I wouldn't really call that "abuse"

I would.

I lost pp due to the update and I still agree with the change because those maps were broken in terms of pp

Sucks for GN and Kuvster but this is probably how it should have been to begin with, it was bound to happen eventually.

I am somewhat surprised that Renai Circulation also got pp taken away from it but I would suppose that's more for completeness (so that people can't claim it's arbitrary) rather than because it's broken

I was wondering - to solve that issue where maps like Talent Shredder give like 160pp for an SS (among others) you could maybe change that "shortest possible distance between 2 notes" thing for sliders that are longer than 1/3 (so, 1/2 or longer), so that those sliders would have distance counted as if the sliderend were a regular hit circle with no slider leniency.

I'm not sure how great of an idea this is but I remember your reasoning for this somewhere was that you didn't want sliderstreams, or other patterns with short sliders at a distance from each other that are pretty damn easy to play, to be overrated. Even though this is a legit argument, slider streams really only happen at 1/4 sliders, so I figured that for any slider that's 1/2 or longer that could never possibly be an issue, and it would ramp up some of the really hard slider maps in terms of pp without overrating the easier sliderstream maps.

It's just a suggestion, but I think it'd be cool if we could focus on solving this issue. It's pretty ridiculous that Talent Shredder gives 160-170pp for a nomod SS while Coloring HDDT gives like 215 pp. It has HDDT SSes all over the leaderboards while Talent Shredder doesn't have a single nomods SS - the highest being 99.92% (1x100) at 159.86pp. If it gave at least 200-210pp for an SS that'd be a lot more fair than the way things are now, so, even if my idea isn't viable, maybe if we focused on finding a reasonable solution with multiple people we might get to something.

Aren't the sliderstreams in big black 1/2s? I suspect it'd get way overrated if this change was impkemented.

Tess wrote:

those sliders would have distance counted as if the sliderend were a regular hit circle with no slider leniency.

point is theres still slider leniency so it would be a bad change and overrate sliders. slider leniency is still there, if someone leaves that slider before the last tick it becomes a 100 not a combo break, and ppv2 still gives him the pp for the play except with a single 100 which it doesnt know where that 100 came from. this cant be implemented until theres more information available.

basically you cant give people credit for hitting jumps that didnt get to.

I'm not even sure how exactly sliders are treated now. Like distance and time between the sliderhead and next note and minimum distance to complete slider is added to the jump? Or min distance to complete slider+distance from that point to next note?

Either way if it's merged into a single jump that is timed between sliderstart and the next note, it'd mean that sometimes the actual jump you have to do is faster, since when you don't have to move much due to leniency, you still have to stay there for half the duration (between sliderstart and next note) before you can move on to the next note, unless the direction of the slider and jump is not too different + distance is kinda even, or they're close enough so that it doesn't matter.

Also being too strict on leniency is like calculating jumps with the shortest possible distance between the very edges of circles, it's not realistic to play like that in practice. You don't wanna risk slider 100s or misses.

Such treatment of leniency is perfect when the path you'd take to play the sliders as singles falls within the sliderfollowcircle anyway, like plenty of big black sliders, but stuff like talent shredder is a different story.

So I heard certain maps don't grant pp anymore, and that's because someone or some people don't approve of a certain playstyle. Why isn't this strange?

Drezi wrote:

I'm not even sure how exactly sliders are treated now. Like distance and time between the sliderhead and next note and minimum distance to complete slider is added to the jump? Or min distance to complete slider+distance from that point to next note?

Either way if it's merged into a single jump that is timed between sliderstart and the next note, it'd mean that sometimes the actual jump you have to do is faster, since when you don't have to move much due to leniency, you still have to stay there for half the duration (between sliderstart and next note) before you can move on to the next note, unless the direction of the slider and jump is not too different + distance is kinda even, or they're close enough so that it doesn't matter.

Also being too strict on leniency is like calculating jumps with the shortest possible distance between the very edges of circles, it's not realistic to play like that in practice. You don't wanna risk slider 100s or misses.

Such treatment of leniency is perfect when the path you'd take to play the sliders as singles falls within the sliderfollowcircle anyway, like plenty of big black sliders, but stuff like talent shredder is a different story.

Tom said that the game calculates the shortest possible route through sliders that the player can take. Obviously that isn't realistic and needs to be changed so sliders are properly weighed.

uberpancake wrote:

Aren't the sliderstreams in big black 1/2s? I suspect it'd get way overrated if this change was impkemented.

http://ask.fm/Tom94/answer/122374575566


Also, I've been reading some things about per-mod highscores. Could anyone give me a summary or link on that? I'm curious.

