Famous Cheaters

From MinesweeperWiki

Jump to: navigation, search

Cheaters occasionally try to get on the world rankings, and sometimes they succeed. This page will collect some of the most famous cheaters. They are a difficult category: not all fake scores are deliberate, and not all cheaters become famous! Let this be a Hall of Shame.

Contents

[edit] João Livramento

João Livramento (ID #2982) was added to the Authoritative Minesweeper with scores of 3-12-55 on 2 Feb 2001. This easily made him the 2nd fastest player in the world. (Lasse Nyholm had 3-16-47). João sent screenshots and his scores were accepted as legitimate. The only surprise was how such a good player could take so long to find the site.

[edit] Strike One

Joao Livramento claimed a 12 one month after his first 28
Joao Livramento claimed a 12 one month after his first 28

The amazement nearly halted when Matt McGinley wrote in the Guestbook on 11 Feb 2001:

Okay, Ive been looking over Joao's 12 second board and Ive come to the conclusion that there is no possible way that anyone can complete that board in 12 seconds. At the fastest, maybe 17 or 18, but surely not 12!!!! I was wrong about Sriram's 2 sec. board and I might be wrong about this one, so respond to this if you disagree with me. But there's just no way!

Lance Votroubek replied:

It looks like his 12 second board is doable in 12 seconds... coming from a person who can't even break 20 in intermediate. But if you are saying that he is faking it (I'm not accusing or anything), it wouldn't be so intelligent of him to claim to have a world record, ya know?

Joe Nuss also believed it to be genuine:

for what its worth, i think joao did it. a lot of the board openened with that one click, and i think if he played a near flawless game he could have done it. however, that would raise the question that if he got a 12 on that kind of crappy board, how did he not get even better on a really good board? [...] i mean, ive gotten better boards than that before.

Things quieted down for a few days. It is easy to call someone a cheater if their scores are better than yours. Perhaps João had made his scores earlier and retired before finding the site. That would explain the sudden appearance of an expert player.

[edit] Strike Two

The retirement scenario fried when Lance Votroubek dropped a thunderbolt in the Guestbook on 14 Feb 2001:

Has anybody else noticed that Joao was already in the record books... with an intermediate 28????

João had earlier submitted a record of 28 seconds made 27 Dec 2000, but his 12 had been made in January! Damien Moore failed to note this because the site did not yet use a database to prevent duplicates. Matt relaunched his attack, claiming that it was impossible to "go from 28 to 12 in a matter of months". However, Joe retorted that Matt was hypocritical: "you of all people should know [...] didnt you go from like 30's to 18 in a few months?"

Matt replied swiftly and sharply:

Joe, I went from 32 in September to 18 on Nov. 21. A considerable jump in a short amount of time, yes, but listen here. I got 17 on Jan 2. I havent been close since. Yes you can go quickly thorugh the 20's, but once you get into the teens, breaking your records every month (or less in this case) is nearly impossible. Yes, you can get very lucky and get a great board and take off three or four seconds off your PR, but Joao's board is not great. A sweeper with an expert time of 38 would have a hard time playing through that board in under 15 seconds. Yes I may sound harsh, but there is no way, ANYBODY can get through that board in 12 seconds!

Joe admitted that his friend once had tricked him by editing screenshots in Photoshop, so the screenshots could have been easily doctored. This led to a brief debate about whether Camtasia videos should be required for rankings. It was quickly realised that many computers could not handle the program, and Lance declared that the only way to discover fakes was to continue scrutinising games.

[edit] Strike Three

The case against João took an amusing turn 16 Feb 2001 when Philip Culp observed:

Not to be a dork or anything but [...] I also doubt it because on the screen shot of his beginner the board was 9x9 and all the others are 8x8 which makes it hard to believe the 12 and his other scores.

This new information finally convinced Joe:

ok, joao=fake. that 9x9 beginner proves it. thats pretty funny, too :). take him off the lists, damien, you're second fastest sweeper in the world now, man! good job! :)

Damien immediately wrote João for an explanation. However, Khor Eng Tat posted in the Guestbook on 19 Feb 2001, shocked that his beginner records were also on a 9x9 grid! This led to the discovery that Windows ME had been released 14 Sep 2000 with the 9x9 grid as standard.

