In Murasa, such a bug has been discovered that can significantly cut down on fighting with Kanako.

I have started this forum to discuss whether or not this bug should be allowed in speedruns.

This bug can be targeted with a certain degree of probability. Therefore, I personally think that it can be considered as a part of the technique and allowed to be used, but if anyone has any objections, I would appreciate it if you would write them here.

I believe it should be allowed. The bug can be replicated, and thus can be routed. Similar to Kogasa's getting stuck bug. Second of all, it might happen accidentally, if we ban it and it happens to someone accidentally then we introduce an unnecessary reset point. It does not skip enough of the game to be considered game-breaking. This is similar to Jo'on&Shion vs Kutaka in 1.02. Meaning that if we decide to ban the Murasa vs Kanako bug, there is grounds to ban the Joon&Shion vs Kutaka bug (I know that in 1.02 the skip bug is unavoidable, but it is similar enough to warrant a comparison, and we know it is a bug, the short Kutaka stage was not intended).

I believe it should belong in a Skip category, as it bypasses having to deplete health bars. 6 IGT seconds is a lot by Murasa's standards, and kuroi_tansan has stated that it is difficult to do consistently. Perhaps this is an appropriate way to handle Kutaka's v1.02 thing as well.

The boat skip should also be possible to do on Murasa as the enemy, which would save a significant amount of time as she spends about 18 seconds under water unable to be hit by attacks. There does seem to be some deliberate choices with her positioning and physics so it's less likely to happen though.

While Reimu's guard break glitch does provide an advantage it is closer to being a timing-based technique that is fairly reproducible.

Since no new opinions seem to be forthcoming, I will offer my personal view.

First, I do not believe that it is desirable to add too many categories. Too many categories can lead to a situation where the speedrun for that game is not as busy as the actual number of runners, or the number of entries itself is actually lower than the actual number of runners. I will not name specific games, but as a member of the speedrun community in Japan, it is not uncommon to hear such complaints in other games.

Also, the bug we are discussing is limited to a single stage, not something that can drastically shorten the entire game, such as BLJ in Super Mario 64. In this game, Any% of Flanders falls under the category of a major bug like BLJ, and while the 6 second reduction is a major component of the overall run, I don't think it can be equated with the Flanders bug. If we were to create a new category for skipping due to this bug, we would have to create a new category each time such a phenomenon is discovered in this game in the future. This policy would not be favorable to the moderators nor to the runners who would have to submit a new run each time.

And even though the technique itself is very difficult, I have received responses that it is possible to increase the probability of it happening somewhat, depending on how you move. If it can be managed with some ingenuity, I think it should be considered as a technique. Also, as per the previous comments, as long as the phenomenon may occur by chance, it is not desirable to be forced to change the category of application when it occurs without aiming for it.

For these reasons, I personally think that it would be better to allow the use of this bug itself, but not to establish a new category. However, please keep in mind that this is not a decision, but only my personal view.

Boat skip is a clearly an unintended bug, which may be fixed in the future. This game has limited controls so it's unlikely that many new skips will be discovered, even then they can be lumped together.

It looks like there is an RNG aspect, especially if you want to do it on the first spell 船上水神嵐, and having to rely on it because the top run did is not fun. Whether or not it was meant to be done the times can still be compared, just accessed on different lists.

Side points:

I agree there should not be too many unnecessary categories, as long as most of them are reasonably used and the board's structure makes them easy to access it's fine.

Retrying through the menu in Flandre's Any% isn't really using a bug, it's an intended mechanic that one only gets so many attempts at training before having to face Yuuma. It is comparable to BLJ in ease of use and time save, but not much else.

I would like to know more about how to reproduce this skip. Is there an estimate of success rate? If it can be strongly influenced by the player, my preference would be to allow it. My foremost concern with disallowing it, or creating a separate category, is the matter of performing it accidentally. "Ah, I went too fast. Now I must restart." ...such a thought seems strange when doing a speedrun. It has been far too long since I played (unfortunately, I need a new PC), but it seems it doesn't save much time. If one category is only ten seconds faster than the other category, one of the categories would surely be ignored.

On the other hand, if the skip is based too much on luck, that could be a problem also. However, luck is part of the game. It can't be avoided. As long as the luck factor is not extreme, I think the skip is fine.

