AGDQ 2014 practice thread

2013-10-25 05:25:40 pm

We should probably go to private messages on here or IRC to avoid cluttering this topic further, but to answer your question:

Quote:

To clarify: Is the basic idea to choreograph cursor movements/selections/whatever in such a way that you know exactly what is going to happen every time, with no random level ups/random hits/random crits?

This is exactly what the run is. I have everything planned out exactly and as long as I don't make any cursor mistakes or move someone in the wrong order, I know everything that's going to happen ahead of time. That's why I can rely on crazy criticals that would mean certain death otherwise (Athos at 1 HP versus Nergal, for example) and other things like that. The "extra restarts" were just because I messed something up in the plan and had to reset to go back and fix it. That's something I hope to eliminate entirely with enough practice.

mikwuyma

2013-10-25 08:18:26 pm

My feelings on The Demon Rush

Molotov, could you try some pure SS/less manipulated runs? The reason why I'm asking is that a run with perfect RNG manipulation might look dry to viewers, and I'm wondering if it's possible to craft a route with less manipulation. While a "pure" run might not be as perfect, it might be more exciting for viewers to see a higher risk of death.

ButtersBB

2013-10-25 09:57:32 pm

To add to what Mike said, donation incentives are something to keep in mind. People might be willing to donate for extra things, like recruit canas or get A support with Eliwood/Ninian or whatever, and if it is so structured that this is not an option it would be kind of a bummer.

yoshifan

Edit history:

yoshifan: 2013-10-25 10:48:22 pm

yoshifan: 2013-10-25 10:47:06 pm

2013-10-25 10:31:04 pm

I can kind of see Molotov's side of things here, that "not manipulating" can end up feeling artificial or feeling like purposely playing poorly. Here's a kind of scenario that can happen if I'm understanding right (please correct me if I'm not):

Say you're on Chapter 1, and say that you're a somewhat heavy resetter, so you play this chapter a lot. On your first move of the chapter, you have to move one space north and one space east to attack the first enemy from a favorable position. After enough resets, you notice that you always critical this enemy if you select your path by moving the cursor north then east; meanwhile, you never critical the enemy if you move the cursor east then north. So from then on, you always end up going north and east.

This would be a really simple form of manipulation, and perhaps this kind of scenario is similar to how RNG manipulation was discovered in the first place. The question is, how far could you go without going overboard - just the first enemy, or just the first chapter? Go as far as you can through the game without ever canceling your path selection, maybe? If you don't go all the way, what you're doing may be seen as arbitrary.

Not trying to take sides here, just trying to point some things out if I can.

Oh, as for my practice:

As of a few weeks ago, I've been working on Galaxy 2 whenever I get the chance, and plan to keep working on it solidly all the way to AGDQ. I did a rusty run last weekend which got 3:32:49; it's certainly not where I want to be, but it's better than I expected to get at this point (PB is 3:28, goal is ~3:15).

Most of my recent work has been using the Dolphin emulator to better understand SMG2's mechanics and revisit strategies that I was never sure about. I figured it was a good idea to do that before resuming overall practice. In the past 2 or 3 weekends, I did heavy testing and grinding of a few key stages in World 1 (progress on my YouTube). That has helped a lot so far, so I'm going to do that on as many stages as I have time for.

I can kind of see Molotov's side of things here, that "not manipulating" can end up feeling artificial or feeling like purposely playing poorly. Here's a kind of scenario that can happen if I'm understanding right (please correct me if I'm not):

Say you're on Chapter 1, and say that you're a somewhat heavy resetter, so you play this chapter a lot. On your first move of the chapter, you have to move one space north and one space east to attack the first enemy from a favorable position. After enough resets, you notice that you always critical this enemy if you select your path by moving the cursor north then east; meanwhile, you never critical the enemy if you move the cursor east then north. So from then on, you always end up going north and east.

This would be a really simple form of manipulation, and perhaps this kind of scenario is similar to how RNG manipulation was discovered in the first place. The question is, how far could you go without going overboard - just the first enemy, or just the first chapter? Go as far as you can through the game without ever canceling your path selection, maybe? If you don't go all the way, what you're doing may be seen as arbitrary.

