abraker's Osu! Log

abraker · ^

[NEW ACHIEVEMENT]: Highest star 4k DT S'd

New:
[12/30/15] pic; replay
4.16*; Hypernite Industries - Speedcore 300 [Hard] by nowsmart

Prev:
[06/21/15] pic; replay
3.83*; LiSA - L.Miranic [Hyper] by Verniy_Chan

abraker · ^

[NEW ACHIEVEMENT]: Highest star 4k DT PF'd

New:
[01/01/16] pic; replay
3.46*; Ray - secret arms [Shiki's HD] by Shirou

Prev:
[05/24/15] pic; replay
2.76*; Jeff Williams & Casey Lee Williams - This Will Be the Day (James Landino's Magical Girl Remix) [4K NM+] by Fullerene-

[NEW ACHIEVEMENT]: Highest star 4k DT FC'd

New:
[01/01/16] pic; replay
3.46*; Ray - secret arms [Shiki's HD] by Shirou

Old:
[09/11/15] pic; replay
3.29*; 3R2 - Next Stop.With You (feat. KSP) [HD] by Sky_Demon

Last edited by abraker (2016-01-01 10:21:04)

abraker · ^

[PLY] 01/01/16 - <MANIA> Finally passed Various - Yolomania Vol. 3B [(Marathon) aran - Graces of Heaven [17]] by Wafles

Last edited by abraker (2016-01-02 03:05:06)

abraker · ^

[NEW ACHIEVEMENT]: Highest star 4k DT S'd

New:
[01/03/16] pic; replay
4.36*; xi - Happy End of the World [4K Hard] by Blocko

Prev:
[12/30/15] pic; replay
4.16*; Hypernite Industries - Speedcore 300 [Hard] by nowsmart

abraker · ^

It's been 10 days since my last post. Nothing new recently because there is nothing new. I lost access to my primary computer and am stuck with something between a 9 year old toaster that takes 3 restarts to get working and a handheld toaster for quick memes. So, I haven't been osu!ing for the past week or so. I also cant accept mod requests right now since the toaster I'm on has dick stick that is 2 gigs in length. Yup, it's that bad, making it very bloatware unfriendly. I still left queue open since the rate at which the requests are being made and the rate at which I handle requests should allow me to catch up fairly quickly. I expect to get my greater toaster back in a week or 2.

Meanwhile, I had time to shift focus on some other things like working more on osu!skills and watching anime. I finally got a lead and momentum with the reading skill. It took some reasearch conserning the field of view and time for eye movements, which would make the backbone of the entire thing. I do see a dead end in the distance concerning edge case patterns, but I'm sure I'll figure it out by the time I get there. At the rate I am going, I think the core theory will be done by March (or June if I'm lazy) and statistical/experimental confirmation would be done by May (or October if I'm verrry lazy).

Research is a slow process. I had to go back and forth between 2-3 cases of non sence before I got a grip on where to start. First I started approaching towards the quantity of notes and how random they are, but that is too vague and inaccurate for a measurement. Next I tried to tackle the patterns themselves. Close notes vs notes far apart, certain arangements, but if I was to do that, the entire algorithm would be full of if-else statements trying to cover every possible pattern. Additionally, the major flaw in that approach is that I dont know how much harder one pattern than the other. I took a step back and decided to work on a different part, unrelated to pattern reading. How fast do human eyes move? What would be the modeled formula based on experimental data? What is the max number notes per second can a human read per second? Well it turns out, for big jumps, a human can track no more than 28 notes per second, and would require an ability to read at an average 24 notes/second to, theoretically, hit 28 notes/sec or more 1/5th of the time.

Next, I will figure out how visual acuity affects the percieved and how approach circles allow the player to get a general sense on where the next note is. Following that, I will combine the acuity and eye movement theories together to come up with a general formula for reading the patterns. Then, derive the note/sec for some patterns and see how far off I am. If all goes well, I'll factor in the AR followed by visual type modifiers like hidden mod. That is the general plan, however. I am not sure if all will go 100% smoothly because I can be wrong. In any case, at least there is still progress and the project is not dead.

Well this is a lengthy log considering not much has happened. I guess I'll end it here.

abraker · ^

Going to resume my regular osu! experience starting next week. Look at my mod request forum, I am wondering how I will catch up. Yup, underestimated the amount of requests if I didnt bump the thread regularly. Before I start taking requests, I think I should update the beatmap I'm trying to rank. It needs hitsounds and it almost has been a month.

