Rarum

Not a whole lot to say. I managed to not have to use AME during the bits where I really wanted to fight in seasons, and that's pretty much all I can ask for. The 120 point difference to the second-best score is a bit on the high side, but I was a bit unsure of tmelloi's average so I played it somewhat safe. Statistics My score: 7303 (100.217% of what my algorithm would expect with completely average luck) Expected score (my algorithm): 7287.2 Expected score (HH++ algorithm): 7265.1 If I had been using the following booster combinations for the duration of the league but with the same fight timing decisions, these would have been my completely average luck scores: AME+3cor 7339.9 AME+2cor,1gin 7319.8 AME+1cor,2gin 7291.7 AME+2gin,1chlo 7219.4 3 cordy 7178.2 2 cordy, 1 gin 7145.2 Perfect precognition score with average luck and AME+3cor throughout: 7396.44 (24.411 average). The AME+3cor expectation was 99.236% of this.

I do believe that's the level 500 line. The bubbles just are that big that you can't really pinpoint to the precision of a single level or hardly even five. But it starts not looking as good if I enlarge the chart even more (that seems to be the way to decrease the relative bubble size).

Well, it looks like I had to go back and fix something from two past league brackets' posts. Namely, the "perfect precognition scores"; they had one extra "illusory opponent". There's always something, isn't there? At least it was easily fixable this time. The very first league where I had it introduced was already correct, though.

News about known potentially notable league happenings: On 14th of April, La Kajira won against Rarum, at a very small margin (24.31 vs. 24.28) but still. Of some note was zoopokemon getting a decent score of 23.87. On 14th of April, sygfried01 broke the record average, with 24.41. It is to be noted that this happened in quite the weak bracket; the next best average in it was 22.47, for some amount of context. On 14th of April, Lgm123 won against Shado with 24.29 vs. 23.82. On 21st of April, Hugh won against Shado with 24.06 vs. 23.44, with zoopokemon as third with 23.3. On 21st of April, Rarum won against sliqguy with 24.21 vs. 23.94. On 21st of April, bla won against RedEagle with 24.03 vs. 23.96. On 28th of April, phyzel won against zoopokemon with 23.94 vs. 23.62. On 28th of April, karistea won against RedEagle with 24.22 vs. 23.79. On 28th of April, bla (24.06) won against sygfried01 (23.95), Thulsa Doom (23.94) and Hugh (23.88). I'd mostly put La Kajira back above me like I promised due to my loss. Otherwise I'm not really seeing weighty reasons to change the rankings yet. I could consider moving sygfried01 above karistea since sygfried's losses seemed to be in really tough brackets but eh, I can wait for more evidence. I'll try and consider if there's sufficient reason to reopen more spots in the ranking next time, too.

Well, no jumbled cables since this is now as one single set of data instead of multiple brackets independently, but your comment did make me add a trend line. I ended up using the Power Series method for that, since it produced the lowest chi-squared number (0.182). Maybe we'll get a statistician telling us what it might mean or if it has even any meaning at all in this context, but it looked like I can test goodness-of-fit by observing which method gave the lowest chi-squared.

I guess I could do that. I also hid the trend lines, even though I think it obscures data in the way that this graph doesn't really show the frequency with which players of different levels reach certain ranks. It just now shows what different leveled people "can" under some circumstances achieve. Maybe the chart design could be improved somehow... Which I do, later on in the message. The amount of data points for so low level people is maybe a bit too low to draw many conclusions, but I think we can say that it's not "always", even beyond the people who are trying to demote. Now, I said I already did think that the frequency of a rank could be shown better somehow. I found this bubble chart type and organized my data to be able to utilize it properly. I even added a few more brackets to it (I only had those "a few more" for level data). It could be worth it to note that I do stop at rank 102, but it's not like we are terribly interested in who is trying to demote, are we? I put bubble opacity as low as 3%.

