Here are my notes after watching the run:
- An enormous amount of time is lost due to TASing from power-on, as opposed to SRAM where you can turn map animations off (and have access to hard mode). This is somewhat understandable given TASes using SRAM are uncommon. However, map animations cost an enormous amount of time, easily over an hour, and is one of the main reasons this TAS is slower than the normal mode transfers RTA WR by over 2 hours.
- The run uses outdated strategies compared to modern RTA runs found here. In particular, the RTA WR for normal mode transfers is 3:10:58 (see here). Again, somewhat understandable given you were going by my segmented run on speeddemosarchive. However, speedrun.com is widely known as the universal leaderboard for speedruns nowadays, so I'm surprised you didn't find it. The Radiant Dawn page on speedrun.com also links to the Fire Emblem speedrunning Discord (see here), which I encourage you to use should you make another TAS.
More general notes:
- Trains units that don't need to be trained (ex/ Micaiah, Nolan). Training Edward in particular cost a massive amount of time.
- Over-use of paragon, leading to lots of unnecessary level-ups
- Lots of unnecessary skill re-assignments, especially with adept and vantage
- Unnecessary supports, which cost a bit of time every instance they're used (ex/ 3-11 Haar x Ike)
- Gets items from the base that aren't needed (ex/ restore staff, renewal)
- Uses coins on forges that don't need stat bonuses
- RNG manipulation favors near perfect level-ups (often overkill) over critical hits
Given all these issues, I unfortunately have to vote no.
I was laughing so hard at certain portions during the run that it felt like some kind of videogame based standup comedy :') I mean, an ostrich bulldozing its way through a wall? Lmao.
Easy yes vote.
The RNG cycles whenever the attack cursor for a ranged attack unit highlights an enemy ~on a diagonal~. RNG can be incremented by going back and forth between the enemy select and weapon select screens (slow). If there are two or more enemies in attack range, jumping between a diagonal enemy and another enemy will cycle RNG (fast). The RNG value is the same for each of these cycles regardless of which unit is used or which diagonal enemy is highlighted.
This is amazing information, might even affect RTA runs. Thanks for discovering this. I'll make a separate post with my constructive feedback for the updated WIP.
Are you serious? Voting no based on entertainment on a 2 minute video with fascinating glitches? Talk about low attention spans.
Resounding yes from me.
I like this movie but the game is glitchless. There are absolutely 0 glitches that I recall being used.
Did you watch the run, and see the part where he skips from world 3 to the credits with some kind of glitch? Despite witnessing that, still voted no/meh based on entertainment?
Yes, that links to my post.
"Are you serious? Voting no based on entertainment on a 2 minute video with fascinating glitches? Talk about low attention spans."
Let's deconstruct this:
- "Are you serious?" Displays surprise, a bit of a knee jerk reaction admittedly
- "Voting no based on entertainment on a 2 minute video with fascinating glitches?" Statement of fact. See all prior posts in the topic.
- "Talk about low attention spans" Seems judgmental at first, but also fairly objective. Why else would so many people be voting no/meh to a video that's less than 180 seconds long?
Link to video
I've gone through the first two maps.
I manipulate the RNG by the order which units are moved.
Something that also alters RNG is visiting a house, canceling the scene quickly or slowly will alter RNG.
This is a new game, no bonus transfers from Path of Radiance, and no option to turn off battle animations. I've searched for real time speedruns and all I've found use NG+ or transfers.
Is this run worth continuing?
I watched your WIP, seems fine, except player phase is inconsistently skipped optimally, and not holding start when enemies attack to avoid the spinning yellow triangle pointer.
Hope you keep it up, as I don't currently have a PC capable of TASing this game.
Self reminder: http://tasvideos.org/forum/viewtopic.php?p=360698#360698
WIP of Ephraim route up to chapter 5, seems optimized upon video inspection, willing to proceed to completion with this. Maximum 1 chapter per day to ensure gradual progression, sustainable.
@ruadath: Why did you not upload your input file?
It’s always fascinating to me how people get offended by other’s opinions
Who exactly got offended?
Subjective responses outlined below:
- unfortunately doesn’t change how boring the game is (no)
- entertainment wise it is too short and slow to really make it all too fun (no)
- game wasn't fun to be honest (meh)
And so on. Don't see emotions engaged beyond that.
Are you serious? Voting no based on entertainment on a 2 minute video with fascinating glitches? Talk about low attention spans.
Resounding yes from me.
What song(s) inspire you in some meaningful way?
