Talk:High scores/Archive 2

Archive 1 (11/23/2008)

The reason why I added ZNA会長's scores back: they were deleted from the Japanese scoreboard voluntarily, in the same vein as AM and UnKnown deleted theirs (Extra and Lunatic modes, respectively, both of which we have kept). I think it makes sense that we keep the old records even if they are no longer present on the official boards, simply because no-one will keep track of them otherwise (which kind of nullifies the point of mirroring the jp scoreboards in the first place). Besides, we've already been doing that anyway. — moozooh 20:23, 11 November 2008 (UTC).

I missed ZNA's score by accident. My apologies. 67.164.3.21 02:17, 12 November 2008 (UTC)

Also, I bolded highest-in-category scores instead of marking them with a single asterisk, because that is undoubtedly easier to read and skim through. Single asterisk now denotes highest-overall scores instead. — moozooh 20:23, 11 November 2008 (UTC).

I decided to use asterisks rather than bold because as per our earlier discussion, the character categories "are created equal." Because Player X achieved a lower score with Marisa-A than Player Y achieved with Marisa-B, player X's achievement isn't necessarily any lesser than Player Y's. By using asterisks, I note that Player Y's score is the highest within a difficulty setting, but not to draw a whole lot of attention to it. For this reason, I think */** works well for difficulty record/overall record. 67.164.3.21 02:17, 12 November 2008 (UTC)

Well, as long as all categories are present, I don't see the problem with bolding highest-per-mode. That way the viewers will be able to appraise the value of each achievement by themselves, but it won't hinder the readability and other statistical purposes. — moozooh 15:49, 12 November 2008 (UTC).

As for the bracket "abuse" with players' names, that's easy: it takes little effort to add brackets to every name, but the navigation is improved in return, and it also saves the effort of only keeping the topmost instance of the name (which is a subject to change) linked. Double names (usually meaning "player A at player B's place") are kept track of separately for that reason. — moozooh 20:23, 11 November 2008 (UTC).

You have a point regarding the navigation. However, I don't think we need to create an article for every player--just the ones who have established their notability through holding overall world records, or numerous indiv character/difficulty records across several games. If a player's only claim to fame is appearing once or twice on that whole records page, I don't think that really warrants more than a mention, especially if he/she doesn't have a blog/personal webpage so that we could write something more than a list of scores.

Of course we don't, and it's not a problem if we won't. :) Though maybe somebody else will have something interesting to say. We may have more people maintaining this part of the wiki in the future. — moozooh 15:49, 12 November 2008 (UTC).

However, how do we deal with 三種の神器・かがみん? Apparently, the replay says the player is ASAPIN, and it suggests that there's some relation between him and Kagamin (but what's the deal with 三種の神器 in this context, then?). — moozooh 20:23, 11 November 2008 (UTC).

I think it's just an inside joke between them. I've also seen かがみん@ZNA (the actual replays were done by Kagamin.) The player who achieved the score is the one that counts. 67.164.3.21 02:17, 12 November 2008 (UTC)

In regards to Kagamin and ZNA, they definitely know each other in reallife. They play the same games, and have records on the same games (like Mushi Futari, DoDonPachi DaiFukkatsu, etc.) as well. If only I could find and contact either of them, I'd just ask. Trying to communicate with the Japanese without knowing their language is pretty hard, though. :\ — moozooh 15:49, 12 November 2008 (UTC).

Also, regarding Imperishable Night, does the game itself keep track of A/B endings for each character? (I don't have IN installed on this comp, since my old comp crashed.) 67.164.3.21 02:17, 12 November 2008 (UTC)

No, but the jp scoreboards do. Not sure what to do here; apparently there are going to be 108 categories for this game because of it: (4x difficulties * 4x modes + extra) * 12 characters. Not that it's anything bad per se, just somewhat hard to read. After all, I've been maintaining a similar score table (except with end-of-game statistics) offline and it didn't kill me, haha. — moozooh 15:49, 12 November 2008 (UTC).