Hard at Work: Test server: Stage 1: 35 shards Stage 2: equipment Stage 3: equipment Stage 4: equipment Stage 1: 35 shards Stage 2: girl Main account: Stage 2: equipment Stage 3: 35 shards Stage 4: 35 shards Stage 1: equipment Stage 2: girl

What @EpicBaconsaid about giving choice about whom to talk to, was basically said on Discord, but not on the feedback channel. I suggested doing it there but to no avail. After that, I've not seen mentions of this idea, so I think it's an important one to have brought up, to see if there's demand. Yeah, I think now it feels far too linear, like you have no control at all, and giving choice about whom to talk to, would be good. It's already been mentioned, but what you choose for your reply doesn't matter, either. It gives a bit of a poor impression of how much they value this feature even themselves, having spent the bare minimum effort on it. The feature was even released with a list of known issues! What kind of an impression does Kinkoid think that's going to leave about how they value their own game? There are people who still like the game and soldier on with the efforts of trying to prevent it going down the drain, but it kind of requires cooperation from the company culture's part, too. Continuing on about the linearity. Currently, if you don't like the sometimes-described-as pervy way of how the conversation happens, it's just going to leave a bad impression for you. I'm personally kind of liking what I read because I wouldn't know how to talk to girls anyway but the criticism is maybe a good one to take into account that many wouldn't talk to girls like that. Somebody suggested designing conversation choices after the classes: hardcore, charm and knowhow. Having a class icon next to them would also kind of let the player know at least the writer's intent about how it's going to be taken.

Fine, fine. You've successfully dragged me into this exchange, too. These two graphs show quite a few league brackets from just this month and how the outcome depended on the player level. They both have the same data, but since I'm no data scientist, I'm giving you the two versions with different trend lines. The first one uses a second-degree polynomial trend line, while the second one uses a logarithmic one. How these look a fair bit different reminds me of the proverb: lies, damned lies, and statistics.

This league I was getting to try out predicting which opponents would be boosted at blessings change and who wouldn't be. I achieved 87.25% accuracy in my prediction. However, I should note that it's still of very limited usefulness due to it only starting to work a day before BC, and reaching peak usefulness just before the BC. Regardless, it's another tool in my toolbelt against the absolute fiercest of opponents. There's definitely something wrong with recording my simulated scores, but fortunately it happens like 1 time every 100, and I can just copy the simulated scores from other fights from that person. So, I'm expecting the statistics data is intact after those changes. I underestimated RustyDuck's average (just about the only one I saw as potential trouble), but fortunately I had been putting in more effort than was needed against the supposed average from Rusty, so I ended up winning. And sort of barely at that, so resource usage was well proportioned with how difficult (easy) the league was. Statistics My score on Mon 19:09:43 (UTC time): 4678 (100.95% of what my algorithm would expect with completely average luck) Expected score (my algorithm): 4634 Expected score (HH++ algorithm): 4580.5 (there were a lot of fights where ego-giving domination was used, where HH++ algorithm, I'm expecting, is wrong in its evaluation) If I had been using the following booster combinations for the duration of the league but with the same fight timing decisions, these would have been my completely average luck scores: AME+3cor 7365.5 AME+2cor,1gin 7345.6 AME+1cor,2gin 7315.3 AME+2gin,1chlo 7255.3 3 cordy 7220.8 2 cordy, 1 gin 7192.1 Perfect precognition score with average luck and AME+3cor throughout: 7405.74 (24.441 average). The AME+3cor expectation was 99.457% of this. No statistics regarding fight time choices this time, since those are a bit more elaborate to make.

300 frames, and I've passed Pantheon floor 1500.

It turns out the "ambassadors" were new permanent Epic Pachinko girls. Actually, I'm kind of not thrilled about that, seeing how (mostly) emptying out Pachinko is tied to gem income through event chests, and so more Pachinko girls means more luck needed to keep getting chests. It's kind of a veteran player mechanic, anyways, widening the gap between new players and those who have been playing for a long time. But I suppose so are a great many other places where you can get gems (which is not saying much since you get them from just about everywhere).

They "kind of" have alluded to it, but not in the context of the chat feature just yet: Source So, there's hope.