[edit] Struck Out

The João scandal became less important when both Damien and Sriram Sridharan passed him in the rankings that weekend. As Damien wrote in the Guestbook about his new 53 on Expert:

That feels good...I finally beat Joao and by two seconds! Too bad that Sriram had to get a 51 yesterday...it's getting hot (Lasse 66 Sriram 67 Myself 68 total).

Matt and Joe then demanded the removal of João from the rankings. Unfortunately, João remained on the rankings until April 2001 because Damien had been banned from the internet by his parents. Matt wrote a rant about João at the Intermediate Hall of Fame to vent his frustration, but ultimately João was reduced to his earlier scores.

[edit] Next Batter

João was the first cheater to be temporarily accepted into the world rankings. His debacle increased usage of video by the sweeping community, although the only program available was of limited usefulness and had been discovered only two months earlier. Estimated board difficulty was proven to vary wildly according to the skill of the player involved.

In the end, João was caught due to the contrast with his earlier submissions and his lack of response to email. (3BV had not yet been invented, but his 12 (47) and 55 (168) both require exceptional skill, especially from someone with 3-28-XXX from a month earlier). A consequence of this was Damien Moore being denied the 2nd world rank for three weeks. By the time João was reduced to his earlier scores, Sriram had passed Damien on the ranking. This demonstrated that an accurate ranking would need better methods of detecting cheats.

  • 3 - Possible real score
  • 12 - Fake 12
  • 55 - Fake 55

[edit] James Shannon

Lasse Nyholm had recently set a new world record of 42 in Expert mode when this email arrived at the Authoritative Minesweeper on 08 Nov 2002 from James Shannon (ID #3233):

A friend of mine told me I should consider looking up the minesweeper records and submitting my times. I was downloading some software on his machine for him and was sweeping while I was waiting. He watched me get a 63 on expert and got very excited. I promised him I would look into it.

I have attached a video of my most recent expert game--tied for my best time (38s). If you are interested, a bit about my minesweeping history follows.

I first learned to play minesweeper when I was seven, but I didn't really play much until I was 11 or 12. I stopped playing on beginner and intermediate about two and a half years ago (personal best of 1s and 19s respectively). In those days I played maybe an hour a week and had an expert time of 119s. Around July of 2000, I developed a bad case of insomnia. I slept on average 3 hours a week for 7 months. It was not that long at all before I was completely bored with staring at the ceiling night after night and I would sweep to pass the time. By the time I started sleeping healthily again, my expert time was down to 44s. I only play about an hour a day now, just to keep my speed up.

James Shannon Vancouver, Canada Born: March 10, 1982

[edit] A Perfect Video

The video is amazingly fast and is on a 137 3BV board. However, Dion Tiu proved in 2005 that sub40 was possible on a sup130 board. The video prevents the game from simply being a doctored screenshot. The timer in the video is accurate when compared to a stopwatch. The game is played on Windows 98 but the Timer Jump bug does not occur; however, this bug does not occur in all games. It could be possible that game speed was increased and a correct timer added, but the level of play is very advanced and would require someone already very skilled. Although he claimed an Expert world record he did not also claim an Intermediate record. As a result, he would not have been ranked 1st in the world (the usual aim of cheaters). He could have used a solver, but the mouse cursor does not move in robotic straight lines. Most solvers completely solve each area but James occasionally misses a solveable area and returns later. Also, most solvers do not use a mixed style of flagging and NF play. Unless he made his own clone, he could not UPK the game. In short, it was the perfect video.

[edit] A Perfect Crime

Damien Moore sent the video to Lasse Nyholm and Dan Cerveny. Both were uncomfortable with the lack of mistakes and the lack of hesitation when guessing. There was not one wasted click in the entire video. Damien wrote James for more evidence and prompted for a better Intermediate score but did not receive a response. The scores were rejected due to suspicion.

It was not until a few years later that Damien spotted a mistake while watching the video for fun. You may want to spot it first yourself:

The mistake? There is an incorrect number in the bottom right corner.

