Hi!

So over the last few weeks, you've probably heard me (or someone else) go on about how we think Rains are probably too hard in the current meta. Well, that's cause they are. But there's no real way to limit what Rains are capable of without making them "Platter+" basically.

While the normal suggestion would probably be "hey BNs stop ranking hard rains" or "start standing up and making people nerf stuff", I think that leaves too much in the air. When I was thinking of ideas, it was a lot of math and way too much work to do for people checking maps. Then, I stumbled upon this guideline, located in the osu!-specific RC for their Insane difficulties (the same level as our rains):

Ranking Criteria wrote:

Avoid cross screen jumps and spaced streams. These are beatmapping techniques primarily reserved for Expert level difficulties.

I feel like if we follow this philosophy, we can similarly limit Rains without making it a rule or by creating a complex solution to a problem that only plagues a few maps.

My proposal is as follows:

Ranking Criteria wrote:

Avoid using large amounts of hyperdashes consecutively. The utilization of longer chains of hyperdashes (especially higher snapped ones) should be saved for Overdose level difficulties.

I'm open to any different suggestions or wordings but this should be a good platform to jump off of in terms of solving a problem that has been noticeably getting worse.

Edit: Since people were having a hard time seeing what I was trying to address, it's basically reducing cases where there is a really large density of higher snapped hdashes.

Some thoughts on the proposed sentences:

Ascendance wrote:

Avoid cross screen jumps and spaced streams. These are beatmapping techniques primarily reserved for Expert level difficulties.

Cross screen jumps can be justified depending on the used snap (thinking mostly on 1/1s or 2/1s), whether they carry a hyperdash or not. We could probably limit the usage of these kind of jumps still allowing them, but making as exception the edge-dash and trigger-distance-over-?.?x scenarios.

Spaced streams shouldn't be an issue in this kind of difficulty, unless they are all hyperdashed, which is already covered with the "If higher-snapped hyperdashes are used, they must be used singularly" line.

Ascendance wrote:

Avoid using large amounts of hyperdashes consecutively. The utilization of longer chains of hyperdashes (especially higher snapped ones) should be saved for Overdose level difficulties.

That should be covered already with "Basic hyperdashes can't be used more than four times between consecutive fruits. [...]". Unless you're referring to chains of hyperdashes, then isolated 1.0x notes, and then more chains of hyperdashes, and so on... In that case we'd need to specify clear limitations.

The current limit of 4 consecutive hyperdashes was set thinking on having a cap on half a stanza with an average BPM. We can either reduce that number to 3, or force a "rest moment" after such pattern if that high amount of hyperdashes is used (preferable the latter option).

The first line was from the standard RC (it’s just what I based the second one off of)

As for the second point, I probably should change the word consecutive since it has a predetermined meaning in our RC already. Basically I’m trying to limit the amount of hyperdashes in a small-ish period of time to prevent rains from getting even harder.

Whilst I'm one of the main people talking about the issue of rain difficulties, I feel it's not something caused by a number of HDashes in most cases, nor do I think the current proposal is a good wording.

Firstly, as mentioned, consecutive sounds more like it's from multiple fruit in a row rather than (as I feel you mean) between sliders. This is obviously a problem that can be solved with a wording change, however most of my other points carrying on from this can not.

Secondly, "large amounts" is such an ambiguous wording that it's essentially just as helpful as the guideline not existing at all. The difference with the standard wording is specifically that it's indicating to avoid the usage of a specific pattern. A good comparison would be if the standard guideline was something along the lines of "avoid using big jumps", because big is such an ambiguous term here, it wouldn't be useful.

Thirdly, the number of HDashes is usually not what leads to the difficulty of rains, but rather the usage of patterns afterwards and such. As an example of what I saw as a problem in a recent rain that led to high difficulty, it wasn't the number of HDashes there, but rather the fact that there were strong HDashes followed by antiflow dash streams. By most opinions, this would be seen as a lot harder of a pattern than the simple 1/4 slider back-forths.

Fourthly, I feel the "especially higher snapped ones" definitely is not necessary, as whilst a 1/2 slider spam with HDashes between may lead to quite a hyper-dense rain, it's by no means difficult for a rain, nor would it be seen as problematic at all, the part that would be a problem would be the number of HDashes in the section.

This brings me onto my last point, though you could say this is a lead-on from my third point. Limiting simply consecutive HDashes doesn't solve the issue much at all, because as well as these generally being easier than some other patterns that can be used following HDashes which lead to an overall difficulty increase in rains, the fact they're consecutive isn't the problem at all, it's the number of them used in a specific section. As an example, you could maybe have 10 HDashes in 5 seconds, something that would be seen as too challenging for a rain by almost everyone (although I'm against limiting the number specifically for mathematical reasons mentioned in your post), yet it's possible none (or very few) of these HDashes are between consecutive objects.

In conclusion, I feel this kind of guideline addition simply isn't helpful in limiting the difficulty of rains due to the fact it fails to really hit any of the main reasons for the increasing difficulty level of rains.

Archiving this because there’s no real clear direction at the moment and discussion has died. It’s probably nearly impossible to limit rains in a way that isn’t a complete numerical mess so we’ll just leave it as-is and tackle problems when they arise regarding rain difficulties