It's difficult to tell what's a heading and what's not, but I think the patch notes are referring to this PoV when they talk about a new objective: - Activites 640 Finish in 1st place 320 Finish in 2-4th place 160 Finish in 5-10th place 80 Finish in 10-25th place 5 Finish Daily Missions It was asked for that they added other activities into it, and it sounds like that feedback got through. A bit "late", but I guess determining whom to compensate what if they did it during the PoV would have been a really complicated task. So, good that at least future Activities PoVs will have that improvement. I have no idea what ambassadors even do. Again they are launching a new feature half-finished. They haven't even touched the boss bangs after their launch, as someone on Discord mentioned. Also, we didn't get to try out on the test server what rewards the frames are going to replace or if they'll be awarded alongside the normal reward. The chat feature is decent by itself, but the worrying direction continues with it that Kinkoid goes for novelty instead of quality. And this many things to juggle hardly even makes sense from the game's current design perspective in my view. If the game's indeed supposed to be a "hentai buffet", then current features are too linked to each other to really allow you to pick and choose a lot of the time which features you'd be using. Also, two new currencies. Wouldn't maybe just frames have sufficed if they are designed to be the bottleneck anyways? Ehh, the timer at least is relatively benevolent on the messages; you don't have to be constantly using the feature. New girls and side quests are welcomed, as always (except when they are so ugly you'd rather be without them).

I went into this league with a bit of an "exhaustion" from fighting hard, but it started to seem like sliqguy was playing kind of seriously. On the first blessing, sliqguy had 13 levels on me, as well as a stronger defense team (183416.8 vs. 189137.11). Two of the big gets from him that I didn't have in a battle-worthy state were Donatella (relatively new) and Selena (a rare girl, and her being sensual, I've felt there have been plenty of better candidates to spend the most valuable gems on). By the end of the first blessing, sliqguy had already fought most opponents. On the second blessing, it was my 186513.26 team power versus sliqguy's 186279.32, and I could start making a comeback. There was some amount of uncertainty in how high up sliqguy was in average, so I decided to play it safe and spend the Kobans on extra fights against players in their weakened states while I could. It was an expensive win, but that was due to sliqguy providing just enough resistance to keep me guessing whether I needed to play it just "really well" or "almost my best". So, good game. Not quite "luck of the draw even at full power" level of a match-up, but it's absolutely fine for my blood pressure to have reprieve from those ones. And I think I got my first error 500 ban. Contests suck because of this kind of a thing as well. Statistics My score: 7336 (100.51% of what my algorithm would expect with completely average luck) Expected score (my algorithm): 7298.9 Expected score (HH++ algorithm): 7289.4 If I had been using the following booster combinations for the duration of the league but with the same fight timing decisions, these would have been my completely average luck scores: AME+3cor 7303.1 AME+2cor,1gin 7278.1 AME+1cor,2gin 7240.9 AME+2gin,1chlo 7171.6 3 cordy 7136.9 2 cordy, 1 gin 7100.1 Perfect precognition score with average luck and AME+3cor throughout: 7356.57 (24.279 average). The AME+3cor expectation was 99.273% of this. My three biggest point losses from fight time choices: * Geidt (478), 5.13 points * Elessar (496), 2.98 points * herger (477), 2.81 points I fought 37 opponents at the exact perfect time, and a further 24 opponents at 0.2 or lower points off from the best time.

This orgy days it will only be the test server account that I need to worry about with the champions. Mansion-3: equipment Mansion-4: equipment Mansion-1: shards Mansion-2: equipment Sky-3: shards Sky-4: equipment