I'd just stick with B only, leaving 60 categories. I don't really see a need to create a sep. category for the "A" path since 5/6ths is the same anyway. 67.164.3.21 12:55, 13 November 2008 (UTC)

You know, something that annoys me about these is that some of them at least are of the score milking variety (including well timed suicides), and that kinda pushes things over the top. I think there should be a section for 'natural' scores (as in scores you'd get playing the game efficiently yet normal) —Preceding unsigned comment added by [[User:{{{1}}}|{{{1}}}]] ([[User talk:{{{1}}}|talk]] • [[Special:Contributions/{{{1}}}|contribs]])

Every score listed on this page is of the "score milking variety." In a score attack, maximising score is the primary objective -- and if a Touhou game's scoring system encourages deathbombing, that's exactly what the pro players do in order to remain competitive. You can set your own arbitrary definitions for what constitutes skill, but from all that I've seen, ZUN's scoring systems usually aren't too far off. Do note that the top players can easily capture all spellcards without bombing, and that all of the overall WRs are effectively no-miss clears, as all deaths are intentional suicides for scoring purposes. 67.180.29.233 11:26, 30 November 2008 (UTC)

Still there should be a notice for natural and milking High score playthroughs. While very skilled, it doesn't make for a good show. That and milking it sets the bar way too high.

I'm not sure you understand what "high score" means. It means high score, not good show. If you're up for compiling a list of "good show" replays, then name it accordingly and gather the relevant material. Everything is in your hands. — moozooh 23:01, 1 March 2009 (UTC).

The royalflare boards are in the process of adopting a new high-rank score labeling system that's already working for EoSD, PCB, and MoF. The #1 scores in each character category are now marked with ☆, #2 with ◇, and #3 with △. (Go figure.) But the real question is, what do the small numbers listed alongside each score mean? It's not a secondary score quantifier like cherrymax count, not a percentage of highest known or theoretical highest scores, and not some kind of a time stamp, either. It currently seems to be maxed out at somewhere around 80.0. Any ideas?

Actually have I no idea what they are. My guess is that they're like Cyberscore points (see http://www.cyberscore.net/chart-38016.php) for a future ranking system that's not just based on number of 1st/2nd/3rd place scores. 76.102.38.97 00:57, 14 December 2008 (UTC)

After deleting the link to LYX's rpy, I've been thinking: maybe we should actually upload all replays in that manner, so that they are preserved right here in the high score table? I've noticed that many top-score holders frequently delete their royalflare board entries for no explained reason, and, in case no-one has saved the replays to their computer, they can as well be considered lost forever. It's obvious that, if we are to preserve these replays, we should have them online in some form. Thoughts? — moozooh 13:15, 1 March 2009 (UTC).

Yes, we should. Personally, I have all Lunatic and Extra WR replays saved, but not the individual character/difficulty runs (I'm not that fanatical, heh.) It's beyond me why so many players feel compelled to delete their replays before they leave (add Slime to the list of departed Touhou veterans; his blog hasn't been updated since November '08.) 67.180.29.240 03:14, 10 March 2009 (UTC)

Re: MSH's 485 million UFO replay. Yes, it is slowdown-assisted. The guy has been notorious for doing so in the past with MoF and SA. All of his other TASes, however, have been obsoleted by legit replays.

http://bbs.thproject.org/read.php?tid=63344&page=3 MSH stated that it is not slowdown-assisted. (KyoriAsh 12:50, 9 March 2009 (UTC))

No, he says that his death on Level 2 was entirely his fault. Where did he make any references to tools? 67.180.29.240 03:12, 10 March 2009 (UTC)

OK, the entire quarrel thread between me and MSH has been deleted by moderator there, so no more answer from MSH (KyoriAsh 06:45, 10 March 2009 (UTC))

Sorry, I'm not going to believe the guy any sooner than the royalflare moderators will. I know he has submitted MoF Marisa-C Lunatic replay to the scoreboard at one point, and it got promptly removed, as I see it, due to obvious traces of tool-assisted play: extremely high level of execution coupled with a clearly suboptimal strategy, noticeable jerkiness of evading moves, being able to near-freely maneuver in clouds of bullets even in very stressful situations. It doesn't make it hard to notice it all because I've been a member of both tool-assisted and unassisted speedrun communities, and having seen hundreds of examples of either type of play, I can at least tell: this is exactly how a replay would look if it was slowdown-assisted. — moozooh 18:59, 9 March 2009 (UTC).