Not trying to take sides here, just trying to point some things out if I can.

Oh, as for my practice:

As of a few weeks ago, I've been working on Galaxy 2 whenever I get the chance, and plan to keep working on it solidly all the way to AGDQ. I did a rusty run last weekend which got 3:32:49; it's certainly not where I want to be, but it's better than I expected to get at this point (PB is 3:28, goal is ~3:15).

Most of my recent work has been using the Dolphin emulator to better understand SMG2's mechanics and revisit strategies that I was never sure about. I figured it was a good idea to do that before resuming overall practice. In the past 2 or 3 weekends, I did heavy testing and grinding of a few key stages in World 1 (progress on [url=http://www.youtube.com/yoshifan28]my YouTube[/url]). That has helped a lot so far, so I'm going to do that on as many stages as I have time for.

Molotov

2013-10-26 05:23:25 pm

Disclaimer: This is a very long-winded post because I hope to clear some stuff up once and for all. I probably repeated myself several times throughout the post unintentionally, but meh. Hopefully my wording came off diplomatically, because I mean no offense with any of this. For those who want the quick version, I've included a short summary at the bottom.

I hate that this is going so off-topic and my run is picking up the wrong kind of attention (to the point where I almost regret offering it), but I feel like I'm in some kind of bizarro world here. Getting through the marathon selection process and then being told how to run my game? Not only that, but I'm being told to add unnecessary risk to the run. Surely I'm not the only one who finds this concept ridiculous. Not to imply he's taking sides, but Yoshifan absolutely hit the nail on the head with his post. Even if I go for a "pure single-segment" or "less manipulation" run, where do I draw the line? If I plan out a no-reset run fully and practice it enough, I could possibly end up with most of the run having the same memorized look. Honestly, it would be stupid if I didn't. If I know that moving Florina to a certain tile during an enemy phase will provoke enough attacks that her next attack on the player phase will be a critical and kill a certain boss, why the heck would I do anything else? Am I supposed to go back to 2006 or something and pretend that I don't know that the RNG patterns can be used to my advantage even if I don't directly manipulate it? The idea is absurd and it's part of why I disagree so strongly with what's being asked of me.

I think there are multiple misunderstandings going on here and that's really unfortunate. The first is that somehow this run involves intense RNG manipulation. This isn't true at all. If I really wanted to, I could've used the cursor to rig far more criticals and level ups in my favor. This is real RNG manipulation in my mind, because I'm literally going out of my way to bend it to my will. If you watch the run, you'll notice that I didn't do that very often. This is partly because I felt that too much direct manipulation would turn the run into a bad segmented run or bad TAS. I purposely avoided doing this even when it might have been helpful because I knew it wouldn't be interesting to watch. Instead, I opted to just achieve most of the results I wanted through movement orders. I'm not really manipulating anything in this case, but instead I'm just recognizing a fixed pattern and using it to my advantage. Again, Yoshifan's post covers this concept perfectly.

Just in case there's any confusion about RNG patterns and manipulation, I mean something like this: Unit A attacks an enemy and misses. Unit B then attacks the same enemy and hits. Had Unit B attacked first, it would've resulted in a critical and killed the enemy. Purposely ignoring this pattern and having Unit A attack first in any run after that would be incredibly stupid.

Alternatively, there's this scenario (this actually happens at the end of Chapter 27, so I'm not making this up): Unit A goes into a horde of enemies that will either result in saving a turn or certain death. I move the unit in and end my turn. The enemies all gang up on this guy and he dies before killing them all, costing me both a turn and a valuable unit. Now go back to before the move and note that I directly manipulate the RNG with one very simple cursor movement. I then move the unit in and end my turn. Not only does the unit survive, but he kills all enemies and ends the map a turn earlier. Am I supposed to pretend that this wouldn't be a good idea for future runs when I know this?