There was one other, less obvious mistake as well. Minesweeper was released with WEP 8 Oct 1990 when James was eight and a half, more than a year after he started playing! The game did not come with Windows until 6 Apr 1992, when he was ten.

[edit] Daniela Weingut

It is difficult to describe the entity known as Daniela. Perhaps it is best to let her do the talking:

28 Oct 2001 in the Guestbook:

WHAT HAVE I DONE??? Yesterday I managed only 7 HOURS of sweeping! I start doing things that have nothing to do with Minesweeper! Today I'll spend 13 hours playing Minesweeper. [...] I still make many mistakes with the new playing technique. When I practice more often, I make less mistakes. My speed has already improved.

1 Nov 2001 in the Guestbook:

I have access to the net only sporadically now, so if you don't hear from me for a long time, don't worry. I'm still going to break my best time because I can still play Minesweeper on my second computer which has no internet access. Only Camtasia doesn't run like a dream there, so I can only record at 5 frames/second. And yes, switching the versions is really better for speed. I tried it out today. I played on a 16*16 board with 39 mines (what if I get the dreamboard and lose it?) and played as fast as possible. I made a 23.

23 Nov 2001 in the Guestbook:

I can tell you how to go below 100. I did it this way: I forced myself to play slowly for about one month (after I got 136). At least a couple of times in a day I really wanted to go fast, but I didn't do it. After some time I couldn't stand it any more, I just couldn't hold myself back, played fast and went from 136 to 96 and after a few minutes to 85. Between the 96 and the 85 I got a 35 on int and a 6 on beg. This may sound silly, but it works. BTW, you can see a 57 no marking video in about 3 days.

24 Nov 2001 in Minesweeper Addicts:

I just got a 77 [...] and finally a 57! It feels so good. I only have the 57 on video, it's in the files section. Oh dear, I'm so happy!

Matt McGinley replied to Addicts congratulating Daniela for her fake Expert video. Owen Fox wrote a strongly worded reply in the Guestbook and accused her of cheating. However, it was impossible for her to have cheated...because someone had watched her play!

[edit] Daniela and Andrea

Andrea wrote the Guestbook and announced that Daniela was her sister. The game had to be real, because she watched Daniela play. Several people continued to attack the video as a fake, which led to an amusing exchange of letters. Andrea claimed that Daniela was honest and did not know how to cheat, and that strange behaviour in the video was due to errors from running Camtasia on an old computer.

Her persistant arguments convinced David Barry and Owen Fox to that Andrea and Daniela were the same person. Andrea had strong arguments against this:

I'm really her sister and not her. She hates Linkin Park, it's my favourite group; she hates loud music and dancing; she has never tried to drink alcoholic drinks, because she thinks it affects her sweeping. she's my opposite, because I like all these things and you see, she hates MTV.

Perhaps Andrea was telling the truth? A month earlier, Daniela apologised for a post in the Guestbook by claiming her sister had written it.

[edit] Cheater or Joker?

Matt McGinley and Owen Fox were convinced Daniela was a cheater. Roelof Smit and David Barry believed the video had to be a joke, because it was so badly made. Admitting that the game was a joke would have saved Daniela. However, she insisted that it was real:

If I really wanted to cheat, I would've... ...written a Visual Basic application that gives me the dream board. ...made a better intermediate time. ...used the League version and I would have written an application that makes the timer go three times slower. ...made a new world record. ...done it a long time before.

[edit] A Cheater

Several mistakes in the video, such as flags appearing after the cursor had left squares, could be attributed to a very slow frame rate in the video. One obvious mistake is when Chording occurs in one part of the board while she is solving elsewhere. Although some of the solving is brilliant, much of it is horribly innefficient. The worst mistake is that the timer stays at 0 for the entire game!

Others factors to consider: Daniela claimed to play more than 12 hours every day; Daniela knew how to cheat and could program; Daniela claimed a jump from 136 to 85 in a single day; Daniela scored 85 and then claimed she would provide a 57 second video in three days using her new technique!

