@Tom94

Tom Long Dong

Ask me anything Follow
Ask @Tom94

Sort by:

LatestTop

Previous

Are there any differences between you in real life and your internet persona?

Not really. When chatting / talking I am very similar to how I'd talk in real life and in public places like ask.fm, the forums or when holding a presentation irl I am very formal.
1
Liked by: DroppedBass

Kann es sein, dass die Plays von einer map nicht mehr richig zusammengezählt werden? Kein einziges Mapset hat über 10k Plays im qualified Status, was mir komisch vorkommt.

Qualified maps werden nicht so offen angezeigt wie ranked maps. Mir kommt der playcount auf den qualified maps richtig vor.
Werde aber nochmal schauen.
1
Liked by: 앤디 - 안드레아스 (Take 1 Name)

Hello Tom, pp's algorythm is actually in place and it works rather well (in my opinion). Do you have any fresh idea in mind in order to improve thie system (even if you do not plan to apply it) ?

I still want to get stochatic-based accuracy difficulty into pp and I have ideas about mania, taiko and CtB. I feel like these need more thinking before implementing them, though. :/

Related users

Bennet_01’s Profile Photo Bennet_01
NFNShowtime’s Profile Photo Noel
RafacarWasTaken’s Profile Photo Rafacar
marcostudios’s Profile Photo desso me la dai visto che mi
Kynan383’s Profile Photo Kynan
P3NGU1NZx3’s Profile Photo Penguin UwU
Brandon Kim
melkisiregar’s Profile Photo Grim Rapper
PastaCats’s Profile Photo PastaCats
NiickAkira’s Profile Photo Akira
Chupachupachupacabrus’s Profile Photo Chupachupachupacabrus
Artaxerx’s Profile Photo Gigo
Njaaaw’s Profile Photo Praw
DroppedBass’s Profile Photo DroppedBass
Shirotsu74’s Profile Photo Shirotsu
Micchaaan’s Profile Photo -Michie
Cyllinus’s Profile Photo what

But as you said at the first answer, that one song is too hard so that every other parts of the maps have almost no effect on star difficulty. So except that part, pp given by other parts is only determined by the amount of hit objects. Isn't this what you said at first, or I just got it wrong?

I said it would be the case on that particular map. That one map has a decent length and the rest of the marathon is fairly trivial in a comparison. If you would take a player that can FC if not SS that one map, then said player really wouldn't have any problems playing the rest of the map apart from maybe silly random mistakes due to the sheer length.
Part of the length bonus is indirectly inferred from the amount of hard parts of the map and another part from amount of hitobjects to account for things like random misses.
1
Liked by: DroppedBass

What benefits from having 2 different difficulty ratings? Why is it not better to have the pp worth and star rating directly proportional? Would remove frustration some people get from the lack of clarity. Also, do you have any insight on when this will happen http://ask.fm/Tom94/answer/116172763342

pp worth already is based on star difficulty. It's just some calculation figuring out how much a given score is worth on a given difficulty.
I depends on star difficulty, OD and to a small degree amount of hitcircles / hitobjects regarding map-specific properties. This is mostly used for accuracy calculation.
1
Liked by: DroppedBass

If a player has 2 scores worth exactly 100pp as their 2 best performance scores (exactly the same even in the decimal digits), would they those 2 scores give 100% weighted each, would one of these be randomly worth 95%, would each be worth 95%, or each would be worth 97.5%?

One would wandomly be worth 95% and the other 100%. Same effect as both 97.5%.
1
Liked by: DroppedBass

Then shouldn't it be changed to number of hard parts(weighted each) rather than just number of hit objects? For example, wouldn't it give like 1000pp when there's Remote Control DT difficulty part at the start and put 10000 of very easy circles after that?

Umm, no. That's not how strain values work. You can try that and compare the star difficulties. You'd only get a similar difficulty to remote control if you are putting an adequate amount of the hard stuff into there.
Length bonus already depends on the amount of hard to hit patterns at a star-difficulty level.
1
Liked by: DroppedBass

If someone removed all other parts except In The Middle Of The Night from the Marathon diff of The Unforgiving and put the same number of objects as easy as possible and it got ranked, does that still give the almost-same pp?

If there would be the same amount of hitcircles and hitobjects, then the pp would be very similar indeed.
1
Liked by: DroppedBass

Hey Tom, ich möchte dir nicht zu nahe treten aber mich beschäftigt da was. Ich habe mich vor einige Zeit bannen lassen durch einen Hack da ich keine Lust mehr auf das Spiel hatte. Gibt es eine möglichkeit diesen Bann irgendwie rückgängig zu machen oder so etwas in der Art. Hoffe ja. LG der Fremde xD

Schicke eine e-mail (in Englisch!) and accounts@ppy.sh und schildere darin deine Situation. Entschuldige dich und sag die Wahrheit, dann hast du womöglich eine Chance. Wenn du etwas Falsches sagst oder Sachen verschweigst kannst du es allerdings vergessen. :p
2
Liked by: DroppedBass 앤디 - 안드레아스 (Take 1 Name)