Basically, I take offense to being asked to purposely pretend that I don't know the RNG is fixed and play based on only an overall plan with a few specifics here and there. This will lead to one or both of two things: A) a sloppy run that chat will jump on, or B) a run so heavily practiced that I noticed favorable RNG patterns and ending up memorizing things anyway (repeating the cycle). I can't stress this enough, because I think there's a disagreement about what makes a strategy RPG run interesting. This varies by person, but I don't think all of the guidelines for other marathon games apply to this genre. In this case, more entertainment and wow factor would come from a low turn count balanced with direct RNG rigging (which I keep to near minimum). Achieving marathon safety by using higher turn, slower strategies will be immediately obvious to a viewer and I'm telling you, they won't like it. They won't like a lot of RNG rigging either, so that's partly why it isn't used much. However, some rigging will help the turn count and demonstrate an in-depth knowledge of the game, which I think viewers will appreciate.

Quote from mikwuyma:

Molotov, could you try some pure SS/less manipulated runs? The reason why I'm asking is that a run with perfect RNG manipulation might look dry to viewers, and I'm wondering if it's possible to craft a route with less manipulation. While a "pure" run might not be as perfect, it might be more exciting for viewers to see a higher risk of death.

I'm not doing "perfect" manipulation at all. Perhaps it isn't the best comparison, but imagine knowing that a standing jump will manipulate a beat 'em up enemy into doing something favorable that lets you get an infinite loop going. This speeds up the run considerably. Now imagine being asked to forget that you know that and play in a "less manipulated" way. This is basically what's being asked here. I'm being told to forget that I know how the RNG works, even though it's critical knowledge to anybody doing a serious run for this game.

Quote from ButtersBB:

To add to what Mike said, donation incentives are something to keep in mind. People might be willing to donate for extra things, like recruit canas or get A support with Eliwood/Ninian or whatever, and if it is so structured that this is not an option it would be kind of a bummer.

I've mentioned donation incentives to Mike more than once, but the problem is the bidding war. It doesn't seem like I can offer as much as I might like because both games would need to share the incentive for it to work. Really, the only realistic incentive they share is "all characters." Support endings like Eliwood/Ninian A wouldn't be a bad idea (in fact, choosing Eliwood's partner at all would be a decent incentive), but it's not common between the games.

I understand that what you're getting at is that I can't use on-the-spot incentives because of the structured approach, but those kind of donations would probably be a bad idea for either of these games anyway. Throwing in new elements to a run in a genre that relies heavily on planning and in a one-shot marathon setting seems like a recipe for disaster to me.

Too long; didn't read version: I'll at least try what you're asking, but for the record, I strongly disapprove of that approach and I make no promises that it will even work out. I'm convinced this intentional disregard of RNG patterns/mechanics wouldn't go over well with viewers at all.

Disclaimer: This is a very long-winded post because I hope to clear some stuff up once and for all. I probably repeated myself several times throughout the post unintentionally, but meh. Hopefully my wording came off diplomatically, because I mean no offense with any of this. For those who want the quick version, I've included a short summary at the bottom.

I hate that this is going so off-topic and my run is picking up the wrong kind of attention (to the point where I almost regret offering it), but I feel like I'm in some kind of bizarro world here. Getting through the marathon selection process and then being told how to run my game? Not only that, but I'm being told to [i]add[/i] unnecessary risk to the run. Surely I'm not the only one who finds this concept ridiculous. Not to imply he's taking sides, but Yoshifan absolutely hit the nail on the head with his post. Even if I go for a "pure single-segment" or "less manipulation" run, where do I draw the line? If I plan out a no-reset run fully and practice it enough, I could possibly end up with most of the run having the same memorized look. Honestly, it would be stupid if I didn't. If I know that moving Florina to a certain tile during an enemy phase will provoke enough attacks that her next attack on the player phase will be a critical and kill a certain boss, why the heck would I do anything else? Am I supposed to go back to 2006 or something and pretend that I don't know that the RNG patterns can be used to my advantage even if I don't directly manipulate it? The idea is absurd and it's part of why I disagree so strongly with what's being asked of me.