Players started getting really annoyed with Daniela. Things finally calmed down when Joe Nuss posted the final comment on the topic:

why doesnt everyone cool out? it's obvious damien won't except daniela's scores; if everyone is so hard core anti-her, why dont you just not give her the satisfaction of attention? duh and im assuming the "obvious flaw" that someone mentioned before is the fact that the timer doesnt move and stays at 0, yes? why anyone is even looking beyond that to minute points such as when the smiley face changes to sunglasses and her style of play and the jerkiness of the animation? rest easy fellas; daniela won't be on the records list next time it's updated.

[edit] Jason Tran

Jason (ID #2509) joined the community from Australia on 9 Aug 2004 with 2-19-67 made with Windows Minesweeper. He took an active part in community discussions and broke his records many times over the next few months. He eventually became the world #4 with scores of 2-10-43 before cheating was discovered by Rodrigo Camargo (Brazil), inventor of the Minesweeper Clone. Grégoire Duffez (France), webmaster of Planet Minesweeper, then made a post in the Guestbook under the pseudonym 'International Minesweeper Committee' on 10 Jan 2005:

"The international Minesweeper committee carried out an anti-doping control, by examinating an urinal sample of Jason Tran. This test has been positive for a substance called "timerslowdownolone", which is strictly prohibited. Jason Tran has immediately been suspended from official lists, waiting for explanations. Here are the results of the analysis : www.planet-minesweeper.com/urinal_sample.txt"

[edit] Slow Capture

It took a long time to notice Jason was cheating because he started with his real scores and then 'improved' over several months. Very few cheaters have that much patience! He also made his 'records' on easy, believable boards. He took part in discussions and congratulated other players on their records. His progress was as follows (scores were Real Time +1 in those days):

24 Aug 2004 in the Guestbook:

"59 ==> 52.51 YESSSSSSSSSS!!!! i scored 52 last night (1am) on a 113 Board i think! anyways im really happy!!! Anybody who wants a copy can email me and ill send them a copy coz i dont have a website [...] Time for me to work on my inter score"

26 Aug 2004:

"I got a nice 36 or 34 (cant remember) board and managed a 13 the yesterday!!!! [...] Overall its been a very good month for me!!!"

Martin Toft Madsen (Denmark) then made a fake 37 Expert video on an old version of the Clone. Jason seemed interested and asked on 28 Aug 2004 "is it still possible now to do fakes with whatever method you were doing???" but Martin replied "No, you can't make a fake the way I did with the new clone. I just found an easy board, practised it, loaded it, began recording and played it. Only very little needed to be done with the video other than that to make it look genuine." Martin also said "And just to make it clear: this video was an attempt to show that it was pretty easy to make a genuine looking fake video with a clone that was developed at that time. It was not an attempt to make anyone believe that I had made a genuine 37."

On 31 Aug 2004 Jason claimed a record of 51 but said the video was corrupted. This added to his credibility as his description fit a newly discovered bug that Rodrigo fixed in the next version of Clone. He also offered to play others on the new Multiplayer feature of the Clone.

16 Sep 2004 Jason posted twice. His new score made him #16 in the world ranking:

"13 --> 11.7 wooo hooo scored 11 seconds on clone the other day!!!! so happy!!! [...] scored 12 seconds the other day!! kinda kool but not as wicked as the 11!!! hehe!!! now time to sub 50 on expert!! hmm....... hopefully by the end of the year!"

Sub50 came earlier than 'expected' on 24 Dec 2004:

"52 ---->>>>>> 46.32 WOOOHOOO First and foremost i would to say MERRY CHRISTMAS to all!! hope you all have a safe and wonderful holiday!!! Well the last two weeks have been a bit of a bummer for me coz i had many sub 50 chances!! but just slipped thru my fingers yet again!!! but wat betta way to celebrate christmas but with a new score!!! 46 seconds!!!! WOOHOO!!!!! 141 3BV Merry Christmas and a Happy New Year!! Now this calls for celebration!!!!"

Georgi Kermekchiev (Bulgaria) then asked why he did not submit his history file for the Clone rankings. Jason replied:

"the reason why i dont submit clone results is because, i play on about three computers, i have two computers at home and i scored my 46 today on my girlfriends computer!! My history is all over the place!!!"

