Video FE10 HM LTC (expected ~125 turns)

[ruadath]

Sorry, I misquoted my notes above; you can in fact clear the chapter in 3 turns without Sniper!Leo, but then you still run into the problem of not having Pass. So maybe it's 3-3 and 3-4 that I should start investigating before I look into actually doing the clear; if anyone has some save files to send along, that would be great.

[ruadath]

Things are looking a little bit grim for the 3-turn clear of 1-E:

It seems impossible to pick up Hammerne in a 4 turn clear of 3-3 without Celerity, and obviously Hammerne is essential to some of the Part 4 strats (more rescue uses). The village is 28 squares away from the closest deployment spot, and Haar has to reach there by turn 3, which means he agonizingly falls 1 space short of reaching his target. I will try to look into this and see if there is any hope, but I doubt it. This is probably good news in some sense in that I don't have to restart the playthrough, but also disappointing for obvious reasons.

Even if I don't end up incorporating the strategy into the playthrough, I will probably record the non Sniper!Leonardo strategy just to document that a 3-turn clear is in fact possible.

[Gradivus.]

God damn it Tellius, should've made mounts rescuable.

[ruadath]

All hope may not be lost. The issue with Haar is that he has to drop someone off on turn 2 in order to burn the NE supplies, which means he can't drop someone off on turn 3 near the Master Crown village -- or so I though. Using Pass!Oscar (or Titania, if I need Axes to break down the door), it might be possible to have him run through the horse kennel (assuming the horses are cooperative) and Give Haar someone on turn 3, so that he can drop them off in the northern area. Assuming this works out, there are two issues to check:

1) In the standard strategy, the person who Haar drops on turn 2 usually gets given Pass, since they get attacked by a bunch of enemies and have to full move the next turn. I'm hoping they can just crit-kill everything.

2) I need to make sure things on the western half of the map work out without a second Paladin helping out. The NW supplies are 37 spaces away from your closest deployment spot, again agonizingly close. Thankfully, having Titania/Oscar fullmove 3 times and drop a person means that if you can shove that person once on part 4, they can reach the supplies. However, the central supplies are 31 squares away, which means Titania probably has to double back to get to those (EDIT: or Titania + Oscar Rescue Drop someone forward on the first turn). If Titania has to double back, that means no Master Crown

The biggest question I need answered right now is this: is the Halberdier who guards the NE supplies passive (does he not move)? And I'm definitely going to need to run through a clear of this to make sure that all of this works before I try anything.

EDIT: Actually I think this still doesn't end up working because I accidentally made the assumption that Haar can drop people off while sitting on the barriers, which isn't true.

Edited March 25, 2016 by ruadath

[ruadath]

On the other hand, we could pursue the train of thought that Hammerne isn't necessary! Let's see how we can go about doing this:

We get the Rescue Staff in 3-E. It has 3 uses.

4-P can be done in 1 turn with no Rescue uses.

4-1 can be done in 2 turns with no Rescue uses.

4-2 can be done in 3 turns with no Rescue uses.

4-3 can be done in 2 turns with 1 Rescue use.

4-E-1 can be done in 2 turns with no Rescue uses.

4-E-2 through 4-E-5 can be done in 1 turn with no Rescue uses.

So this leaves me 1 use of Rescue for each of 4-4 and 4-5. It is unclear whether I will even need both of those, since I imagine 4-5 can be done without Rescue by shoving Reyson a bunch of times and having Pass on a flier (not sure) and 4-4 might be more easily 3-turned with a deployed Seraph Knight, which LD has shown to be possible. I suppose I'll ask him for his Rescue use distribution.

[ruadath]

God damn it Tellius, should've made mounts rescuable.

I don't think even this would save you, since Haar has to drop someone in the NE corner of the map on turn 2 by flying, which just puts you in the same scenario as before.

More annoying is the Indoor Move Penalty for flying units. Without this, we wouldn't need Celerity on Jill for 1-E, which make this whole thing a non-issue (no Pass or Celerity required!)

[ruadath]

... but I need Sniper!Leonardo... sigh... and training him (without a BEXP dump) means starting back from like 1-2...

to give Leo a nice kill in 1-2, it's probably a good idea to go back to 1-P and rig Edward to not get that Def proc (easy now that I'm using tools, especially since it's a player phase level) so he goes into Wrath mode, so this basically just means starting the playthrough over from the beginning.