I think there are multiple misunderstandings going on here and that's really unfortunate. The first is that somehow this run involves intense RNG manipulation. This isn't true at all. If I really wanted to, I could've used the cursor to rig far more criticals and level ups in my favor. This is real RNG manipulation in my mind, because I'm literally going out of my way to bend it to my will. If you watch the run, you'll notice that I didn't do that very often. This is partly because I felt that too much direct manipulation would turn the run into a bad segmented run or bad TAS. I purposely avoided doing this even when it might have been helpful [i]because[/i] I knew it wouldn't be interesting to watch. Instead, I opted to just achieve most of the results I wanted through movement orders. I'm not really manipulating anything in this case, but instead I'm just recognizing a fixed pattern and using it to my advantage. Again, Yoshifan's post covers this concept perfectly.

Just in case there's any confusion about RNG patterns and manipulation, I mean something like this: Unit A attacks an enemy and misses. Unit B then attacks the same enemy and hits. Had Unit B attacked first, it would've resulted in a critical and killed the enemy. Purposely ignoring this pattern and having Unit A attack first in any run after that would be incredibly stupid.

Alternatively, there's this scenario (this actually happens at the end of Chapter 27, so I'm not making this up): Unit A goes into a horde of enemies that will either result in saving a turn or certain death. I move the unit in and end my turn. The enemies all gang up on this guy and he dies before killing them all, costing me both a turn and a valuable unit. Now go back to before the move and note that I directly manipulate the RNG with one very simple cursor movement. I then move the unit in and end my turn. Not only does the unit survive, but he kills all enemies and ends the map a turn earlier. Am I supposed to pretend that this wouldn't be a good idea for future runs when I know this?

Basically, I take offense to being asked to purposely pretend that I don't know the RNG is fixed and play based on only an overall plan with a few specifics here and there. This will lead to one or both of two things: A) a sloppy run that chat [i]will[/i] jump on, or B) a run so heavily practiced that I noticed favorable RNG patterns and ending up memorizing things anyway (repeating the cycle). I can't stress this enough, because I think there's a disagreement about what makes a strategy RPG run interesting. This varies by person, but I don't think all of the guidelines for other marathon games apply to this genre. In this case, more entertainment and wow factor would come from a low turn count balanced with direct RNG rigging (which I keep to near minimum). Achieving marathon safety by using higher turn, slower strategies will be immediately obvious to a viewer and I'm telling you, they won't like it. They won't like a lot of RNG rigging either, so that's partly why it isn't used much. However, some rigging will help the turn count and demonstrate an in-depth knowledge of the game, which I think viewers will appreciate.

[quote="mikwuyma"]Molotov, could you try some pure SS/less manipulated runs? The reason why I'm asking is that a run with perfect RNG manipulation might look dry to viewers, and I'm wondering if it's possible to craft a route with less manipulation. While a "pure" run might not be as perfect, it might be more exciting for viewers to see a higher risk of death.[/quote]
I'm not doing "perfect" manipulation at all. Perhaps it isn't the best comparison, but imagine knowing that a standing jump will manipulate a beat 'em up enemy into doing something favorable that lets you get an infinite loop going. This speeds up the run considerably. Now imagine being asked to forget that you know that and play in a "less manipulated" way. This is basically what's being asked here. I'm being told to forget that I know how the RNG works, even though it's critical knowledge to anybody doing a serious run for this game.

[quote="ButtersBB"]To add to what Mike said, donation incentives are something to keep in mind. People might be willing to donate for extra things, like recruit canas or get A support with Eliwood/Ninian or whatever, and if it is so structured that this is not an option it would be kind of a bummer.[/quote]
I've mentioned donation incentives to Mike more than once, but the problem is the bidding war. It doesn't seem like I can offer as much as I might like because both games would need to share the incentive for it to work. Really, the only realistic incentive they share is "all characters." Support endings like Eliwood/Ninian A wouldn't be a bad idea (in fact, choosing Eliwood's partner at all would be a decent incentive), but it's not common between the games.

I understand that what you're getting at is that I can't use on-the-spot incentives because of the structured approach, but those kind of donations would probably be a bad idea for either of these games anyway. Throwing in new elements to a run in a genre that relies heavily on planning and in a one-shot marathon setting seems like a recipe for disaster to me.