この現象に関しましては、調査してくださった方からフィードバックをいただいております。 成功率は現状では低いようで、研究次第で5割くらいまで成功率が上がるかもしれない、といった感触のようです。

ただ、成功パターンの画像を9つ並べたところ、神奈子の位置にはフレーム単位で見てもほぼ差がなかったため、見てから攻撃を出すのではなく、タイミングをリズムで把握すれば成功率は上がるかもしれない、とのことです。 新たなパターンを閃く可能性もあるので引き続き練習してみる、とのお言葉も頂いており、少なくとも完全に運に左右されるテクニックではなさそうです。

懸念点としては、現状では、狙って発生させられるのは最後から2番目の攻撃である「船上水神嵐」の部分であり、その前の「揺曳弾」で発生させられるかは運任せというところが挙げられます。 当然ながら「揺曳弾」で発生したほうがタイムは早くなるため、その点では非常に運に左右されると言えます。 その一方で、「揺曳弾」においてこのバグが起こった場合に通常カテゴリーで申請できなくなることに対する懸念は、検証者の方も口にしていらっしゃいました。

個人的には他のゲームにもよくあるグリッチのうちの一つという認識ですが、自分一人の見解だけで裁定を下すことへの抵抗感もありまして… 多数決にしても現状では2：1にすぎませんし、この状況で結論を出すのは専横的かもしれないと感じていました。 お忙しいところ恐縮ですが、これらの情報を勘案したうえで、加賀美さんのご意見も頂ければ幸甚に存じます。

I have received feedback from those who have investigated this phenomenon. The success rate seems to be low at present, and the feeling seems to be that the success rate may increase to about 50% depending on the research.

However, when we lined up 9 images of successful patterns, there was almost no difference in Kannako's position even when viewed frame by frame, so the success rate may increase if the rhythm of timing is grasped instead of launching the attack after seeing it. He also said that he would continue to practice the technique as it may spark a new pattern, and it does not seem to be a technique that is completely dependent on luck, at least not yet.

One concern is that, currently, it is only the second-to-last attack, the "船上水神嵐" that can be targeted and generated, and it is up to luck whether it can be generated by the "揺曳弾" that precedes it. Naturally, the time will be faster if it is triggered by a "揺曳弾" so it is very much a matter of luck in this respect. On the other hand, the verifier also expressed concern that if this bug occurs in the "揺曳弾" category, it would not be possible to apply in the normal category.

Personally, I recognize that this is one of the glitches that are common in other games as well, but I am not willing to make a ruling based on my own opinion alone... Even if it is a majority decision, it is currently only 2:1, and I felt that it might be tyrannical to draw a conclusion in this situation. I am sorry to take up your time, but I would be very grateful if you could give me your opinion as well, taking this information into consideration.

On reproducing it, one has to have knock down the enemy at the the end of an attack so they land on the edge of the boat and slide off with the rest of the horizontal movement from the knockback. This makes it so that they will attempt to stand up while they're out of bounds and any other attacks end automatically. It will softlock if done on Kanako's non-spell, possibly since that is followed up with some intermission dialogue.

I am not completely against allowing it. It is simpler to take a game as is and use whatever it provides to your advantage for a speedrun. I have not done much work with Murasa, so I'm not one to talk much about whether this should be a part of character mastery. So far there haven't been any runs that have relied on it for a decisive advantage so it's not really relevant at the moment.

I did notice that after the start of "船上水神嵐" vertical positioning can vary before the attempt to land on the player, but maybe it is influenced by the player's vertical at the time of checking like it is with the Yorigami Sisters' "貧乏神的な原油流出汚染". Perhaps that slight difference doesn't actually matter, especially if you aren't planning to attempt it in that timeframe.

I don't have a lot of experience playing Murasa myself, so there is not much I can say definitively about this matter. I apologize for the fuzzy situation, but I also do not want to make a poor decision.

Although it is a separate issue from this agenda, I think it would be better to split the categories by version, since the Jyonn&Shion bug that uadopen pointed out is a definite occurrence, and since the record cannot be updated in 1.04 in any way.

This is because we don't want a situation where people who bought the Switch version from now on, when they try to do a speedrun, will not be able to get the record unless they use ver 1.02, which can only be reproduced on Steam. I have been a bit busy myself and apologize for the delay in responding to this issue.

Jyoon&shion currently had only ver1.02 records for Normal and 1.04 records for Hard, so I thought it would be better to separate them now to avoid problems, so I implemented categorization by version for this character only.

As for Kanako, uadopen's research has shown that she has strengths unique to ver1.04, and my own research has shown that 1.04 is more likely to have better records, so I think that Jyoon&shion are the only characters where version makes a difference.