According to MSH, that replay is not meant to beat high score, it just serves as a guide of how to clear Lunatic and he didn't sent the replay to Royalflare (which leads me scolded and threaten by MSH in Touhou Town BBS as he thinks ME submit his replay to RoyalFlare). I'm not sure if he really use slowdown-assisted tools to play... quote from someone from Touhou Town BBS: "If he really use, why he still can die in Stage 2". Link (not sure it is in page 2 or 3 or maybe reply deleted). (KyoriAsh 21:11, 9 March 2009 (UTC))

"If he used tools, why did he still die on Stage 2?" Low-fps tools wouldn't have been able to prevent such deaths or mistakes; only re-recording tools would be able to cover them up. Currently, there are no known re-recording tools for Touhou games, or we'd be seeing truly flawless runs that are perfect to the last frame, and such replays do not exist.

I suspected that it was tool-assisted for the same reasons as moozooh. If you've watched replays by long-time famed players such as AM, YASU, Kagamin, etc., their character handling/maneuvering all have a certain degree of smoothness that sets them apart from others.

Maybe MSH didn't actually use an external tool, but has a slow computer and/or heavy CPU load in the background that causes the game to slow down naturally. I have a five-year-old computer and if I overload it with lots of apps in the background, TH naturally runs below its usual frame rate.

However, I am going to leave his achievement in the world records section anyway, until it is beaten. Historically, slowdown-assisted replays have never surpassed real-time replays in the long run, but they serve as great references, and their valuable contributions to the Touhou high score community should not be discounted. In fact, they are always appreciated, as long as their distinction from real-time runs is duly noted.

That said, I would really like to see all Touhou games played with re-recording one day, just to find out each game's true scoring limit.

67.180.29.240 02:55, 10 March 2009 (UTC)

BTW, I didn't spot any laginess except for some heavy graphics like some SpellCard, but the replay is 0.0%, which means did not drop any frame, how you guys can spot it? (KyoriAsh 06:45, 10 March 2009 (UTC))

Additional Reply:

May you guys can provide some Undoubted Evidence: Proof that perfectly proves the maneuver of MSH was created by External or other means that will slow down the game but won't affect the slow down rate Ie: which part of which stage of maneuver cause you suspect this to happened.and why, or how "unnaturally" movement. (KyoriAsh 10:33, 11 March 2009 (UTC))

Ok, here's an example. As you can see, the slowdown rate is 0.0% (and I even die at the end, haha), but the rest is something I bet you would want to do for real. :) I would have believed MSH if I didn't know something like this was possible. I didn't try to do something extraordinary, though, and I was playing at a pretty low speed which made my movements look really unnatural. Higher speeds will naturally result in more natural-looking movement due to their proximity to legit conditions of both reaction to danger and the process of decision-making (which is the reason even veterans make mistakes when overwhelmed with choices to be made in the time frame of milliseconds).

For the record, the tool I used was SpeedGear v5.00. I wasn't able to make it work with UFO, but I'm reasonably sure it's not the only speedhack utility available, and probably isn't even the best. I haven't made research on this yet, although it may prove useful for full-fledged Touhou TASing in some (currently unforeseeable) future. — moozooh 20:05, 11 March 2009 (UTC).

Well, MSH has released his new Lunatic replay of UFO, current record is 685 million... (KyoriAsh 13:24, 18 March 2009 (UTC))