of them are valued at 8. There are some differences, like the elemental ToFR floors (9, 10, 11, 12) and of course the Floating Castle 5F, with its singularly high value of 24. So, not all the same, just mostly. Smiley

I disassembled the code for this once upon a time, but I don't have access to the info right now. Hopefully I'm remembering this correctly. The process is entirely predictable, so there's no percentage chance of flipping patterns. Internally, the way this works is that there is a byte counter for the steps taken and one for the directionality. The steps counter is really more of an index into the pseudo-random number table (PRNG) and the directionality controls which way the steps counter moves. Each time the steps counter passes 255 (or falls below 0), 160 is added to the directionality (discarding multiples of 256), while the steps counter rolls over to 0 (or back to 255). If the directionality value is less than 128, then the counter moves forwards; if 128 or more, then the counter moves backwards. If calculated out, one can derive the A, B pattern mentioned in the gf sticky topic.

When the PRNG table is accessed, if the value is less than the threshold value for that area, then a battle will take place. So, the higher the threshold, the more often this will happen. All of the step-counts listed previously and in the gf sticky can be derived by moving through the PRNG table (taking directionality into account), finding values smaller than the area threshold, and counting the difference between occurrences.

Hmm, I think most, of not all, of that battle-steps information has been available in the "FF Formulas" sticky topic on GameFAQs for some time now.  :-}  There's lots of other interesting mechanics info in there, as well.

However, even the steps info there isn't totally complete, from what I remember.  Not all dungeon floors are the same.  The game uses a table of thresholds for the different dungeon floors (the first of which correspond with overworld terrain types, I think; see FF Hackster for the corresponding floor order).  I don't have the table handy, so the exact values may be off slightly.  IIRC, a large portion of them are valued at 8.  There are some differences, like the elemental ToFR floors (9, 10, 11, 12) and of course the Floating Castle 5F, with its singularly high value of 24.  So, not all the same, just mostly.  :)

I disassembled the code for this once upon a time, but I don't have access to the info right now.  Hopefully I'm remembering this correctly.  The process is entirely predictable, so there's no percentage chance of flipping patterns.  Internally, the way this works is that there is a byte counter for the steps taken and one for the directionality.  The steps counter is really more of an index into the pseudo-random number table (PRNG) and the directionality controls which way the steps counter moves.  Each time the steps counter passes 255 (or falls below 0), 160 is added to the directionality (discarding multiples of 256), while the steps counter rolls over to 0 (or back to 255).  If the directionality value is less than 128, then the counter moves forwards; if 128 or more, then the counter moves backwards.  If calculated out, one can derive the A, B pattern mentioned in the gf sticky topic.

When the PRNG table is accessed, if the value is less than the threshold value for that area, then a battle will take place.  So, the higher the threshold, the more often this will happen.  All of the step-counts listed previously and in the gf sticky can be derived by moving through the PRNG table (taking directionality into account), finding values smaller than the area threshold, and counting the difference between occurrences.

Winkwonle

2009-12-09 11:22:41 pm

You are right, Paulygon. I was unaware of that information although I had looked at the "FF Formulas" some time ago.

Your explanation on the thresholds is enlightening. Not only can the numbers be derived from it, but it also justifies why the step patterns were divided in chunks of 256. Also like you said, adding 160 when crossing table boundaries leads to the same 'don't flip, flip, flip, flip, repeat' rule. If you've been following the directionality for some time, then its next value is entirely predictable indeed.

More on the topic of optimizing a multi-segment run though, it is likely that you're going to be messing around with the reset button for some time without counting steps between battles, just following enemy groups until you are close to your preferred encounter table index. Then you start counting steps if you wish to optimize a little more. What you want to know is that the step pattern "might flip, with a 75% chance". Only when it doesn't "flip" will you know where you are in the cycle and be able to predict when it'll "flip" again. That's what I wanted to express with my "chance of flipping" phrasing.

I haven't found the floor-specific threshold information in Disch's FF Hackster, nor in your Companion, which by the way are both impressive pieces of work! Would it be possible to indicate where that information is situated in Hackster's interface?

I must say that I have much admiration for people of your stature in the FF1 community. There are a lot of things that you know that none of us do here. Please continue to spread that wisdom.

You are right, Paulygon. I was unaware of that information although I had looked at the "FF Formulas" some time ago.

Your explanation on the thresholds is enlightening. Not only can the numbers be derived from it, but it also justifies why the step patterns were divided in chunks of 256. Also like you said, adding 160 when crossing table boundaries leads to the same 'don't flip, flip, flip, flip, repeat' rule. If you've been following the directionality for some time, then its next value is entirely predictable indeed.