Also I'm with b1rd here. I highly doubt that my suggestion would overrate longer sliders, but hey, I'm not a programmer or whatever.

those wrote:

So I heard certain maps don't grant pp anymore, and that's because someone or some people don't approve of a certain playstyle. Why isn't this strange?

It's not like we (or I) disagree with that playstyle. Heck, I have scores on tag4 maps with touch myself. The reason it doesn't give pp is that it's inherently impossible to make a difficulty algorithm that works correctly with touch and cursor-based input at the same time.

Jumps are trivial on touch while insanely hard cursor-based. That's a fact. Same with spaced streams being almost impossible using touch due to not being able to move the cursor while hovering. (That being said, theoretically when using some method that allows hovering during touch it should be superior to cursor-based input unconditionally.)

There are only 3 options. One of them being to rate touch plays differently than cursor plays. Which is impossible to do reliably, because we can't reliably detect input devices used. People can even switch them mid-play during pause. The other 2 options are

1. Make large jumps give little pp for both cursor and touch plays, basically making it impossible to gain good pp while playing with a cursor and reducing pp to be speed/acc based.
2. Remove the only maps (in the case of osu!, that's exactly 2!) where there are patterns which become problematic. Those patterns don't exist in the rest of osu! simply due to not being allowed by the ranking rules - fortunate coincidence.

I think it's obvious to choose the 2nd option here since there is no alternative. I don't like it either. I love gimmicky maps, I love playing with touch.
Now lastly, why did all tag maps lose their pp rather than only the 2 problematic ones? The answer is being consistent. tag maps were never meant to be played by a single player and would have never passed the ranking rules - even back then - when being considered as single-player maps. This simply was a good opportunity to remove them from ranked play altogether. Scoreboards have been kept to let people keep the amazing scores they put tons of effort into and to keep the niche competition that obviously many enjoy.

Something like a tag-ladder would be an amazing thing to happen. There are in fact so many things that would be amazing to happen and which would be allowed to happen if there was the manpower to implement them properly.

Question: If you took a map and removed all the slider ticks from every slider so the combo is lower, would it award less PP?

No, because the potential PP a map can award is not based off of max combo, but on object count.

Well, what if there were separate boards for touch and non-touch players? That way, you could respect the scores that mouse/tablet players earn while keeping track of the touch scores (and the touch scores would not give pp, or give pp on a different scale.)

Also, wouldn't touch players jump directly to the spot at the same time they hit, whereas mouse and tablet players would have a gradual cursor movement? There should be a way to tell the two apart. (Of course, there is dragging, but in that case, the touchscreen has the same limitations as a mouse or tablet.)

Even if it's too hard to program a detection method inside the game, there is always the option of osu scanning for touchscreen inputs. Or, people could just voluntarily say that they are using touchscreens instead of mice or tablets.

gameon123321 wrote:

Well, what if there were separate boards for touch and non-touch players? That way, you could respect the scores that mouse/tablet players earn while keeping track of the touch scores (and the touch scores would not give pp, or give pp on a different scale.)

Also, wouldn't touch players jump directly to the spot at the same time they hit, whereas mouse and tablet players would have a gradual cursor movement? There should be a way to tell the two apart. (Of course, there is dragging, but in that case, the touchscreen has the same limitations as a mouse or tablet.)

Even if it's too hard to program a detection method inside the game, there is always the option of osu scanning for touchscreen inputs. Or, people could just voluntarily say that they are using touchscreens instead of mice or tablets.

I still don't think it can be done reliably, Tom has said so anyways. Also, what can you do about people swapping devices part-way through a play?

Rewben2 wrote:

gameon123321 wrote:

Well, what if there were separate boards for touch and non-touch players? That way, you could respect the scores that mouse/tablet players earn while keeping track of the touch scores (and the touch scores would not give pp, or give pp on a different scale.)

Also, wouldn't touch players jump directly to the spot at the same time they hit, whereas mouse and tablet players would have a gradual cursor movement? There should be a way to tell the two apart. (Of course, there is dragging, but in that case, the touchscreen has the same limitations as a mouse or tablet.)

Even if it's too hard to program a detection method inside the game, there is always the option of osu scanning for touchscreen inputs. Or, people could just voluntarily say that they are using touchscreens instead of mice or tablets.

I still don't think it can be done reliably, Tom has said so anyways. Also, what can you do about people swapping devices part-way through a play?

Why not seperate mouse and tablet rankings too?

ZenithPhantasm wrote:

Why not seperate mouse and tablet rankings too?

Yep

marcostudios wrote:

ZenithPhantasm wrote:

Why not seperate mouse and tablet rankings too?