This league saw my second Super Serious Showdown, this time with La Kajira. Both of us had the same team power defense teams during the first blessing, while I had 183416.8 and La Kajira had 178761.8 during the second blessing. So very equal in that regard, but La Kajira of course was at level 498 and I was at 485. Actually, I could have been level 486 because I neglected to spend the experience rewards from PoV during an XP contest. I was out at reset time then (I kind of foresaw that but still) and I didn't want to be sniped. Thus I didn't really participate in that contest. It turns out it was even an easy one, so it's definitely a mistake on my part. I started this bracket with an AME, 2 cordyceps and 1 ginseng. This was not according to plan, but I decided to make the best of it and fight mostly opponents who had a better score prediction at AME+2cor+1gin than AME+3cor. My management of this blunder reduced its negative effect to just 0.9 points. It could have been a couple of points if I didn't do this disaster management. Or, I could have waited for those boosters to run out before starting battles. Anyway, not a big deal. Even an acceptable booster combination to start with, I'd say. This time I feel like I had learned from my battle with bla not to take a serious opponent roughly at my power lightly. From the get-go I went with a mindset I would be willing to spend, spend, spend, as long as it gave even a small theoretical advantage. And even though La Kajira stayed at 0 score for quite a while at first, when she did start fighting, it looked like she wanted it rough as well. This being only my second truly all-out fight, I was absolutely still honing my strategy. This time I pivoted a bit more toward taking a riskier approach to blessings change, leaving even some opponents, that I would have considered important targets with my previous prioritization, unfought until BC. It might not even look like that since I already fought a whole lot of people before BC due to there being quite a few who would actually get stronger with their current team at BC, and I'd also be stronger on top of that. But it's not just about my strength which blessing is the better one to do fights; it's also what kind of power I expect my opponents to have. This mythic blessing really separated out those who have been focusing on mythics and those who haven't, so I focused my fights on the ones who had seemingly been focusing on legendaries and fought quite a lot before BC beyond just the most obvious people. I also tried to permanently semiboost in this league. There could have been a midsnapshot of me unboosted while I was in the process of pivoting to that system, but beyond that I don't think I gave La Kajira a whole lot of good opportunities to get good points off me. When she got to 7371 points, I was thinking she might have one more unfought opponent left still and if so, I figured it almost had to be me, so I kept my boosts on even after having finished with the league myself. After all, if she had to fight me boosted, I at least had a chance, but if she managed to catch me unboosted, I'd lose for sure. I kind of hate that close of battles; there's always that nagging feeling of "what if my opponent does have fights left". My fears were realized on Thursday morning. La Kajira had gotten 68 points from her last three fights, thus winning her the league. I expect it was me fully boosted that she had fought. Statistics My score: 7430 (100.027% of what my algorithm would expect with completely average luck) Expected score (my algorithm): 7428.0 Expected score (HH++ algorithm): 7422.3 (though I expect my algorithm to give more correct results) If I had been using the following booster combinations for the duration of the league but with the same fight timing decisions, these would have been my completely average luck scores: AME+3cor 7428.9 AME+2cor,1gin 7409.4 AME+1cor,2gin 7378.5 AME+2gin,1chlo 7311.4 3 cordy 7282.9 2 cordy, 1 gin 7249.5 Perfect precognition score with average luck and AME+3cor throughout: 7462.38 (24.387 average). The AME+3cor expectation was 99.551% of this. My biggest fight time choice "mistakes" * Yami (485), 2.45 points. I fought them on Saturday when they had never been unboosted before. The best time to strike would have been on blessings change. Even in retrospect what I did seems like it was at least an understandable decision, though. * Diller9 (479), 1.98 points. I fought them already on Thursday, but it turned out they weren't boosting a whole lot. If I was prioritizing attacking on blessings change even more, this could have been avoided, since I even knew their team was only going to be decent at the blessings change. Maybe this is the clearest big mistake I made; something that could have been avoided with a little bit more patience in seeing their boosting pattern. * Kaitosan (457), 1.91 points. I fought them at the blessings change where they were boosted but in the last moments before a snapshot I figured it might be for the best to fight them then. They had been weaker in the early portion of the league and grew more powerful later on. This is really the opposite case from Diller9. It would have been avoidable, had I prioritized fighting people relatively early despite being unboosted relatively often. * lilalala (500), 1.91 points. I fought them during second blessing. They had changed team for the BC so that wasn't an option unless I wanted to fight them boosted. Best would have been to strike during the first blessing. * Jhadaen (388), 1.87 points. I fought them on Friday. They were prior to that giving 23.11 when unboosted and 20.43 when boosted. Furthermore, at the moment of battle, they were in an even more weakened state at 23.50. The huge difference between unboosted and boosted score and the even better unboosted score than prior to that made them look like a lucrative target. However, they were even more lucrative at blessings change. * moleface (500), 1.76 points. I fought them on Tuesday. They were boosted on blessings change which I had been been gambling on for them, but it turns out to have been a mistake. The opposite lesson from Jhadaen. * Tirek (472), 1.74 points. I didn't fight them yet on the blessings change, but they became stronger afterwards unboosted than they had been boosted at the blessings change. However, this wasn't even the high for them; that was before the BC. * memelas (500), 1.66 points. I fought them on blessings change. However, at that point they had already switched to the new blessing's team. Before that I only had two chances to strike them for the best score, one on Thursday and another one on Friday. Perhaps it would have made sense to fight them on Friday when I had collected enough information to know they were boosted a fair bit and they also had a nature team. But at the same time, as far as I can tell, 51 opponents were fought at the exact perfect time and a further 22 opponents were 0.2 or less points off from the perfect time to strike. Importantly, I lost just a fraction of a point from the timing decisions from opponents whom before BC I had deemed to be less powerful than at BC. So that part of my strategy seems to be working according to this small sample size. The rest of the fighting decisions seem to be more about luck whether they work out, unfortunately. Closing thoughts La Kajira has in my view reasserted her place back above me in the toughest player rankings. But maybe I'd caution reading too much into a win with such a small margin. When bla and I fought, I remember him saying: "may the luckier of us both win". It's at least not due to a singular player that I lost this time, but if a couple more had unboosted during blessings change with which I took a somewhat risky approach, I might have even won. I didn't expect the BC to be quite "a string of 25s" good, but it's also the time when it all started going wrong. Before that point I think I had even good luck, but it turned kind of bad afterwards. A single bad loss against me by La Kajira, which was certainly possible, could also have been enough to make me the winner. Or... maybe even that one player level I could have had would have given me a few extra points (I didn't actually calculate this) and then it would really have been pretty much a tie. Or maybe there was someone I could have leveled up in auxiliary teams; I had already pretty strong ones, though. But enough excuses. I lost fair and square, and at least as far as my own score is considered, it was all in all with very much average luck. I can't be salty about average luck and going in with a mindset of all-in, now can I? Thus my winning streak of 9 comes to an end. I'm honored that it happened in such a way; it could have been against someone very undeserving due to me using too little boosting in an effort to save resources. So, if anyone was to cut my winning streak, I'm happy that it was against an equal with us both (probably) doing our best. I'd like to hear what it was like from the other side if you still visit the forums, @La Kajira. Also I'd note that zoopokemon seems to have been growing in power considerably lately. It's not even that bad of a score for mostly level 650 teams.