Some new mania maps came out, so that should put me 100 ranks higher, mabe 200 of I'm lucky. I will be mostly focusing completely and making new achievements for now. I also had a try at stepmania. Since I have a specialized skin I use for mania, I want to figure out how to customize skins in that game. Maybe I'll talk about what in my skin boosts my performance sometime.

I came into a minor roadblock with the reading skill. One of the formulas gave me xe^x and solving for x is not possible using any of existting mathematical concepts. I decided to spend a week on it to try figure a way around, and still am. This is one of the reasons it will take months to complete, I have no idea what I will run into.

When I'm done with osu!skills, I'm thinking to start researching on how to detect whether a human is playing the game or an AI. osu!skills's formulas should provide me the backbone of normalizing skills to "take out" the skill part within replays. This would require a system capable of generating a simulated play of a person as if they had a lower or higher skill. This will allow me to see their nervous system's fingerprint and detect imperfections. The nervous system fingerprint would be what dictates the person's motor function behavior. A person would never have flawless motor output especially when playing harder maps. There is always a struggle to stay close to perfection even though it's impossible. The nervous system finger print will also never have psuodo random output. There would be some consistant imperfection (and improvement over time) such as hesitating to move 2 fingers into a certain position by some mere 3-5 ms. While just hypothsis right now, I think it's possible to pull off. Also, it would be my first actual research paper. I havent even found any research on this nor an idea like this, so I think I'm crazy for even suggesting as such. If it would be true, this would be an interesting discovery.

Whelp, I'm done talking nonsense. This was getting long enough.

Last edited by abraker (2016-01-24 08:09:40)

abraker · ^

Finally started playing osu! regularly again. Guess what? I suck now even more, which nothing new. I still make records in some maps and am still able to grind pp, but my skill in the stamina department has crashed like 1929. I fail maps which I used to pass. I B'd a map I S'd. Weirdly, still I passed a the hardest 4k I am able to do on the front page. I guess my decline in stamina has to do with me using a different keyboard that has less actuation force. I did not use the keyboard I am using now for a good month. It sort acts like a weight for my fingers. I will probably take around a couple months to build up my fingers to what they were used to.

On the modding queue front: Finally started taking mod requests again. Revisited the pending statuses to see if they were completed then processed some new ones. Gave a couple stars away to worthy maps and modded maps which needed work. The M4S option isn't proving to be terrible popular, and neither is M4M. I've put a priority note next to both just in case they miss it (who knows? >.>) Perhaps staring a map is not worth spending hours for a mod request because you get stars for modding either way? Maybe I should increase the star amount to 5 stars per 3 edits? Though experienced mappers would quickly starve me out of stars. Dunno. I'll just leave it as is.

I decided to get a head start on researching skills for mania, so I made a thread asking about how the difficulty between keymodes vary. What bothers me in the conversation is that players think it's impossible to account for so many variables, that to calculate the difficulty is impossible. First, yes, second, no. It is possible to calculate the skill required to achieve a certain score or accuracy on a map. By examining the replay, it is possible to determine the player's skill and, therefore, the difficulty of the map. Players think that difficulty can be absolute. It's not. It's based on the player's perception, their bias. Required skill is absolute. It generally gives me a headache when someone says a map cannot be rated. I understand there were multiple failed attempts, but none had the right approach. None I saw/read even tried to consult research material. They are too short sighted, focusing in a primary think like keys/second. Hopefully the article shoegazer linked is more relevant.

osu!skills: It is fun seeing players hunt down for tags/badges. Most have been with a couple yet to achieved by anyone. I think the current plan is to release stuff to keep users busy while I work on the reading skill, which I mentioned will take a loooong while. The next thing planned *spoilers for the future* is a system for players to search maps by patterns. Kert suggested it to be a recommendation like system like Tillerino. Guess you will have to wait which one it will be or whether it will be >:D

I am wondering whether I should start blogging my findings relating to my research pertaining to osu!skills or not. I have determined estimated results based on formulas I derived. For example, For a CS 7, AR 8.5 map @ 30 FOV, 60 FPS and 1:1 aspect ratio, the furthest significant angle at which the approach circles going inward would be noticeable would be 2.09 degrees from the center of vision. At 2.09 degrees, when using the formula for eye movement latency it would take 25.65 ms to look at each consecutive note. At 2.09 degrees, the notes would be 171 osu!px. This is just one example calculation I did to check if the result makes sense. Reading 1 note/25.67ms at 171 osu!px apart makes sense to me. It might be a rough estimate due to possible errors within methods used to calculate, but it could be used to engineer maps based on wanted constraints.