Yep

Why not separate mouse only/tap only/tap x/click x/ touchpad/ trackpoint/ps3 controller/this too

You can't (not yet, anyway). This applies to mouse and tablet as well.

I was thinking about the problem mentioned on the previous page of sliders being underweighted and I had an idea: What if the difficulty algorithm were to minimize strain across 3 notes at a time instead of 2? Let's call the objects A, B, and C. The idea would be to find the lowest strain way of going from A to B to C such that B and C are each hit within their respective hit windows and assume that the player plays hit object B in that way. This would give you a new starting time and location and you would then consider objects B, C, and D in order to find when and where C should be hit. The main benefit to this approach is that it allows slider starts, ends, and ticks to be treated as individual hit objects without causing problems, which should give more accurate difficulty judgments in maps that use sliders in tricky ways. I think it would also help to make short fast streams more valuable compared to longer slower streams, since higher bpm streams would gain some value from being harder to transition out of.

The biggest downside I would see to it is that the relative value of patterns like this would go down when they're already undervalued as is:

Making that worth an appropriate amount would probably at a minimum require increasing the importance of time over distance for angular cursor movements and I don't feel like thinking about that right now.

gameon123321 wrote:

Well, what if there were separate boards for touch and non-touch players? That way, you could respect the scores that mouse/tablet players earn while keeping track of the touch scores (and the touch scores would not give pp, or give pp on a different scale.)

Also, wouldn't touch players jump directly to the spot at the same time they hit, whereas mouse and tablet players would have a gradual cursor movement? There should be a way to tell the two apart. (Of course, there is dragging, but in that case, the touchscreen has the same limitations as a mouse or tablet.)

Even if it's too hard to program a detection method inside the game, there is always the option of osu scanning for touchscreen inputs. Or, people could just voluntarily say that they are using touchscreens instead of mice or tablets.

How would you do that?...

Gumpyyy wrote:

gameon123321 wrote:

Well, what if there were separate boards for touch and non-touch players? That way, you could respect the scores that mouse/tablet players earn while keeping track of the touch scores (and the touch scores would not give pp, or give pp on a different scale.)

Also, wouldn't touch players jump directly to the spot at the same time they hit, whereas mouse and tablet players would have a gradual cursor movement? There should be a way to tell the two apart. (Of course, there is dragging, but in that case, the touchscreen has the same limitations as a mouse or tablet.)

Even if it's too hard to program a detection method inside the game, there is always the option of osu scanning for touchscreen inputs. Or, people could just voluntarily say that they are using touchscreens instead of mice or tablets.

How would you do that?...

Rewben2 wrote:

gameon123321 wrote:

Well, what if there were separate boards for touch and non-touch players? That way, you could respect the scores that mouse/tablet players earn while keeping track of the touch scores (and the touch scores would not give pp, or give pp on a different scale.)

Also, wouldn't touch players jump directly to the spot at the same time they hit, whereas mouse and tablet players would have a gradual cursor movement? There should be a way to tell the two apart. (Of course, there is dragging, but in that case, the touchscreen has the same limitations as a mouse or tablet.)

Even if it's too hard to program a detection method inside the game, there is always the option of osu scanning for touchscreen inputs. Or, people could just voluntarily say that they are using touchscreens instead of mice or tablets.

I still don't think it can be done reliably, Tom has said so anyways. Also, what can you do about people swapping devices part-way through a play?

Well, a touchscreen can't register a finger unless it's touching the screen. Therefore, when you tap with different fingers to aim, the cursor would skip directly from one position to another, instead of gradually moving there. That's what I meant.

Plus, when an input device is plugged in, it usually has a name. What if osu checked for keywords, or complied a list of touch devices? Then, when a device is connected, osu can see if it's a touchscreen.

gameon123321 wrote:

Well, a touchscreen can't register a finger unless it's touching the screen. Therefore, when you tap with different fingers to aim, the cursor would skip directly from one position to another, instead of gradually moving there. That's what I meant.

Plus, when an input device is plugged in, it usually has a name. What if osu checked for keywords, or complied a list of touch devices? Then, when a device is connected, osu can see if it's a touchscreen.

Guess how rapid motions at low framerate look like? Right, like teleportation, just like touch screens. Not reliable.

I think I don't even need to begin explaining why the keyword method you suggest is also not reliable.

Anything not reliable can not be used for something like scores that require absolute reliability. A leaderboard only makes sense if it's integrity is out of the question. That being said, even if we could detect input device types with 100% certainty - how should we tackle scores that switch between input devices mid-play?

Tom94 wrote:

That being said, even if we could detect input device types with 100% certainty - how should we tackle scores that switch between input devices mid-play?