[b]Too long; didn't read version[/b]: I'll at least try what you're asking, but for the record, I [b]strongly[/b] disapprove of that approach and I make no promises that it will even work out. I'm convinced this intentional disregard of RNG patterns/mechanics wouldn't go over well with viewers at all.

quinntendopower

2013-10-26 06:47:05 pm

Takahashi Meijin

Coming from a viewer standpoint, I think I'd rather see someone with a mastery of the RNG and commenting on specific things then winging it and hoping for the best instead.

ButtersBB

2013-10-26 07:30:50 pm

Quote from quinntendopower:

Coming from a viewer standpoint, I think I'd rather see someone with a mastery of the RNG and commenting on specific things then winging it and hoping for the best instead.

The Sacred Stones run being offered doesn't manipulate RNG, but is far from "winging it." There are plenty of strategies and things to comment on, it is just a different type of run.

For the record though, Seth is a far different character from Marcus, which makes a No RNG Abuse run much more doable. I'm sure it can be done for FE7, but it would probably be harder, and the amount of time needed to route it out makes it not really viable at this stage, especially if Molotov isn't really feeling it.

I think it is important to note that choosing not to manipulate RNG isn't "intentionally playing badly." If you aren't manipulating RNG, then it allows you to come up with different types of strategies instead.

After thinking about it a bit, I think the fact that the runs aren't just different games but also different categories can make the bid war a bit more interesting. After all, a popular debate amongst FE fans is RNG Abuse vs. No RNG Abuse (some swear by it, some scoff at it).

[quote="quinntendopower"]Coming from a viewer standpoint, I think I'd rather see someone with a mastery of the RNG and commenting on specific things then winging it and hoping for the best instead. [/quote]

The Sacred Stones run being offered doesn't manipulate RNG, but is far from "winging it." There are plenty of strategies and things to comment on, it is just a different type of run.

For the record though, Seth is a far different character from Marcus, which makes a No RNG Abuse run much more doable. I'm sure it can be done for FE7, but it would probably be harder, and the amount of time needed to route it out makes it not really viable at this stage, especially if Molotov isn't really feeling it.

I think it is important to note that choosing not to manipulate RNG isn't "intentionally playing badly." If you aren't manipulating RNG, then it allows you to come up with different types of strategies instead.

After thinking about it a bit, I think the fact that the runs aren't just different games but also different categories can make the bid war a bit more interesting. After all, a popular debate amongst FE fans is RNG Abuse vs. No RNG Abuse (some swear by it, some scoff at it).

TheMG2

2013-10-26 07:41:35 pm

Molotov literally just said he isn't heavily manipulating RNG though so I don't get what you're saying.

Patashu

2013-10-26 07:55:37 pm

I am on Molotov's side. It would be like asking for someone to make an RPG TAS with no luck manipulation in it - the very idea is self-contradictory to the point of self-destruction, because the better a TAS gets in terms of optimization, the more it manipulates by default. To forego easy manipulation is to say 'do a worse run than you could'.

ButtersBB

2013-10-26 07:56:25 pm

Quote from TheMG2:

Molotov literally just said he isn't heavily manipulating RNG though so I don't get what you're saying.

He is, just not as heavily as he could be.

Basically everything is fixed, he knows what will happen, but if he wanted to really go in hard on RNG manipulation, then there would be far more moving the cursor around to get better levels and crits and whatnot.

His route is fixed and removes randomness, but it is more simple for marathon purposes. Reason for this is that if he tries to abuse the RNG TOO heavily, then the run becomes way less safe for a marathon setting. There is already the risk of messing up a cursor movement and restarting as it is, and this is amplified if you are trying to be too optimal.

Bismuth

2013-10-27 11:35:32 am

Worthless categories WR master

There are some very good news for the Minecraft speedrun.

1.

2. I found a stupidly simple way of backing up a disastrous mistake. Since each duplication glitch iteration doubles the duped items, I can dupe once more before nether, and once more before The End, and throw the second copy of all my useful items in the nearest chest. That way, if I burn in lava or get knocked off by an asshole dragon, I get an instant back up. That would probably take about 45 seconds and save me a lot of headaches to find a way to come back from being knocked off the map instantly in The End (improbable but disastrous scenario).