Yeah, I have also been taking note of this imbalance. But, not to derail this too much, to say these are both new girls is incorrect. The wiki says about Lapis: "Random drop by beating Ninja Spy in Ninja Village between April 10th-14th 2021."

These two weeks have brought about some very significant developments for the rankings: On March 31st, Rarum beat bla. It was a very close one, too, so while normally I'd put the victor above the defeated, I've understood bla is able to get their girls to level 750 at a moment's notice, so another victory by me against bla might not be in the cards unless the blessings are disadvantageous for bla On 31/3, karistea won against RedEagle. No surprise there, but by how much is of significance: karistea got 24.12 and RedEagle got 23.4. On 7/4, Shado won against karistea and RedEagle. A real surprise win, since karistea was highly ranked and Shado wasn't even in the top 10. Here RedEagle did a little bit better but still not all that great: 23.68. Shado got 24.03 and karistea got 23.8. On 7/4, bla won against sygfried01, La Kajira and SJL996. The last bracket especially shakes things up quite a lot, so let's dissect it a bit further: Name Level Score Average bla 499 7453 24.36 sygfried01 500 7437 24.3 La Kajira 497 7386 24.14 SJL996 500 7380 24.12 First of all, the expected winner, SJL996, came in fourth. Word of mouth says SJL "made a mistake in the beginning and wont try hard this week, but still 3 AME and only one window to be unboosted". So, I'll excuse it somewhat, but even still bla got the highest recorded average yet in a league that was very difficult. Not only the top 4, but top 15 also looked quite tough. So, even if SJL was fighting their best, I'm not sure they'd have won regardless. I would say that I've been underestimating bla in some previous rankings where I still had level 650 girls while bla had level 700s. So with this new record average, they jump in front of me and I'd say karistea and La Kajira as well. Maybe not SJL still; I don't think bla has the variety in their battle teams yet to win consistently against some of the consistently super tough players. Also I'd note that while certainly an impressive average by bla, it was also based on a bit of a gambit to my understanding. Regardless, it's clear bla's strategic and tactical skills are also very high aside from just the raw strength. But another surprise was sygfried01. They used to rank really low, but now got a superb score. I'd be a bit careful about giving them too many rank ups because they've not yet shown consistency, but it'll be interesting to see if they continue doing well in the league. Judging from this week where I'm up against La Kajira and she only has a level 700 team, I'd actually predict myself as the winner. We'll see if I regret this, but proactively I'm going to advance myself past La Kajira, actually, and by extension karistea, who also had another reason of dropping in ranking because of the loss against Shado. Furthermore La Kajira has demonstrated the ability to get really high scores before, but I've not seen any notable ones in a fair bit. Even if her scores are consistently high, they don't tend to be in the >=24.2 range high. With a lot of consideration, I'm updating the ranking to the following: Lgm123 (still on a winning streak that's lasted the entirety of time I've had information on all D3 winners, so from 20/1/2022 at least) SJL996 bla Rarum La Kajira karistea sygfried01 sliqguy And honestly I'm for now going to delete two ranks altogether since I don't think it's very doable to rank these. I see inconsistently good plays and sometimes good win streaks but no idea if they are against any decent players or not. RedEagle also gets dropped because of this, and even without that they've been doing so poorly recently that I'm not sure they are fitting in a top 10. I suppose instead I'll give you: Honorable mentions Based on demonstrated power and win rate, but not in any particular order. HenNaYatsu (recent top 1s: 7/4, 31/3, 24/3, 10/3) Hugh (recent top 1s: 7/4, 31/3, 24/3, 17/3) RedEagle (recent top 1s: 24/3, 17/3, 3/3) Shad0w (recent top 1s: 31/3, 24/3, 17/3) Shado (recent top 1s: 7/4, 10/3) Thulsa Doom (recent top 1s: 7/4, 31/3, 24/3, 17/3, 10/3)