I'm burned out from playing the same Part 1 chapters over and over again, so I don't think I'm going to do this anytime soon. I'll probably come back to this run around the start of summer, maybe a bit sooner or later, but the point is I'm once again shelving this.

If you have interesting stuff to contribute to the discussions I mentioned above (or really, anything FE10 HM LTC related), feel free to post here, as I'll definitely check in to see what you have to say!

[Vykan_12]

"4-1 can be done in 2 turns with no Rescue uses."

You can 1 turn 4-1 with rescue, although it's really fucking hard to do.

"I'll probably come back to this run around the start of summer, maybe a bit sooner or later, but the point is I'm once again shelving this."

Damn, that sucks. I would honestly just forget about Sniper Leo and continue the run. You're always going to find stuff later on that forces you to go back a bunch of chapters to correct, if you give in to that too much in search of the perfect playthrough, you'll be stuck in this near endless cycle of restarting runs.

[ruadath]

You can 1 turn 4-1 with rescue, although it's really fucking hard to do.

I haven't actually looked at Chiki's strategy yet, but based on his own calculations (odds of <<1/10^{15}), this would require at least 50 bits of RNG to even encode, and in fact since RNGs are pseudorandom generators, it would really require ~75 bits or something like that. Considering that in both GBAFE and FEDS we only have 48 bits (and they aren't even fully utilized; in GBA there is only something <2^32~1 billion state spaces), it's highly unlikely that a seed for that strat exists (even if RD has 64 bits, which I doubt), and furthermore even if it did, it would be nearly impossible for me to find in a reasonable amount of time.

Chiki claims to have edited RNs in Dolphin, something that I'm rather skeptical he actually figured out how to do, but if so he could tell us the memory address and how many bits are allocated to the RNG which would allow us to resolve this issue.

Again, I haven't looked at the strategy yet, but if it turns out some +3 Mt or +whatever crit boosts to Javelins/hand Axes from the Coin forges would reduce those odds to a reasonable amount (maybe something like 1 in a billion), it might actually be an interesting project to see if I can generate a Coin forge seed that allows for getting a good number of these (for those of you who don't know, Coin forge bonuses are randomly generated when you create a new game), and it should be not too difficult to test out by a combination of automation plus cheats that allow for level skipping and dropping a bunch of Coin and gold into the inventory.

In any case, I think 4-5 (not sure about 4-4) can be done without Rescue, so that should leave me a use there if I want/need it. If not, I think incorporating a significantly more reliable (and probably more aesthetically appealing) turn shave in 1-E is superior to incorporating the 4-E-1 turn shave.

EDIT: I didn't see your post on the other thread before I read this, but if you have your "reliable" strat (and remember "reliable" can be as bad as about 1/10,000,000 odds for the EP) for a 4-E-1 turn shave on NM, you could send it to me or post it here and I'll consider whether or not it can be replicated.

Damn, that sucks. I would honestly just forget about Sniper Leo and continue the run. You're always going to find stuff later on that forces you to go back a bunch of chapters to correct, if you give in to that too much in search of the perfect playthrough, you'll be stuck in this near endless cycle of restarting runs.

On the other hand, with a 3-turn clear of 1-E, we can be almost certain that Part 1 is optimized (at least in terms of turn count). It's pretty obvious from map geometry that no other shaves are possible, except for perhaps a 3 turn of 1-7, though coupling indoor move penalties for mounts with the fact that your deploy spots are limited (and we already know the advanced Muarim can't reach the boss by turn 3), this seems rather quite unlikely.

One thing you (or anyone else) might be able to help with though. If you have a save file either near or at 3-4, it would be great if you could either send it to me or check for me if the mage who blocks Ranulf's path on the eastern ledges will attack Provoke!Wind Edge!Ike over Shade!Ranulf who faces ledge disadvantage. If this is the case (and maybe it's obvious, I don't know exactly how Provoke/Shade work in this game), then the GMs actually don't need Pass and I can just continue on from where I left off using the alternative strat.

Edited March 27, 2016 by ruadath

[dondon151]

laguz can't equip shade

[ruadath]

OK fine just regular Ranulf then

[Vykan_12]

I was talking about 4-1, not 4-E-1. Both are 1 turnable.