I think this is enough writing for one day. I might leak something unwanted if I continue... oh wait

Last edited by abraker (2016-01-31 06:23:51)

abraker · ^

Just something mildly interesting

▼Hidden text

abraker · ^

I can finally not get as tired mashing buttons on the keyboard. Still not making many records though. Apperently mania is not the only gamemode a crashed in. There are standard maps which I can't pass anymore. I feel that my coordination is degrading or something. Perhaps it's due to another factor. I any case, I am spending my time fixing all the maps I got C's in, then will do B's for the meantime.

The collages I posted on the off-topic subforum here really got me thinking on making a personal art thread in the osu! forums. I hear that's a legal thing to do now, so why not? I think it will consist of image dumps of stuff I make or put together. Still thinking about it.

Modding queue is moving along a bit slower than intended, but at least there is progress. Had to blacklist someone for the first time for not putting "[REQUEST COMPLETE]" in their modding request. Stupid reason it may be, but it had to happen. Why do I enforce the "[REQUEST COMPLETE]" thing anyway? It's an agreement that the mapper is satisfied with my mod and doesn't have anymore requests, comments, concerns, questions, etc regarding that specific mapset. Not only it's an agreement, but it's a hand shake and a display of satisfaction on the forums to the public as well. So what do I interpret neglecting to put "[REQUEST COMPLETE]" as? Unsatisfaction. The mapper I sent 3 PM's to has not responded, giving me unsatisfaction as well as ignoring me. For that he is blacklisted.

I managed to finish hitsounding the Dense difficulty of my map and update it the day it became graveyarded. YAY! I guess I'll finish hitsounding the other diffs while I wait for something to happen. I seriously either need to post on that new general mania mod queue or do some m4m. I am also not sure whether I should throw a portion of the stars at the map now or later. Yea...

My research in the reading skill is entering the next stage. Until now I have been researching micropatterns, which is how one or two notes affect reading. Its importance gives my a good idea on what the players' response time should be on noticing notes appearing and accuracy of estimation when to tap according the approach circle. The next phase will be a bit harder, dealing with macropatterns. So far, I came up with three subskills: Path Reading, Spacial Reading, and Rhythmic Reading. For each of those I have to figure out the odds of the player succeeding at each of those skills.

While I don't have the formulas for the subskills, I came up with a general formula to calculate the odds of a player succeeding at achieving a certain %, and almost combo as well. The main problem I ran into is that Desmos is not powerful enough to compute a product-sum of a function that has a factorial in it. The sum itself needs to be large to be accurate, but since the factorial is dependent on that large number, evaluating something like 1000! is very unrealistic. I will need to come up with some shortcuts or workarounds to this.

Damn this is getting long already. I still wanted to talk about what kinds of maps I want to make in the future. I guess it will have to wait. yawn~ I need sleep ,`=.=)//zz

abraker · ^

[PLY] 02/12/16 - <MANIA> Finally passed ueotan - Mario Paint (Time Regression Mix for BMS) [Lv.02] by Cache with a score of 659k (91.18%)

Last edited by abraker (2016-02-12 19:03:17)

abraker · ^

[NEW ACHIEVEMENT]: Highest star 7k B'd/Passed

New:
[02/12/16] pic; replay
5.8*; Team Grimoire - C18H27NO3(extend) [7K Axities] by [Shana Lesus]

Prev:
[09/08/15] pic; replay
5.48*; DARK SAMBA MASTER -SATOH- - DARK SAMBA LAND [7K Lv.0] by -Kamikaze-

Last edited by abraker (2016-02-17 13:41:32)

abraker · ^

[PLY] 02/16/16 - <MANIA> Finally passed Shiraishi - Shinsekai [Insane] by Xay with a score of 681k (90.16%)

abraker · ^

I missed last Thursdays regular modding schedule I set up myself due to work that needed to be done and I expect me having less time for osu! is going to become regular. I had a snow day and classes got canceled, so I was able to do some modding on Tuesday. Though it was mostly management. I went over the list of users yet to put "[REQUEST COMPLETE]" on their post or users that never gave kudos. The maps I went through were 5* or more that session. Either I was too lazy or the maps just drained me that time. Either way, I would expect the next wave of mods next week or the week after that. I'd really hate to make it every other week, but if it will come to that, then it will come to that.