IF it was possible to detect device used on certain parts with 100% certainty - you could simply disallow multi-device scores for submission. Normally people play through a single map without changing from touchscreen to tablet/mouse midway, so enforcing that doesn't sound unfair for simply playing as intended with whatever playstyle.

Drezi wrote:

IF it was possible to detect device used on certain parts with 100% certainty - you could simply disallow multi-device scores for submission. Normally people play through a single map without changing from touchscreen to tablet/mouse midway, so enforcing that doesn't sound unfair for simply playing as intended with whatever playstyle.

How would it know if you both a tablet and a mouse connected to the PC at the same time?
Your suggestion would make me have 0 scores since I have both connected at the same time.

This is only about switching between touchscreen and non-touchscreen (cursor input) midgame. And while that's easy for us to see by looking at a replay, as I understand it's hard for a computer to detect with 100% certainty, so it's purely hypothetical anyway.

As said before something which isn't 100% certainty is like it's 0% certainty. As a programmer I can assure that something client-side (on the side of the player) is not 100% verifiable since only data are sent at the end and they can be altered anyway.

x

AR11 buff please

edit: wow i was way off, tom's ask.fm proved that.

Riince wrote:

that speed buff <3 exactly +50 pp, i wonder what my scores will look like once they're re-weighted...

Yay for 3000pp.

R.I.P hvick's 9000pp O_O

XgenSlayer wrote:

R.I.P hvick's 9000pp O_O

Seems like TV Size doesn't pay off anymore.

I can't wait for rrtyui playing a map.

[-Cloud-] wrote:

XgenSlayer wrote:

R.I.P hvick's 9000pp O_O

Seems like TV Size doesn't pay off anymore.

I can't wait for rrtyui playing a map.

Doesn't make any sense since nothing in the changelogs was said about nerfing tv sizes. if anything they wanted to nerf longer maps. they just nerfed ar10.3 since the scaling starts at 10.3 now which started 10 before. this means they nerfed all dt scores on ar 9 maps.

Guess what hvick farmed the most. Yes, the answer is TV Size AR9 maps. Good job, billy.

[-Cloud-] wrote:

Guess what hvick farmed the most. Yes, the answer is TV Size AR9 maps. Good job, billy.

How does tv size have to do anything with him losing rank? Explain please. TV Size doesn't automatically equal AR9. On the other hand let me add they buffed smaller cs which makes HR scores on CS5 maps more interesting. Again, does that mean Tv sizes are worth less in general? No. In fact, CS5 HR plays are weighted stronger.

The changes also resulted in a smaller pp reward for www's Sana DT play. Now he lost 15pp of 506. Can you really say it didn't pay off? 3% less pp not worth anymore? Well...

Infevo wrote:

[-Cloud-] wrote:

Guess what hvick farmed the most. Yes, the answer is TV Size AR9 maps. Good job, billy.

How does tv size have to do anything with him losing rank? Explain please. TV Size doesn't automatically equal AR9. On the other hand let me add they buffed smaller cs which makes HR scores on CS5 maps more interesting. Again, does that mean Tv sizes are worth less in general? No. In fact, CS5 HR plays are weighted stronger.

AR9 TV Sizes + DT are indeed worth slightly less in general, at least that's what it looks like, 513 -> 489pp on hoshizora for example, despite ALL maps giving more pp overall.

On his grind to #1, hvick played mostly TV Size DT (Not only, but a lot) and now he's basically completely nerfed.
In case you didn't get it yet. hvick spammed a shitton 10.3, that's it. DH will drop soon aynways.

[-Cloud-] wrote:

On his grind to #1, hvick played mostly TV Size DT (Not only, but a lot) and now he's basically completely nerfed.
In case you didn't get it yet. hvick spammed a shitton 10.3, that's it. DH will drop soon aynways.

Totally aware of his plays. It is just your generalization is not right as you made it. Some of the best plays are still short versions/short maps/tv sizes with AR9 and dt. They didn't get any less impressive than before the update. It is just they balanced dt out with hr and I explained in how far.

edit: sayo is also probably gonna drop as well once he is gonna get calced. also notice how many ranks azer went up considering how much hr he plays.

You asked about his drop in the rankings, not about how impressive his plays are. They are indeed impressive, but the played way too much of short AR9, which got nerfed right now. So it's no wonder he lost that much PP.

[-Cloud-] wrote:

You asked about his drop in the rankings, not about how impressive his plays are. They are indeed impressive, but the played way too much of short AR9, which got nerfed right now. So it's no wonder he lost that much PP.

No, I didnt ask anything about him dropping in particular. I wanted you to give reasons for why short sized maps would not "pay off" anymore which is plainly wrong considering my reasoning.