I used a mixture of these booster combinations: * AME, two ginseng and one chlorella * 2 cordyceps and 1 ginseng A very reasonable cost for a top 1. It seems I might have finally fixed the issue of not all data points being registered correctly (halfway through this league), and with the vastly improved speed from the new battle algorithm I'm able to do some serious amounts of predictions. Not that it is really expected to improve my score, but I think you'll see even more data about this just starting league. This starting week I'm up against the mighty La Kajira. And I already made a mistake of not doing the season battles for the kinky cumpetition before putting on an AME (not that I was planning on finishing it anyway), and I went into the league with the slightly less than best set of boosters. Oh, well. It'll be an interesting one, I predict.

Oh! Zooming works by the mouse wheel. Now I get it. I wouldn't have tried if I didn't know there was a way to zoom.

You did kind of also move the post to QA, with the reasoning being that "it's not a valid suggestion". Usage of mod powers to enforce that statement is a more powerful signal than just an opinion about what is acceptable feedback, and I could maybe retort to that focusing on the wrong detail that it seems you're here dodging the issue of it being far from obvious that OP's post shouldn't be accepted as feedback. It's fine and all to have moderation, but in this case it feels a bit arbitrary instead of based on a set of rules. I don't think anybody's criticized you being helpful telling about the script.

It would still be better for the probably still majority of players who don't use any user script if quality of life improvements were inside the game proper. So, I don't agree that it's an invalid suggestion. Maybe "unrealistic" due to Kinkoid's current approach of being dependent on unpaid outsiders to make their game playable. But in a different world where HH developers were focused on user experience instead of cramming new things to do everywhere, maybe a community script wouldn't be at least "mandatory" for play. And maybe a refocus on user experience would get them to see clearer what features can even be made to work and what's a usability nightmare.

I guess I could edit this post with further data points, if I remember. Test server: Stage 2: 35 shards Stage 3: nope Stage 4: 35 shards Stage 1: girl Main account: Stage 1: 35 shards Stage 2: nope Stage 3: 35 shards Stage 4: nope Stage 1: nope Stage 2: ... I'm getting a bit annoyed here. No shards. Stage 3: girl

I need to spend roughly 65000 gems to unlock level 750 girls myself, but this is more like a lower limit since it doesn't differentiate between gem types. I have 22420. There's like no way if blessings aren't bad for you that I'd win against a full level 750 arsenal combined with your skill, I'm thinking. So choose whatever suits you, more realistically than that boasting. 13 mythics are at level 700, 1 at 350. 9 legendary 5-stars are at or have unlocked level 700. 17 are at or have unlocked level 650. 7 are level 600, 1 is at level 550 and 7 are at or have unlocked level 500. Epic 5-stars are at 650 and 600. One rare 5-star is at 700, three at 650 and the rest are hardly of consequence. Of course there's a fair bit of common 5-stars at 650.