More on the topic of optimizing a multi-segment run though, it is likely that you're going to be messing around with the reset button for some time without counting steps between battles, just following enemy groups until you are close to your preferred encounter table index. Then you start counting steps if you wish to optimize a little more. What you want to know is that the step pattern "might flip, with a 75% chance". Only when it doesn't "flip" will you know where you are in the cycle and be able to predict when it'll "flip" again. That's what I wanted to express with my "chance of flipping" phrasing.

I haven't found the floor-specific threshold information in Disch's FF Hackster, nor in your Companion, which by the way are both impressive pieces of work! Would it be possible to indicate where that information is situated in Hackster's interface?

I must say that I have much admiration for people of your stature in the FF1 community. There are a lot of things that you know that none of us do here. Please continue to spread that wisdom.

Paulygon

2009-12-12 01:23:45 am

I like to watch

Thanks! I've been out of the loop for some time, so I'm not sure how much help I can be, but we'll see...

Aha, I now see what you mean about the 75%, thanks for the explanation. I guess I was too used to being able to view the counter values in real time during emulation, heh.

Hmm, sorry if I implied that the threshold information was in Hackster. What I meant to say was that the order of values in the table of thresholds directly corresponds to the list of Standard Maps in Hackster. It's too bad, though.. all it would need is a simple editing box for the number. AFAIK, there is no program that allows editing of those values. Undecided

WarMech

2009-12-14 06:51:15 pm

I haven't abandoned you all again... I promise.

I've spent the last three weeks doing a preliminary test run of a FFWB group to see whether or not this group would save time, and I'm pretty sure that it does, although one run is not enough to draw a firm conclusion.

Here's basically how the run went:
- Instead of Kazokus I did groups of 3 groups of 4 geists. The goal was to kill them all in one round... doing it with one blast of harm was virtually impossible. This means that each segment lasts about 30 seconds rather than 90 for each of the Kz battles.
- I then go up to the PoP at level 3 and manipulate 1 fight of 2 ZomBulls rather than doing 4 battles of 1. Here's the catch though: I let everyone in the battle die except the black mage, and finish off the last ZomBull with him. He gets all the XP and the group gets 200 gold. The probability of success is astronomically low, but the benefit is huge! The black mage is now level 5 and can get Lit2. I'm done my early leveling and am att least 5 minutes ahead of my last run.
- The Lit2 before the March Cave means I manipulate the maximum number of Wizards and kill them all with a Lit2. I still have 3 Ice charges for Astos, so I still haven't lost anything from the Red Mage run.
- This extra gold lets me get Fire2 and Harm2 since the White Mage will be level 5 by now. After Astos, the power leveling segment with 4 ZomBulls goes TONS faster because I can use 1 Fire2 and 1 Harm2 per battle. I can avoid that trip back to the inn if I get the house in the dwarf cave first which I think I'll do.
- The rest of the run up to the Ice Cave is virtually identical.
- When I go to the Ice Cave the first time, I let the white mage die because s/he is useless at that point and will just be a walking liability.
- After getting the Floater and ProRings, I go back to the ice cave and power level my 3 characters... FFB. I thought that without the 3rd fighter and the red mage, this segment could be a bit slower, but that is false. It was actually 18 minutes rather than 24 because I needed 25% fewer fights, and there were 25% fewer level-ups, etc. I keep the Zeus gauntlet because I don't need the 20K gold for the 4th ProRing, so the Black Mage actually does 50 damage to the EYE, which is about what the Red Mage did!
- The rest of the run is again nearly identical
- TOFR is a massive bitch. There's no way around it. My first 300 attempts did not make it past Lich2, and this is the one and only place where having the Red Mage is actually better. Lich almost always uses Nuke in the first round which will always kill the Black Mage and not the Red Mage. This means that a successful Lich2 fight depends entirely upon having all three of my chars attack before Lich... AND kill him. The alternative is hoping that Lich2 doesn't use Nuke AND doesn't 1HKO the Black Mage. It really did take 300 attempts to get past Lich2. Finishing the segment without the Black Mage is impossible. Not having the 3rd fighter made no difference in the times of the boss battles, and it saved some time in healing. However, it was one fewer character to take hits, so a bad encounter (aka: a death) is an instant reset. I also could not find the ideal place in the route to begin this segment. There are SO many possibilities and trade-offs that I was going crazy. Basically, I good route would minimize the number of inescapable encounters (especially Gas D's) and also minimize the risky battles (such as Sorcerers... those little fuckers). It honestly reached the point where even a good part of the segment didn't matter because of crappy luck, so there might actually be a number of places that I try and just accept a few crappy battles if good luck compensates for it.

BUT, it is possible to beat it with this group. After 600-700 attempts (only 3 of which made it past Lich2), I made it to Chaos and killed him on the first attempt. Roll Eyes

RRWB ftw.