Two more records arrived on 26 Dec 2004 and Rodrigo immediately asked for the videos:

"46---->>>>> 45 ---->>>> 43.06 (137 3bv) hehehe!!! woot woot!!! i dont believe it!! I scored a 45 this morning and then ten minutes later!! a 43!!!!!! better watch out im coming!! nahh... still a way to go yet to catch up to oli!!"

On 5 Jan 2005 Jason played a 3BV=23 Intermediate board and lost on the last click, at 9.03 seconds.

His final post was 6 Jan 2005 with a new world record of 10 seconds on Intermediate and world rank of #4:

"11-->>>> 10.96 YEP! 10 seconds! 10 seconds only just though, i did it on a 41 3bv board crazily enuf! but hey its a 10!! hehe :P I did miss out on a very possible 9 seconds earlier just one click away only one ........ click away, so disappointed it was a 23 3bv board i thinks! @dion: hehe one second closer dion! watch ur back!!"

On 5 Jan 2005, Jason lost on the last click at 3BV=23 boardWhile part of the community Jason submitted scores to the Active Ranking, which at the time did not require picture or video evidence. Starting in Period 19 (10-24 September) he submitted 11-54 (4th), 14-53 (6th), 15-54 (8th), 14-54 (8th) and 12-43 (2nd).

[edit] Happy Ending

The quick pace of improvement led Rodrigo to suspect possible cheating. He requested all videos from Jason and persuaded him to provide the Clone history file. A discussion on IRC led to his removal from the rankings and Grégoire's post. The history file covered less than 2 weeks of scores. (Note: Clone history is in RT.) A few days later, Rodrigo identified the exact program used. This program slowed the clock, so on replay Clone played at the 'correct' speed. He immediately fixed the Clone to detect similar programs. Jason apologised on IRC and left the community.

But this story has a happy ending! In May 2006, Jason wrote to apologise again and asked if he could ever be accepted. After providing his real scores (2-18-65) he was reinstated.

[edit] Benson

Benson (ID #1462) is more of a story than a person. He joined the site in April 2004 and accused the world records of being fake. This led to many players trying to convince him scores were real. The best player in the world, Lasse Nyholm (Denmark) then made the famous "To Benson" videos. After accepting the evidence his scores improved and he was welcomed into the community - until Damien Moore (Canada) outed him as a player already on the world ranking!

[edit] Fun At First

The Guestbook pages where Benson arrived are lost, but on 4 Apr 2005 Lasse made two videos of himself scoring a 49 second game on Expert and wrote:

"Never liked being called a cheater - still don't. take a look: www.lanyje.dk/vids/tobenson.AVI and www.lanyje.dk/vids/tobenson2.avi Happy sweeping everyone"

Benson repled was quickly:

"ok so Lasse Nylom or whatever that was a nice high quality video. I couldn't see your hand at all, all I saw was your head for a few seconds and what looked to me like a mouse program. No offense, but what exactly was the point of that? "

Stephen Arnason (Canada) strongly rebuked Benson:

"Benson: What on earth is the matter with you? What you are saying is absolutely ignorant. Like, it actually doesn't take very much to get below even 85 as you suggested as the maximum limit. Like are you joking that you've been playing for years and can only get like 190 on expert? I started here at like 83 on expert, something like 25 on intermediate and like 5 on beginner or something. The minesweepers here are the best of their class [...] You talk about using logic. What kind of logic is "if i can't do it nobody can"?"

This prompted Lasse to respond twice on 5 Apr 2005:

"Benson: these guys have already tried to argue rationally, I thought I'd try with a little humor. I knew very well you wouldn't take it as proof, but... not everything is about you. Hope someone enjoyed the vid - it was pretty fun to make. Maybe someday I'll make another one - where you see my hands, and a clock moving, and an audio-file of the mouse-clicks, with realtime-music in the back - or perhaps a sportsannouncer, or some cheerleaders.....We'll see - happy sweeping [...] Okay - I couldn't help it - had to try again. This time only 51, but with my hand, the sound of the clicking and music in the back. www.lanyje.dk/vids/withhands.AVI No cheerleaders though. Grr, never could keep them in my room for more than 40 secs anyways, so I would have had to break my record..."