[ruadath]

I was talking about 4-1. Both are 1 turnable.

Oh lol, I misread that. I think I know how to 1 turn 4-1 with Rescue too. But if 4-5 doesn't require Rescue (and I don't think it does), then this shouldn't be an issue.

In any case, instead of trying to theorycraft 4-4/4-5 clears without Rescue (and more importantly, without good map data available online), I think the most reasonable course of action (especially since I'm taking at least a small break from the run regardless) is just to wait until LD gets his next couple of videos up so we can see what he does. He has already dramatically simplified some of these Part 4 clears, and if we're lucky maybe he'll be able to demonstrate how to do these without Rescue.

Edited March 27, 2016 by ruadath

[ruadath]

FWIW, the alternative 1-E strat the requires Pass on Micaiah but not Sniper!Leo doesn't actually work; I thought it did, but I didn't visualize something correctly. I discovered this while coming painfully close to successfully executing a (new) 3 turn strat of 1-E without Celerity on Jill, but with Pass on Micaiah, and realizing that no mater which route you take, you need to go through 2 enemies on the upper floor to get to Jarod.

[Vykan_12]

I wish you would film some of these near successes, maybe as unlisted videos or something, and I could watch them and help brainstorm ways to pull off your desired turn counts ;)

[ruadath]

I will get something up eventually. Right now I'm a little bit too busy with stuff to get a recording done (I record at like 5fps in order to get the quality as good as it is since Dolphin can only record on a single core and I'm using a laptop), I'm happy to PM you (a rough outline of) the 1-E strat as it currently stands if you want to take a look. Just let me know (send me a PM).

[Junkhead]

FE10 TAS Boyz

[ruadath]

I stumbled upon this great website (http://www7a.biglobe.ne.jp/~hyunkell/index.html) which has map data for all of the Part 4 maps (something that FE Planet is lacking, and which previously made planning very difficult). To recap what we were discussing last time:

On the other hand, we could pursue the train of thought that Hammerne isn't necessary! Let's see how we can go about doing this:

We get the Rescue Staff in 3-E. It has 3 uses.

4-P can be done in 1 turn with no Rescue uses.

4-1 can be done in 2 turns with no Rescue uses.

4-2 can be done in 3 turns with no Rescue uses.

4-3 can be done in 2 turns with 1 Rescue use.

4-E-1 can be done in 2 turns with no Rescue uses.

4-E-2 through 4-E-5 can be done in 1 turn with no Rescue uses.

So this leaves me 1 use of Rescue for each of 4-4 and 4-5. It is unclear whether I will even need both of those, since I imagine 4-5 can be done without Rescue by shoving Reyson a bunch of times and having Pass on a flier (not sure) and 4-4 might be more easily 3-turned with a deployed Seraph Knight, which LD has shown to be possible. I suppose I'll ask him for his Rescue use distribution.

56282058 says that 4-4 can be done without using Rescue, though I've yet to figure out how this will be done (although now that I have map data this is something I can probably work on by myself). In addition, I can now confirm that 4-5 can be done without Rescue as long as I either have Celerity or Smite + Wildheart (I'll need 2 if Tibarn can't shove Marcia, but I think he can) available for the map.

Assuming I can figure out a way to not use Rescue in either map, this means I end up using Rescue only once in 4-3! So I'll have a couple of uses left to play around with potential 1-turns in 4-1 and 4-E-1. Even if I don't manage to do both, it'll still be good enough to equal the current record for Part 4, so this is some great news for the viability of the 1-E 3-turn clear!

[ruadath]

I'm still not planning to return to serious work on this playthrough for some time, but the first 4 chapters (1-P through 1-3, completed in 13 turns) are pretty much dictated by the 3-turn clear of 1-3, so I did those. The minor changes to the strats aren't worth making new videos but here are my notes on the differences:

1-P: Same as before, except Edward's level rigged to not get Def.

1-1: Same as before, Edward needs to dodge one of the attacks because of lower Def.