Therefore, this being up an interesting dilemma which someone mentioned: Do I formally rewrite the route and abandon the old one in the fear that I lose motivation or ability to complete this run? I'm 90% sure that this FFWB group is faster than the FFFR, but because I have not written the final route and tested it, I cannot be sure. I could also modify the FFFR route to incorporate some of these latest strategies such as killing off the 3rd fighter before power-leveling.

I'm also still looking for a DVD recorder that I described in a previous post. Any ideas? Links?

I'd appreciate some opinions!

-WarMech

I haven't abandoned you all again... I promise.

I've spent the last three weeks doing a preliminary test run of a FFWB group to see whether or not this group would save time, and I'm pretty sure that it does, although one run is not enough to draw a firm conclusion.

Here's basically how the run went:
- Instead of Kazokus I did groups of 3 groups of 4 geists.  The goal was to kill them all in one round... doing it with one blast of harm was virtually impossible.  This means that each segment lasts about 30 seconds rather than 90 for each of the Kz battles.
- I then go up to the PoP at level 3 and manipulate 1 fight of 2 ZomBulls rather than doing 4 battles of 1.  Here's the catch though: I let everyone in the battle die except the black mage, and finish off the last ZomBull with him.  He gets all the XP and the group gets 200 gold.  The probability of success is astronomically low, but the benefit is huge!  The black mage is now level 5 and can get Lit2.  I'm done my early leveling and am att least 5 minutes ahead of my last run.
- The Lit2 before the March Cave means I manipulate the maximum number of Wizards and kill them all with a Lit2.  I still have 3 Ice charges for Astos, so I still haven't lost anything from the Red Mage run.
- This extra gold lets me get Fire2 and Harm2 since the White Mage will be level 5 by now.  After Astos, the power leveling segment with 4 ZomBulls goes TONS faster because I can use 1 Fire2 and 1 Harm2 per battle.  I can avoid that trip back to the inn if I get the house in the dwarf cave first which I think I'll do.
- The rest of the run up to the Ice Cave is virtually identical.
- When I go to the Ice Cave the first time, I let the white mage die because s/he is useless at that point and will just be a walking liability.
- After getting the Floater and ProRings, I go back to the ice cave and power level my 3 characters... FFB.  I thought that without the 3rd fighter and the red mage, this segment could be a bit slower, but that is false.  It was actually 18 minutes rather than 24 because I needed 25% fewer fights, and there were 25% fewer level-ups, etc.  I keep the Zeus gauntlet because I don't need the 20K gold for the 4th ProRing, so the Black Mage actually does 50 damage to the EYE, which is about what the Red Mage did!
- The rest of the run is again nearly identical
- TOFR is a massive bitch.  There's no way around it.  My first 300 attempts did not make it past Lich2, and this is the one and only place where having the Red Mage is actually better.  Lich almost always uses Nuke in the first round which will always kill the Black Mage and not the Red Mage.  This means that a successful Lich2 fight depends entirely upon having all three of my chars attack before Lich... AND kill him.  The alternative is hoping that Lich2 doesn't use Nuke AND doesn't 1HKO the Black Mage.  It really did take 300 attempts to get past Lich2.  Finishing the segment without the Black Mage is impossible.  Not having the 3rd fighter made no difference in the times of the boss battles, and it saved some time in healing.  However, it was one fewer character to take hits, so a bad encounter (aka: a death) is an instant reset.  I also could not find the ideal place in the route to begin this segment.  There are SO many possibilities and trade-offs that I was going crazy.  Basically, I good route would minimize the number of inescapable encounters (especially Gas D's) and also minimize the risky battles (such as Sorcerers... those little fuckers).  It honestly reached the point where even a good part of the segment didn't matter because of crappy luck, so there might actually be a number of places that I try and just accept a few crappy battles if good luck compensates for it.

BUT, it is possible to beat it with this group.  After 600-700 attempts (only 3 of which made it past Lich2), I made it to Chaos and killed him on the first attempt.   ::)  RRWB ftw.

Therefore, this being up an interesting dilemma which someone mentioned: Do I formally rewrite the route and abandon the old one in the fear that I lose motivation or ability to complete this run?  I'm 90% sure that this FFWB group is faster than the FFFR, but because I have not written the final route and tested it, I cannot be sure.  I could also modify the FFFR route to incorporate some of these latest strategies such as killing off the 3rd fighter before power-leveling.

I'm also still looking for a DVD recorder that I described in a previous post.  Any ideas?  Links?

I'd appreciate some opinions!

-WarMech

DoubleThink

2009-12-14 08:04:32 pm

There's a word for that

Will you still test FFWR instead of FFWB or not? I guess you could try keeping the White Mage alive on the way to Lich and praying that they can get a HRM2 in or something.