Some people are getting banned for multi accounting, is it safe for my brother to make an account? He has been playing offline for a while, because we thought that you can't have 2 accounts on the same computer, until I saw this: http://puu.sh/aAlPu/c810a03bec.png

If he is the one making the account and only he will be using it, then it should be fine. Please ask the support instead of me, though, next time. I'm technically not qualified to answer these kind of questions officially.
2
Liked by: DroppedBass no

A response to your question: http://ask.fm/Tom94/answer/116461549262 http://pastebin.com/dqQGAphs I think I've asked this before, maybe I didn't, maybe you ignored. Asking again just in case I didn't before

Anusmen’s Profile PhotoI am no longer Angusman on osu
True, that could work. The thing I am uncomfortable with is, that the training data would have to be human-made and can thus be pretty biased and / or inconsistent. Especially inconsistent if you try to overcome a bias by having multiple people generate the sample data.
1
Liked by: DroppedBass

FC should be rated even higher and acc less... The miss someone gets at the end of a song is gonna put you ahead of them in score, even if you have twice as many 100s

Really depends on where in the song you made the 100s. Often such an accuracy difference can make up for a miss a couple of notes before the end.
1
Liked by: DroppedBass

Why does Combo give so much? I just S'd a 5.04 song with 97.90% (isn't "great" but also isn't "bad") with a combo brake in the middle, but all I got for pp was 0.05.. But If I FC'd with a bit less acc I would of ended up with about 10pp from it. I really can't understand why it's like this.

Please understand, that the amount of pp you actually gain does NOT linearly scale with the score's pp. Your non-FC gave more than half of the pp your FC gave.
1
Liked by: Ahmed Badawy

Here's a small concept of combo-unreliant aim pp calculation: http://puu.sh/axBtZ/1f0e77d739.png. It seems like a pretty good algorithm, since ignoring a few jumps won't influence the map difficulty much, while doing the same with a dozen of them will. If it won't work, I would like to hear why. Ty!

Lahvuun
Unfortunately this is a worst-case analysis. It's something I considered and actually tried with tp, but didn't end up going so well since suddenly things like missing the last not of a map or having 100s at random places greatly reduce your pp. This is basically how per-hitobject data would be used, only that instead of weighting the corresponding pair of hitobjects (the appropriate strain values!) it takes a pessimistic approach and gives the hardest parts of the map the worst possible ratings.
1
Liked by: DroppedBass

I think BPM changes can make a map harder to play, for example https://osu.ppy.sh/s/99244 The BPM changes are so violent, it is harder to follow the rhythm. I think the effects of the bpm changes should be considered with pattern complexity if it is ever implemented.

They would in all likelihood indirectly be incorporated. But honestly, even that maps doesn't have too harsh BPM changes. I could understand someone complaining about Roze - but even there I think the BPM flows really well with intuition.
1
Liked by: DroppedBass

Would you mind answering what your favorite burger is then? :D

For sure the homemade ones my mom makes. Despite having nothing special on them.
1
Liked by: DroppedBass

Hi Tom. I would like to know is there any scaling pp of bpm changes? I mean, if song include 120 and 270 bpm its harder to play than song with 120 or 270 bpm only.

Can't really agree with your statement. I don't find BPM changes hard in any way.
6
Liked by: Rivastyx Stephen Brown Kynan Mindfang

http://pastebin.com/1mMmm9zE A long question regarding the calculation of accuracy rating on maps.

Anusmen’s Profile PhotoI am no longer Angusman on osu
Hmm... I can't think of any good way of incorporating machine learning into the current algorithm without changing it completely.
Am I correctly assuming, that you suggest letting a machine learn to play the song in X steps and see how good it got? Alternatively see how many steps are needed to reach a score equally good or better than the achieved one. This would be a lot more computationally challenging since it'd have to be done for each score being evaluated or alternatively for many possible discrete scores beforehand which would result in huge data amounts.
1
Liked by: DroppedBass

you keep talking about how per hitoject data would make everything so much better in pp. are there actually plans to make that available?

There are steps taken into the direction. Don't forget the downsides, though. They include a lot bigger performance implications and the fact, that it is impossible to obtain such data for scores made up to this point which do not have a replay stored.
1
Liked by: DroppedBass

It'd be cool if it was possible to decide not to submit a score. This would be especially helpful in multiplayer, because I'm playing competitively and I obviously can't restart , so I sometimes miss and I don't want that score to be submited. Could something like this be implemented?

This is intentionally not allowed so statistics can not easily be falsified.
1
Liked by: DroppedBass

If there were ranks for every mod combination you would be able to do score to all of them. In example I could do a nomod score and hdhr score or whatever combination. And of your all scores on that specific map the score which gives the most pp will be counted to your profile.

Exactly.
1
Liked by: DroppedBass

Suppose there's a long map that, for the very most part, is easy. There is one part at the end that is extremely difficult, not humanely possible. You use nf and fc the song all of the song except the very hard part. Could the very hard part up the difficulty lot and make the play worth a lot of pp?

Yes. Sadly this can only be circumvented with per hitobject data. Currently this is avoided by deducting pp based on misses quite heavily.
1
Liked by: DroppedBass

Next

Language: English

4.1

1,266,418 reviews
© Ask.fm 2024