原文转自国际网
https://minesweepergame.com/forum/viewtopic.php?f=9&t=1316&sid=1f56e4b51e75715ade239ef2c3763b68
中文翻译在2楼，感谢某雷圣翻译，濮天羿校对。
Re: World Record 29.43 (Ze-En Ju)
Post by qqwref » Tue Sep 22, 2020 5:06 am
I'm not posting this to try to stir up drama or controversy, but I feel like someone needs to bring these kinds of things up, and nobody has yet. I am suspicious of JZE's scores, for a lot of reasons, and I think something unusual has to be going on. It's not necessarily cheating, but in my opinion, things don't add up. Here's a write-up of what seems bizarre to me:
- A lot of the expert records have quite high 3BV. From 37.87 to 32.93, all records were 140+ 3BV, including a 167 which is practically an average board. The 31.53 that he had just before the current record was 160 3BV. I know flagging players can be a little more tolerant of high 3BV than NF players (like me), but this doesn't feel like the behavior of a top-level player who is pushing their limits and going for the best time they can get. Even in NF, you can see his last 3BVs of his PBs: 152, 149, 159, 135. Those are still very high! If you look at his 2019 times, he had PBs with 3BV like 117, 127, 104 which are closer to what I would expect for a top player.
- As I previously mentioned, the WR in Intermediate 3BV/s is about 10% better than the previous WR. It's a nice board, but you shouldn't see jumps like that on a world record that so many people are competing over. Also, he plays the last 4 seconds without flags at 6.8 left clicks/sec, which IMO is more physically challenging than 6.8 total clicks/sec with two buttons. This includes a 50-50 guess made with no hesitation. There is nothing impossible here, but it feels too good to be true.
- The fast improvement rate is weird. It was 3 months from sub50 to sub40, then 5 months to sub35 and 6 months to sub30. Again, it seems too good to be true. Normally top players play for years to get good results and optimize their times. I played a lot starting 2009, got sub50 in 2 years, then 42.x in another year, after which I didn't play as consistently. Kamil took a year from sub50 to sub40, then 10 months to sub35, and 9 more months to the first 31.xx. Perhaps JZE may have had a lot of skill in mobile play before this, but just getting used to high-level mouse movement that fast is unusual.
- Related to the above, he is also supposedly world class in mobile minesweeper; sliding puzzle solving; and schulte table, including a blindfolded variation. I have spent years on sliding puzzle solving, and I know other top sliding puzzle solvers in the English speaking community, who have typically spent years and hundreds or thousands of hours solving sliding puzzles to get to this level - and even so, some of the single solve times on the popular Chinese app are much better than the English program's world records. (To be fair, many of the English program's players focus on improving their skills at large puzzles. Also, it is possible that dragging on a touchscreen is faster than the English program's mouse or keyboard controls.) Similarly, I know it takes many hundreds of hours to achieve 2.7s on an app called Touch the Numbers, which is actually *easier* than the Chinese app's version because the numbers disappear when you touch them. There are a few players under 2.1s on Touch the Numbers, but they have played for something like 7 years, with some of them having 500,000+ attempts. My point is that these are several *distinct skill sets*, each requiring hundreds or thousands of hours to get to a top level. Even mouse controls and touchscreen controls are distinct skill sets, and getting a world-class Minesweeper time with mouse requires a very high level of mouse skill that often takes years to develop. You cannot train mouse skill by practicing on a touch screen. Anyway, I do not see how someone who is 12 years old could have had time to spend so many hours on so many distinct skills.
|
|
最近一次修改:2020-9-24 17:21:31
|
|
|
|
- Many of JZE's videos include weird mouse behavior. The 31.62 has some long jumps that seem too fast, as well as some situations where the mouse rapidly goes in the wrong direction and then returns. The int 3BV/s record has very odd clicking order, and while I understand that this is more efficient for mobile play, clicking in this way with a mouse is more difficult than clicking in a normal order, which is why top players usually don't move like this. The 31.53 has fast flicks for no reason, and sometimes moves from one area to another too quickly. The 29.43 looks more natural to me, but also has some weird jumps and back-and-forth movements, and in addition the ending seems to be a very risky way to finish off such a fast game.
- Speaking of mouse skills, my friend Kaori, who does not play Minesweeper but has himself spent thousands of hours on various games involving fast aiming and clicking with the mouse, and is the administrator of a community about those kinds of games, believes that JZE's Minesweeper replays are "obviously" sped up. He said "His acceleration on direction changes is absurd and I've never seen anything remotely close to it ever", and "The entire plays have absurdly consistent rhythm, except for the occasional long flicks which are .. unbelievably fast" as typically "rhythm, cleanliness, and smoothness are sacrificed to boost speed at a top level".