This finally convinced Benson, who wrote:

"Well, it appears as though I may have been proven wrong? I'm sorry. I told my mom about this and she looked through and I showed her the video and she told me I was probably wrong. It just seems so impossible. I started playing this game many years ago but I haven't practiced hardcore like you guys. I just thought that maybe you guys made a program that let all of you get good scores and I wanted to too. Sorry for accusations. Can I have help with how to get better? I really want to."

The community accepted the apology, and over the next week they gave him tips. Benson improved from 194 to 169 on Expert, 81 to 74 on Intermediate and scored a 12 on Beginner. He was excited to discover Chording and seemed eager to learn, asking other players many questions. In a way, he became a favourite pet due to his naive enthusiasm. Then came a confusing time in the Guestbook where a spammer posted fake posts pretending to be Lasse, Damien and Elmar. This prompted Damien to investigate, and on 14 Apr 2005 he posted:

"the fake damien, lasse, elamr posts are all done by the IP used by all of benson's previous posts...however, the last three posts by stephen use the same IP as well (where he breaks his int score, 3bv/s values, and one lecture against benson). just putting this info out..."

Now discovered to be Stephen Arnason (Canada), Benson replied:

"Well it was fun while it lasted. To be quite honest, I was a little bored of the regular talk - wanted to spark up a bit of a controversy. Obviously when Damien had my IP addy he knew who I was. I did break my int record. I also made sure I 'pseudonymed' people who come on here often - I knew that people would know. No apologies here, just some fun as damien even mentioned - but my streak of benson/others is over."

[edit] Outcome

Stephen was already ranked with 1-16-65 when he started posting as Benson. Looking back, the fact he even posted as Stephen and criticised Benson is funny. No punishment was given, as many players found him amusing and he did not fake any videos or pictures. In the end, the community was more than pleased to have videos of Lasse playing minesweeper!

[edit] Justin DiFebo

Website of Justin DiFebo
Website of Justin DiFebo

Justin (ID #3292) was a spammer known by many names including "Michael Kenyon", "The Illinois Enema Bandit", "Al Kilyu" and possibly others. He mainly posted in the Guestbook from June 2001 to November 2001. Most posts were deleted for inappropriate comments of a sexual nature, but some players found him amusing. As Justin he claimed scores of 2-19-51. There were two highlights to his career: on 2 Nov 2001 he pretended to be Sriram Sridharan (world rank #2) and claimed that David Barry (world rank #4) had died; on 29 Nov 2001 Damien Moore (world rank #3) posted a picture and a few hours later a porn site was created featuring Damien (his head edited into pictures of men and women) performing sexual acts with men, women and animals. As Damien recollects, he showed the site to a friend at work who responded, "Nice T*ts!"

[edit] Wendy Schopieray

Wendy (ID #4108) joined the site with MSX videos of 1.41-7.60-21.69 on 15 Dec 2010. It was clear a human was playing the games due to misflags and other mistakes, and it was obvious the video speed had been increased. Innocent until proven guilty! Damien asked her to send Clone videos and explained more evidence was needed for scores this fast. On 28 Dec 2010 and the next day she sent Clone videos, the fastest being 1.31-7.46-24.93. She was then asked to send Arbiter48.5 videos because no one had succesfully hacked this program. At this point she stopped answering her emails.

There were many reasons why her videos were rejected as fake. First, her skill level was amateur and suggested the videos were 4 or 5 times faster than her actual times. Second, she claimed scores much faster than the current world records on Intermediate and Expert. Third, her 3bvs scores were insane; for example, she claimed 8.82-7.80-6.09 on the three levels with her 'slowest' games being 5.30-5.89-4.99. This was explained to her and it took her by surprise, because she was a newb and had not done her research.

[edit] Brandon Stitt

Brandon Stitt (ID 3764) is also known as “Webs Man”, “BJSwimmer”, “Silver” and “Axecutioner”. He joined the site in September 2009 and in less than a month scored 40-40-90-170-240 on the standard custom levels. The games were on MSX but he only sent pictures and YouTube videos. It looked like he was using MSX 0.34, which has cheat features but no video function. Damien wrote him in October 2009, explaining that official videos were needed and asking about his strange mouse movement (which looked like the XYZZY cheat was being used). Brandon gave a feasible explanation for the mouse movement but then disappeared.