I'll begin with one of my favorites, a song from the iconic game, Final Fantasy 6 for the Super Nintendo Entertainment System. I played this game as a child, also giving it nostalgic value.
The song is very melodramatic, yet simultaneously so beautifully well composed. The patterns in the song are just so pleasing, I find myself returning to it periodically.
https://www.youtube.com/watch?v=JbXVNKtmWnc
Reading back the last page, I highly advise using SRAM because holding A makes characters move twice as fast. Over the course of a run, this accumulates to quite a few seconds, even minutes saved.
I'm interesting in Fire Emblem and Advance Wars TASing, this time for good. Had to deal with mental health issues, which have since been resolved. Now work as a frontend web developer, even have a separate github account for gaming related repositories.
Heard that Toothache is working on an improvement for this game, interested in collaborating in any capacity.
Glad to see this game finally TASed after all these years. It's too bad the jumping mechanics ruin any kind of fluidity in Luke's movement, this could've been a much better game had they made that simple change.
Just wondering, which game? I made some scripts to fish out information in Desmume for that kind of stuff.
FE: Shadow Dragon and FE: Heroes of Light & Shadow, the latter of which is more interesting.
For FE12, I've found that the RNG appears to change every frame, but is cycling between the same 4 outcomes over and over. Beyond that, I haven't even been able to identify the RNG's address in memory after a couple hours trying.
Well, looks like I'll have a go at Nintendo DS disassembly and see what I can come up with. Anyone have some experience with DS disassembly (esp. which tools to use) or disassembly in general that can provide some general tips & guidelines?
Suppose you find the address(es) that govern the RNG, and know how to advance the RNG one RN at a time, indefinitely. This would effectively allow you to have a sample size of consecutive RNs as large as you want. Further assume that later RNs are generated entirely from previous RNs (say R_n+3 = (R_n+2 + R_n+1 + R_n)/3 or something similar) and not dictated by something more complex such as the character's position on the screen.
Is there a way to reverse engineer a huge data sample to find the actual formula that generates the RNs?
As a specific example, the random number generator for Fire Emblems 6 to 8 is as follows:
function nextrng(r1, r2, r3)
return AND(XOR(SHIFT(r3, 5), SHIFT(r2, -11), SHIFT(r1, -1), SHIFT(r2, 15)),0xFFFF)
end
It took me a while to even understand what the formula meant, but I basically convert the 3 RNs to a binary value, perform the binary operations (shift, xor, and), and convert the number back to decimal format to yield the correct result.
Is there a way a formula like that could be determined from a sample set of say, 100,000 consecutive RNs? Is there some kind of pattern recognition program you could run? Do you have to disassemble the ROM?
(the 5 prologue chapters don't show up in difficulties above normal)
Wow, really? Well, that would make Hard mode faster, though…
H5 is most certainly not faster than normal mode, even without the 2 minute prologue. In fact, it is difficult to the point of being blatantly unfair, at least in the first 3 chapters.
Link to video
Notice how you need to have basically every unit in your army take javelin potshots on the pirates just to reach the boss, and the obscene amount of effort it takes to hit the boss, let alone kill him, and contrast that to Marth just blindly charging to the throne. The prologue also serves an important purpose in that Marth gains 4 levels, which means a potential 4 str, spd & def, which is the difference between him being a chump & being a godly unit.
Granted, I would agree that if me or someone else makes an H5 TAS, it should obsolete the current normal mode run.
I’m not fond of how you say “This mistake probably cost about 5 seconds over the course of the run, nothing to lose sleep over”. I’d think 5 seconds are a substantial amount for a TAS.
I simply meant it was a mistake that was minor enough that I wouldn't be motivated to re-do the run from scratch. I edited the submission text accordingly.
ALAKTORN wrote:
Also, you don’t really explain why you use Normal mode over Hard mode in the submission text… you should use Hard mode unless the only changes are irrelevant; are they?
There are actually 6 difficulties available, normal and hard 1 through hard 5. Normal mode is the only difficulty where Marth can solo the game without much RNG manipulation (he basically needs average-ish strength & speed and above average defence). On the higher difficulties, you would either need a godly Marth (think +def on every single level-up) or would have to warp Catria with a forged wing spear to 1HKO the boss on player phase before warping in Marth for the same-turn seize. The latter would involve a lot more warp usage, which would cause significant changes in how warp staff usage is rationed. At any rate, I wouldn't dare attempt an H5 TAS until cracking the RNG formula.