I set some new records as seen in the previous posts. Team Grimoire - C18H27NO3 was hard to me before due to the hold note cords the map has, but now I learned to read them a bit better. Shiraishi - Shinsekai was hard for me due to its high BPM streaming and high HP. The player would need to have flow to pass that map, and for once in a rarity I managed to feel no or little strain in my fingers while playing up to 4.5 stars that day. It takes a certain mental state to relax and have fine control over how your fingers tap, and in edge cases, vibrate. Couldn't pull off the 4* diff in Empress though. I came close with 3/4 of the map done before my stamina killed me. That tells me I'm about 3-7 months away from passing it.

There is some progress on the reading skill, but not a whole a lot. With work load, I get little mental stamina remaining to come up with ideas for how to approach things. This is why I wasn't certain if it will be done by June or November; shit like this happens. At least it's foreseen. I mean if I finish in 2017, then that idk what laziness would overwhelm me considering I will have summer break and stuff to get this done.

Anyway, time to shift topic to mapping. My map, "Density of Air", still needs to get hitsounds done in the Hard, Normal and EZ diffs. And more mods are needed as well. I'll probably go on a mod request spamming spree when I het those hitsounds done. I also started work on a new map, "F-777 - Ludicrous Speed" and I am mapping it for the standard gamemode. It's the first map I'll map in standard and so far it's taken shape of a style I really look forward bringing forward for others to play. No, not symbolic mapping, though I use that a bit everywhere. Hopefully 2016 will be the year of hold jumps, and I don't want to spoil anything more.

Following those 2 maps, I will probably map one of my favorite songs, do a SB/vid based map, or visualization map. I did mention I wanted to make a Geometry Dash kind of map using the 1k keymode in mania. I don't know how possible it is, but I'll give it a try at one point. Concerning the visualization map, I think osu! needs more Centipede-like stuff. I see myself going into visualization mapping later on. I hope in the future there will be beatmap visualization contests. The ideas I have on how to visualize some songs would make Centipede look like a 2007 map compared to modern day maps. Ok, maybe to far of an exaggeration, but I do have sweet ideas in mind.

Oh yea I also got a bunch of older things a mapped. I will prob need to take care of those at some point. One is a fully mapped marathon map which needs some polishing. Another is comparable to a newb 3* shitmap. Needs to be redone from scratch. 2k-18k short touhou mapset. That needs extending somehow. The melody repeats, so it's hard to come up with much. And then I have an 8* mistake which In which I will remap impossible part to make it more playable.

Now that I wrote about what I want to map, I have no idea what I will write about next. I guess I will find out soon. Time to end this before I start writing gibberish I write in my private entries on here.

abraker · ^

Something I came across recently. The low HP on this map let me put my stamina to the test, unlike Empress where you do or die. I tried DT'ing recently, but couldn't. Strange how my skill doesn't improve in some aspects. I guess my stamina has not gotten better.

4577726

abraker · ^

This is 2.24*, the same diff as the other map that I got the highest DT 4K SS'd. The skill barrier is real. Also so many maps a got single 200's on. Probabilistic math is just screwing with me now.

4590243

^ #46 2015-12-31 03:00:03

Re: abraker's Osu! Log

^ #47 2016-01-01 10:17:40

Re: abraker's Osu! Log

^ #48 2016-01-02 03:04:05

Re: abraker's Osu! Log

^ #49 2016-01-04 01:57:12

Re: abraker's Osu! Log

^ #50 2016-01-13 20:25:42

Re: abraker's Osu! Log

^ #51 2016-01-24 08:06:56

Re: abraker's Osu! Log

^ #52 2016-01-31 06:17:59

Re: abraker's Osu! Log

^ #53 2016-02-04 03:09:46

Re: abraker's Osu! Log

^ #54 2016-02-06 08:46:28

Re: abraker's Osu! Log

^ #55 2016-02-12 19:02:32

Re: abraker's Osu! Log

^ #56 2016-02-12 19:49:48

Re: abraker's Osu! Log

^ #57 2016-02-16 19:51:00

Re: abraker's Osu! Log

^ #58 2016-02-18 08:16:57

Re: abraker's Osu! Log

^ #59 2016-02-21 08:32:08

Re: abraker's Osu! Log

^ #60 2016-02-23 05:37:16

Re: abraker's Osu! Log

Board footer