NF player. Best scores 1-10-39.
|
|
以下是某雷圣友情翻译：
我发帖不是想要带节奏或者挑起争端，但我觉得确实要有人来把这件事提出来，现在还没人提。出于很多原因，我对鞠泽恩的成绩很怀疑。我觉得这里面一定有故事，下面是我觉得奇怪的地方：
1、很多的纪录局是大bv的。从他37.87到32.93，很多纪录局都是140+的图。包括一个167，几乎是平均bv了。他上个纪录31.53是160的图。我知道FL玩家大图比我这样的NF玩家强一点。但是这确实不像一个顶级选手冲击自己极限纪录的样子。包括NF纪录在内，他的纪录局BV也是152,149,159,135这些数，还是很高。你们如果看他的19年的成绩，纪录局都是117,127,104这种局，这更符合我对顶级选手的印象。
2、我之前说过了，中级bvs纪录局比原纪录好了10%。图确实好，但你爆了一个那么多年都没人破的纪录这么多就不科学了。而且他最后
4秒nfcls高达6.8，在我看来这比用双指cls点到6.8还难（这句话可能不对）。并且还有个地方二选一死猜丝毫没有犹豫。这也不是不可能的事，不过it feels too good to be true。
3、这个进步速度过于奇怪。他三个月从sub50到sub40，又用了五个月sub35，再过了六个月30。it seems too good to be true。一般来说顶级玩家都是用了很多年才扫到这么好的成绩，达到自己的极限。我从09年开始爆肝，两年才sub50，又用了一年才43，以后就没一直肝了。母狼sub50到sub40也用了一年，然后10个月sub35,9个月32。可能鞠帝之前已经是手机扫雷大佬了，但是这么快就有了这么强的操作鼠标的水平，不太正常。
4、和上面相关，鞠帝也是手机扫雷，华容道和舒尔特的顶级选手，包括盲玩，我认识一些顶级华容道玩家，他们花了成百上千小时玩了很多年，还是被鞠帝的单次纪录完爆了。虽然他们玩高阶玩得比较多，并且触屏应该比鼠标更快。我也认识一些玩舒尔特2.7秒以内的，最强有几个进了2.1秒，他们玩了7年，开了50万把以上，并且他们玩那个版本还要简单一些（对应联盟里的“简单”模式）。我的意思是这些游戏存在一些独立的技术要领，每一样都要花成百上千个小时训练，甚至触屏和鼠标的控制都是独立的技术，不能同时训练，我觉得一个12岁的小朋友也不可能花了太多时间来训练这么多独立的技术。
5、鞠帝很多录像鼠标轨迹不太对，31.62那盘几个长距离移动太快了，有几个地方鼠标往错误的方向去了又马上折回来。那个中级bvs纪录的点击顺序太奇怪。虽然我后来知道了手机扫雷这样点效率更高，但是电脑上这样移动鼠标很难受，顶级玩家也不喜欢这样移动。那个31.53有一些快速轻点我不太懂，长距离移动也太快。29.43挺正常的，但也有一些奇怪的鼠标跳跃和来回移动，而且结尾这么快太冒险了（鞠帝：谨慎收尾）
6、说到鼠标技巧，我有个不玩扫雷的朋友，他别的用鼠标的游戏（包括快速瞄准）玩了很多，是很多这类游戏社区的管理员，他相信鞠帝的这个录像显然是加速了，他说“这个人的准确度和变向能力很扯，我甚至从来没见过鼠标操作能接近这个水平的“，“整个录像的节奏异常稳定，除了几处超快的远距离瞬移。这和典型的顶级玩家为了提速而牺牲节奏、操作精度、移动稳定性不同。
|
|
|
说到鼠标技巧我觉得有必要做一个鼠标技巧测试器，但还没有成熟的想法。
|
|
|
我也不敢朋友圈里公然找人啊，怕被鞠帝的爹妈打，理论上我跟鞠帝的物理距离应该不超10公里，都不知道鞠帝的爹妈是不是知道儿子是世界第一。。。。
|
|
老卡也发言了，说最好多录几个高水平录像（而且不用阿比特），我就怂了。（毕竟我也是这么过来的）
Re: World Record 29.43 (Ze-En Ju)
Post by KamilSaper » Wed Sep 23, 2020 9:40 pm
If Ze-En Ju can get sub33/sup5 3bvs expert/sup7 3bvs int every 2-3 days then indeed lets give him a camera so that he could record him playing for 100% time. If he records himself getting any of these scores within a week since he receives the camera on a version different than arbiter then i will believe his scores are not fake. I'd also like to have a view on a full day recording.
I mean, I believe he can be actually fast, but if it's not a problem for him to get better scores than my records every few days, then it shouldn't be a problem to record one with a camera. 1 hour recording of just random games is not enough.
0.49 - 7.03 - 31.13
NF: 0.49 - 7.03 - 31.51
|
|
|
Umm通篇看下来感觉提出的点都好主观..（尤其最后一句是什么鬼
|
|
|
就是就是，本来我还想帮某雷圣校对校对，读到一半就酸得读不下去了。谢谢扑地了。
|
|
|
|