Sloth

2013-10-27 12:50:34 pm

Did some Any% Dishonored practice yesterday. A couple good runs, faster than my record at many points but perfect RNG is just too hard to beat.
http://www.twitch.tv/decidedsloth/b/473825931

Dime

2013-10-27 03:38:11 pm

http://www.twitch.tv/kingdime/c/3150714

New PB for Plutonia

the total time comes out to 29 minutes and 33 seconds. Obviously it's not a single segment
and I died once on Omen but still it shows that this game can be pushed below 30 minutes.
Just need to hit that God SS run Sad

And of course been practicing Ultimate doom as well.

cyghfer

Edit history:

cyghfer: 2013-10-27 11:21:00 pm

2013-10-27 10:44:12 pm

i got the mm2 record. http://www.twitch.tv/cyghfer/c/3152838

blechy & i have been steadily working on chip n dale, we will probably stream that sometime in november.

i will start playing bucky in november also.

mikwuyma

2013-10-28 03:36:36 pm

My feelings on The Demon Rush

Molotov:

Fire Emblem RNG stuff: Okay that's a good explanation, I thought you were going to segmented/tas-levels of rng manipulation, that was my worry because that would be marathon unsafe (on top of looking kind of ridiculous). You do make a good point about knowing the game well for a speedrun.

Donation Incentives: Yeah...it's too bad that the bid war for the games is good, but also cuts into other incentives. I'm guessing there's no similar incentives for the two games (like all recruits or special endings)?

luigimeister

2013-10-29 05:44:55 am

Likes SoM far too much

http://www.twitch.tv/luigimeister/b/474584482 I got a 2:19 in Secret of Mana. Run starts at like 20 mins in. RNG was so strong in this run, but I'm still PBing somehow >_>

Vykan12

Edit history:

Vykan12: 2013-10-29 01:02:28 pm

2013-10-29 01:01:15 pm

Uwee Hee Hee

As someone who's heavily invested in how Fire Emblem gets portrayed in the marathon, I have to side with Molotov as well.

First, it's very important to understand some recent developments in GBA Fire Emblem speedrunning. From TASing the game it has been discovered that the game's RNG can be manipulated through path drawing, and is time independent. Quoting from my own Sacred Stones TAS:

Quote:

Luck is manipulated in the run by abusing the game’s path drawing algorithm. When there are multiple ways to reach the same location (eg/ by pressing a diagonal input), the game’s random number generator determines which path to take. In fact, the path drawing algorithm is quite straightforward. For a simple diagonal path, random numbers between 0-49 will force the path to end horizontally, whereas RNs between 50-99 will force the path to end vertically.

It should come as no surprise that the most efficient way to burn RNs is to draw a diagonal path as far as a character can move, then “wobble” the cursor back and forth between 2 spots, forcing the RNG to make as many decisions about the path drawn as possible.

It’s also known that the RNG is dictated by a seed of 3 values that uses some complicated formula to generate a 4th value, then the 1st through 4th value are used to generate a 5th, and so on indefinitely. There’s a widely used emulator lua script that actually shows a table of RN values on the right side of the screen that updates in real-time.

The most recent discovery is that saving & resetting the game always causes the same RNG seed to be used- 8, 56, 21. This is extremely useful since you can make every chapter in the game have the same RNG sequence by deliberately saving & resetting and planning out your movements accordingly.

The implications of this are that you can plan out every last move of your run on emulator, memorize that as best as possible, then copy those same moves on console to achieve the desired result. This is a labor intensive process, but completely worth it compared to “winging it”. In an RNG planned run, if you make a cursor mistake, you only have to re-do said chapter since the RNG always resets to the 8,56,21 state as mentioned earlier. In a winged run, entire strategies can crumble if certain units don’t meet certain stat benchmarks or important hits miss or a player unit gets hit one too many times on a long enemy phase and dies. If, for instance, Seth gets str and spd screwed in a Sacred Stones run, your whole run falls apart since the run relies so heavily on him carrying the offensive load.

Another thing to consider- It’s actually pretty difficult to mess up a cursor movement since the only problematic movement type is a diagonal movement in a wide open map. Most movement is either straight movement or in a confined space, where diagonal movement can be avoided by taking advantage of what I’d call cursor ramming (the cursor stops automatically at some kind of boundary such as a wall unless you let go and re-press the direction you were aiming). I’d be more worried about remembering things like weapon trades and character movement order, to be honest.