Brandon joined the site again in February 2011 and sent official MSX videos. He did not mention his old fake scores and the YouTube videos had been deleted. His first videos were 22-35-64-125-160 on the standard custom levels. He made 25 on 8x8 in March and LC'd a 170 on 24x30 in early April. The videos looked genuine because solving was amateur and the density was low, except for the 35 on 9x9 which was a lucky game. He stopped sending scores for several months.

In July he suddenly improved and started sending scores every few days:

July 13 200 (24x30)
July 23 75 (16x16)
July 23 140 (16x30)
July 24 31 (8x8) 8 second pause at start
July 28 39 (9x9) 7 second pause at start
July 29 80 (16x16)
July 30 36 (8x8) 8 second pause at start
Aug 6 85 (16x16)
Aug 8 90 (16x16) 4 second pause at start
Aug 11 215 (24x30)

When asked about the pauses at the start of games he replied, "My strategy for openings is to click-look-F2 as quick as I can, and when I find a large opening like those 2 boards I stop and figure out how many open spaces are left which helps me know what to guess and how to play the board.” He also claimed to play 5-8 hours a day.

On 17 August, name removed contacted Damien about these games, pointing out the huge odds of getting the large openings on the Beginner games. His rough calculations suggested the chance of these openings was less than 1 in 250000. Starting a game each second and hitting F2 would take about 70 hours to find one, and then you would have to make all the correct guesses. If Brandon was using a program to search for openings, the pauses would be from him being late to notice his program had stopped.

Examples of UPK screenshots. You can see "JitBit Macro Recorder" on the Taskbar.
Examples of UPK screenshots. You can see "JitBit Macro Recorder" on the Taskbar.

On 20 August it was discovered that Brandon was using a macro program called JitBit to look for openings and then using UPK. After finding a good opening, the macro would be customised to look for that opening again. Using screenshots from failed solving attempts, eventually the game could be solved. The 8x8, 9x9 and most recent 16x16 games were definitely cheated. There were several pictures (now deleted) at his Photobucket page (s666.photobucket.com/albums/vv25/BJSwimmer) showing screenshots of games played with UPK and the JitBit program open. One picture shows a blast of his 36 mines on 8x8. Another picture shows examples of lost 9x9 games (the blasted game in the top corner is identical to his 39 mines on 9x9 for example).

The same day a post by Brandon (as "Axecutioner") was found at www.hackforums.net dated 17 July: "Long story short, I need the source code of a specific minesweeper game. I've tried decompiling it and that didn't work. I'm very basic as far as computer knowledge so I'd appreciate it is it could be done for me or at least give me very detailed steps on how to do it. All I need is the source code. Thanks!" He provided a link to MSX. This suggests his scores up to this date were legitimate, as he was still looking for a hack.

On 25 August there was a chat on IRC. Damien Moore, name removed, Daniel Marden and Brandon were present amongst others. Brandon was challeged to a "26 mines on 8x8" competition using Arbiter. Unknown to Brandon the others were not playing (except Damien who fluked a 14 second win in less than 2 minutes but kept the result quiet while waiting for Brandon). Brandon’s game showed a small pause after a large opening (consistent with the JitBit macro) and then several strange guesses (consistent with lack of skill or UPK). Brandon also "forgot" to use Arbiter and submitted the video using MSX.

Two days later, name removed challenged him to another Arbiter density competition. Brandon emailed Damien afterwards, "Long story short, we did a density tournament in the IRC chat today on 8x8 and I could not for the life of me get 25 mines using Arbiter. It took me like an hour and I still don't know why it seemed so hard." He said all future records would be made on Arbiter.

More evidence of cheating was found the next day. Back in November 2009, Brandon (as "Silver") had posted in a forum at s6.invisionfree.com that he was hacking Minesweeper X to put mines where he wanted, but was having problems hacking the mine counter and the timer. Someone pointed out it was easier to use Clone “UPK Mode” or to use MSX version 0.34 to do the same thing. Here are some of his posts put together:

"I tried hacking into Minesweeper X, and this is what I have so far" --- "Yeah, I just need to fix it to look like a finished minesweeper board. Then it'll be great." --- "Trust me, hacking into a Minesweeper X game is harder than I thought it would be. I'm getting closer now...” --- "Yeah, I now have the ones, twos, and threes working, but it puts eights everywhere else. Getting the mines in place were hard since I had to learn the code to set it up right. This part is just trial and error...”