1-2: Biggest change, basically Leo gets a kill instead of Nolan, and having Edward go into Wrath mode was very helpful. Net changes were +27 EXP Leo, -1 EXP Sothe (he won't miss it), -something Nolan which doesn't matter. Leonardo ends the chapter on 4.73 EXP

1-3: Same as before, except I had Ilyana use Thunder instead of Elthunder, and on turn 2 EP I had Aimee and Kurth charge into the enemy reinforcements, since you don't get anything for them surviving, and every round of combat they survive is an additional point of BEXP. Probably only a gain of like 4 or 5 BEXP, but whatever.

1-4 is where the big question of Part 1 comes into play: how do we optimize BEXP allocation with regards to Laura vs Leonardo? This is unclear to me and something I'm still thinking about, and would appreciate some feedback on. Obviously, it's best to spend BEXP every time we get it, but unclear who will benefit the most from it. Laura certainly gets the most immediate benefit, since taking her from level 1 to level 3 literally triples her CEXP gain, and I also imagine that I would not want to take Leo above level 4 for 1-4, since my current plan will have him steal a kill from Ilyana, which will allow Laura to heal him when he levels up. On the other hand, it's unclear if having a stronger Leo will allow for me to get him more levels in 1-5, which could prove an excellent training ground for him; he will probably stealing kills from Ilyana, but she can probably weaken them down to like 1-2 HP before hand, so I don't know. Leo doesn't actually get a ton of time to train after that; I imagine his best bet is in Chapter 1-8, where he can take out some of the mages near the starting position on the EP (by letting Micaiah and Ilyana wall for him). Maybe he can pick up a kill or two in 1-6-2 and/or 1-7 but it doesn't seem all that likely, so... yeah. His target is to reach level 11 before promoting at the start 1-E, by the way.

This brings up another question somewhat related to the first: is it still worth it to purchase Mend? The benefits seemed pretty decent in my last run through Part 1, especially since it let Laura reach a level higher than she would have without it (for net savings of somewhere between 100-200 BEXP I think). However, if Laura isn't getting her full share of the BEXP than the benefits might be less, but more importantly I will be spending 10,000 more gold in this run on a Secret Book and Iron Longbow, so it could be more important to preserve cash over BEXP. That figure isn't even counting the Olivi Grass (1500 G) I need in 1-E (I thought I could weasel my way around that in a 4-turn clear), or whether it may be useful to forge an Iron Bow (~1000 G) for Leo at some point so he can kill things.

So definitely a lot to think about, and as always, I appreciate some comments regarding these issues.

[ruadath]

So this is reasonably embarrassing, but...

After I made the above post, I spent some time thinking about how to have Leo kill the boss in 1-4 and I actually came up with a pretty cool idea which involved Shade!Nolan (obviously, I didn't check to see if it actually worked). However, while thinking about this further, I've actually come to the realization that... Sniper!Leo might actually not be necessary for the 3-turn clear of 1-E. Unfortunately, I'm still busy with stuff regardless, so I'm not sure when I'll have time to sit down and check this out (it may rely on some enemy AI stuff), but hopefully it works!

Note: I'm assuming in this new strat that Nailah can Shove Volug. Obviously I'm going to check this when I do testing, but if one of you guys knows for sure that she can't do this, then I won't bother.

Edited April 5, 2016 by ruadath

[Vapo]

For Nailah shoving Volug I'm assuming you mean untransformed Volug since I don't think transformed laguz can be shoved: Nailah's transformed Cn is 19 and Volug's untransformed weight is 18, so you should be fine.

[ruadath]

I mean both transformed (you can't untransform Volug in part 1). I'm pretty sure I've seen transformed Muarim shove transformed Vika and the like at some point, but I'm not sure. If she can't shove him, then yeah I guess we still have to go with Sniper!Leo

[Vapo]

Oh, yeah, forgot about that. I'm pretty sure that any laguz that isn't flying has a weight too high to be pushed, since Volug's transformed weight is 40 and Nailah's transformed Cn is only 19.

[ruadath]

Ah, OK, just checked and you're right. That's a real shame, but I guess that makes this more fun. Thanks for the help!

[Vapo]

No problem. I'll be interested to see how you're going to get sub 130 (on Hard Mode of all difficulties) since the last time I was looking at LTC for this game (which was a while ago) people thought getting sub 140 was either impossible or very hard. I guess part of it is you not objecting to extreme RNG rigging.