I do admit there are some advantages to winging it, although all of it comes from an entertainment perspective. When you wing it and a strategy goes wrong, it is fun to have to improvise on the fly since you have to formulate an alternative strategy as quickly as possible, which keeps you on your toes since you don’t have the time to devise an optimal solution. For the sake of a marathon though, I think this approach is way too risky so long as an RNG planned run can realistically be pulled off, which it can assuming Molotov’s 1:35:21 youtube run can be reproduced with strong enough consistency.

Now assuming an RNG planned run takes place, there is still an endless amount to discuss about these games, which can easily fill in 1.5-2 hours of pure commenting. And without having to worry about RNG, all focus can be placed on execution. I’m sure most people have no idea how fast-paced this game can be when optimizing cursor movements that hold B (doubles cursor speed) and efficiently switches between characters by pressing L (or X for the gamecube/wii games). That’s where a lot of the entertainment value will come from.

I’m going to put some effort into whipping up an RNG planned Sacred Stones run to show what it would look like compared to Bertin’s playstyle, and you can decide what you prefer.

As someone who's heavily invested in how Fire Emblem gets portrayed in the marathon, I have to side with Molotov as well.

First, it's very important to understand some recent developments in GBA Fire Emblem speedrunning. From TASing the game it has been discovered that the game's RNG can be manipulated through path drawing, and is time independent. Quoting from my own Sacred Stones TAS:

[quote]Luck is manipulated in the run by abusing the game’s path drawing algorithm. When there are multiple ways to reach the same location (eg/ by pressing a diagonal input), the game’s random number generator determines which path to take. In fact, the path drawing algorithm is quite straightforward. For a simple diagonal path, random numbers between 0-49 will force the path to end horizontally, whereas RNs between 50-99 will force the path to end vertically.

It should come as no surprise that the most efficient way to burn RNs is to draw a diagonal path as far as a character can move, then “wobble” the cursor back and forth between 2 spots, forcing the RNG to make as many decisions about the path drawn as possible.[/quote]

It’s also known that the RNG is dictated by a seed of 3 values that uses some complicated formula to generate a 4th value, then the 1st through 4th value are used to generate a 5th, and so on indefinitely. There’s a widely used emulator lua script that actually shows a table of RN values on the right side of the screen that updates in real-time.

The most recent discovery is that saving & resetting the game always causes the same RNG seed to be used- 8, 56, 21. This is extremely useful since you can make every chapter in the game have the same RNG sequence by [u]deliberately[/u] saving & resetting and planning out your movements accordingly.

The implications of this are that you can plan out every last move of your run on emulator, memorize that as best as possible, then copy those same moves on console to achieve the desired result. This is a labor intensive process, but completely worth it compared to “winging it”. In an RNG planned run, if you make a cursor mistake, you only have to re-do said chapter since the RNG always resets to the 8,56,21 state as mentioned earlier. In a winged run, entire strategies can crumble if certain units don’t meet certain stat benchmarks or important hits miss or a player unit gets hit one too many times on a long enemy phase and dies. If, for instance, Seth gets str and spd screwed in a Sacred Stones run, your whole run falls apart since the run relies so heavily on him carrying the offensive load.

Another thing to consider- It’s actually pretty difficult to mess up a cursor movement since the only problematic movement type is a diagonal movement in a wide open map. Most movement is either straight movement or in a confined space, where diagonal movement can be avoided by taking advantage of what I’d call cursor ramming (the cursor stops automatically at some kind of boundary such as a wall unless you let go and re-press the direction you were aiming). I’d be more worried about remembering things like weapon trades and character movement order, to be honest.

I do admit there are some advantages to winging it, although all of it comes from an entertainment perspective. When you wing it and a strategy goes wrong, it is fun to have to improvise on the fly since you have to formulate an alternative strategy as quickly as possible, which keeps you on your toes since you don’t have the time to devise an optimal solution. For the sake of a marathon though, I think this approach is way too risky so long as an RNG planned run can realistically be pulled off, which it can assuming Molotov’s 1:35:21 youtube run can be reproduced with strong enough consistency.