He had also posted in April 2011 at sparkworkz.com (as “Axecutioner”) that he was hacking MSX using the command prompt. The goal was to manipulate mine locations but get the numbers, counter and timer to display properly. Here is his post on 8 Apr 2011 from www.sparkworkz.com/forums/viewtopic.php?t=7101:

"I got bored earlier today, and thought about making a Sparkworkz logo from Minesweeper. It started with getting the .exe file to show it's functions and calculations in the command prompt. From there, I just had to edit the mine locations from the partially random setup previously in the coding to the set coordinates of the letters I wanted. But then when I went to finish the board normally, it wouldn't recognize any of the numbers, as if there were no mines at all, and clicking a mine wouldn't do anything. That, and the timer wouldn't start and the mine counter went into the negatives. So from there, I had to reprogram the numbers to show the number of mines bordering them, get the mine counter to count normally (which was much harder than expected) and the clock to count right. I couldn't get the clock working though, my best result was it going up by a second each time I clicked a square. So I just stuck it at lucky 777 and kept it there. Also, clicking a mine didn't do anything. All in all, a nice waste of time, and I'm happy with the result."

Damien wrote an investigative report on 28 August listing the evidence but did not publish it. The report became this article.

In one more strange incidence, on 28 September Brandon signed into IRC and also under a fake name (Tom Sarango) and thanked himself (Brandon) for teaching himself (Tom) how to play minesweeper. (Brandon was not aware that IRC showed the IP address, and "both" players had signed in from the same computer).

[edit] Conclusion

It was clear his 2009 pictures were fake, and they were never allowed on any ranking. His 2011 games were allowed on the ranking but removed in October 2011 after Brandon was allowed time to defend himself. His March and April records are probably real, but his July and August records are all (or mostly) fake. Because of a consistent history of faking scores, none of his scores are now allowed on the rankings. His last post in the Guestbook (26 Oct 2011) states that he is trying to break his records but finding it difficult.

[edit] Brandon Explains

Brandon wrote the site on 3 Nov 2011 and explained everything. The 31 (8x8), 36 (8x8), 39 (9x9) and 90 (16x16) are fake but the other scores are real. The sparkworkz photos were him trying to be cool in a hacker forum, but he actually edited the pictures with Photoshop and did not hack MSX. JitBit was used to start games every 5ms but it was unable to find repeated boards on 16x30 and 24x30 so he did not ever cheat on those levels. In his words:

I didn't hack MSX, the sparkworkz posts were just simple photoshops and me trying to make it look like a real hack. I never did get the source code of MSX and if I did, I could have coded in a like that worked like xyzzy but still let me record video. The 8x8, 9x9, and 16x16 were done using the macro, and upk after finding multiple of the same board. The 16x30 and 24x30 were done the correct honest way. I did try 16x30 on the macro but there were too many different boards to cycle through, even at 5ms per reset that I was running at. I never tried to cheat 24x30.

The only thing I can say about the article is get rid of the sparkworkz posts - those were simple photoshops and me trying to act like it was a real hack. I was never able to actually hack MSX, or get the source code for that matter. To prevent against macros, I'd say, don't have repeating boards. Make mine placement totally 100% random. I could get 3 or 4 of the same board in a day on 8x8 and 9x9 with a refined macro program.

The only thing I ask is that you accept my 24x30 record and put it back up. I spent many hours trying to get it, with no cheating in any way. I know you have no reason to believe me at this point, but hey, it's worth a shot. If there's anything else you want to know, I'll be happy to help. I'd also appreciate it if this email was posted with the article.

I hope your leaderboards are running well for many more years to come. Thanks for good time while it lasted.

[edit] Videos and Pics

  • Brandon Stitt - All of Brandon's videos and pictures and the evidence of cheating.
Views
Personal tools