Now assuming an RNG planned run takes place, there is still an endless amount to discuss about these games, which can easily fill in 1.5-2 hours of pure commenting. And without having to worry about RNG, all focus can be placed on execution. I’m sure most people have no idea how fast-paced this game can be when optimizing cursor movements that hold B (doubles cursor speed) and efficiently switches between characters by pressing L (or X for the gamecube/wii games). That’s where a lot of the entertainment value will come from.

I’m going to put some effort into whipping up an RNG planned Sacred Stones run to show what it would look like compared to Bertin’s playstyle, and you can decide what you prefer.

kirbymastah

2013-10-29 01:27:02 pm

<(^_^)>

Quote from mikwuyma:

Molotov:

Fire Emblem RNG stuff: Okay that's a good explanation, I thought you were going to segmented/tas-levels of rng manipulation, that was my worry because that would be marathon unsafe (on top of looking kind of ridiculous). You do make a good point about knowing the game well for a speedrun.

Donation Incentives: Yeah...it's too bad that the bid war for the games is good, but also cuts into other incentives. I'm guessing there's no similar incentives for the two games (like all recruits or special endings)?

The donation incentive of FE7 vs FE8 will almost definitely raise a buttload of money Tongue

mikwuyma

2013-10-29 07:30:32 pm

My feelings on The Demon Rush

Yeah Kirbymastah I'm sure the FE bid war will do well on its own, but you can't ignore the other incentives.

Joka

Edit history:

Joka: 2013-10-30 01:19:06 pm

2013-10-30 01:18:50 pm

Personal text

Played some Little Nemo today --> http://sv.twitch.tv/jokaah/c/3165706 (ღ˘⌣˘ღ)
Runs went ok.

I'll be playing it on the Ludendi Halloween marathon, so be sure to watch! Wink

PJ

2013-10-30 11:49:38 pm

Is PJ

I did my first full "speed run" of Arcade's Revenge tonight and came up with a 30:52 or so. Estimate is 40 min, so I should definitely be good to go. Run was really sloppy, as expected, and had one death because Cyclops decided to see what was at the bottom of a pit. Should be np.

I'll be getting back into ActRaiser 2 after I move back to NY but tbh that run is no problem.

quinntendopower

2013-10-31 02:30:25 pm

Takahashi Meijin

I look forward to hearing the spider sense noise PJ

Raelcun

2013-10-31 06:21:02 pm

We require more minerals

Not going to post every time I set a record because at this point every time I get a practice run with a better time it will be a new record.

Nobody is really doing true single segment besides myself in Heart of the Swarm at the moment, though last night's stream with an actual not horrible time seems to have stirred the interest in a few people.

Despite wasting close to 30 minutes and another 30 minutes being lost to loading by my horrid computer I still came in under 4 hours at 3:51. So things are looking good with a lot more room for improvement.

dntn31

2013-10-31 06:57:36 pm

Did 3 full runs today of MK64 150cc GP with skips on stream: http://www.twitch.tv/dntn31/b/475351756 (34:03 starts right away)
run 1: 34:03
run 2: 34:09
run 3: 34:38

PB is 32:11

Pretty consistent considering that I haven't really been practicing much lately.

These runs were done for the first time while also practicing commentary - which does take some getting used to. I'm going to sit down and construct a rough outline of things to say and try to come up with a plan of when to say them and work on getting the commentary as concise as possible while not sacrificing being informative/interesting. Will also try to work out contingency plans for when certain skips go wrong and have a backup fail-safe to minimize cringe.

Barbarossa89

2013-11-02 12:48:29 am

@Molotov I am quite a fan of the Fire Emblem series, and suggest that you do as much manipulation as you can remember. Just make sure to make frequent comments, or if you have a commentator, make him aware ahead of time of some of the more impressive tricks. Knowing that leaving Florina in extreme danger will get the RNG to allow a critical on the boss would definitely be something viewers would want to know. Those with an intimate knowledge of the RNG in the GBA FE games might be able to figure it out after the fact, but letting them know ahead of time wouldn't hurt.

I am very hyped that a Fire Emblem game has finally been accepted at a speedrun.