[PROPOSAL]Remove Beginner From Time Rankings

Suggest ideas for the World Ranking

[PROPOSAL]Remove Beginner From Time Rankings

Postby EWQMinesweeper » Fri Mar 20, 2015 11:12 am

The rules for which beginner records were eligible for rankings have changed a lot over the years. First the timer started at random between 0.00 and 0.99s, then 9x9 boards were allowed for a short time, then boards had to have 2 3BV or more, then we introduced rt and a few years ago we added the requirement that all records had to have at least 1 controlled and intentional click, ruling out all lucky solves.

Although there is a rules page above the rankings, I am not sure how many people are aware of it and have read and understood it.

While this last few changes in rules did at some purpose to practicing beginner, there is something that has been bothering me ever since my investigation of the then top100 players' beginner records: macros.

As we have seen in the case of Brandon Stitt, macros can be used on smaller boards to gain an advantage in density mode. On beginner lever and in normal time mode however, macros can be used to create solves that are impossible to tell apart from real solves.

The process to create such a macro is very simple. Record two clicks on a beginner board and the mouse path in between. Make sure to take a small break before second click and have the path between the clicks not be a straight line. When editing the macro, cut out frames until the actual solve is around 0.3s. The macro itself doesn't need to be much longer. Have the macro hit F2 in its first frame, make sure that mouseDown and mouseUp events for the first click ar in the same frame (since the recording of the replay starts on mouseUp of the first click) and set mouseDown for the second click at 0.1s of the solve (also useful in normal solving, just saying). If you want to save some more time, make sure that the first click is in either (1,1), (1,6), (6,1) or (6,6) on the board and make the macro check if this hits an opening. If it is, the opening will be at least 9 squares big, thus eliminating a lot of boards very early and saving time. The second click could be at (0,7) or (1,7) or similar.

One could also have the macro check whether the board has a 3BV of 2 (if 11 unopened squares are left after first click) and then let it choose a prerecorded path to a random unopened square.




Pro:
+beginner times are usually a neglible part of the sum of times - changes in ranking position would be small.
+removes an easy opportunity to cheat - see above
+current rules are already difficult to execute and only few people know about 2 3BV and non-lucky requirements

Con:
-people with sub1 or low 1 times have put a lot of effort into playing beginner
-beginner is a standard level of windows minesweeper
„Das perlt jetzt aber richtig über, ma sagn. Mach ma' noch'n Bier! Wie heißt das? Biddä! Bidddää! Biddddäää! Reiner Weltladen!“
EWQMinesweeper
 
Posts: 410
Joined: Sun Nov 30, 2008 11:50 pm

Re: [PROPOSAL]Remove Beginner From Time Rankings

Postby pauv » Fri Mar 20, 2015 10:54 pm

Seems the next of viewtopic.php?f=12&t=137, but you raise the evidence that "1 click on purpose" rule can be faked in this post.

So I will do a little conclusion of all proposal of that post below. Edit a little for simplification.
--------------------------------------------------------------
proposal:
by qqwref:
1) Keep it as is.
2) Remove beg time entirely from the world rank.
3) Remove lucky games only, allow so-called "non-lucky" 1s games. This involves checking all 1s games; borderline ones would have to be voted on.
4) Set the 3bv limit higher, like 3 or even 4, so that completely lucky games are very rare.
5) Decide that there is no skill difference between 0.99 and 0.10, and simply record any 1s game as "1" for the purposes of world ranking.
by EWQ:
6) no lucky beg game played after eg march 2011 is allowed on the rankings, players who got a lucky beg record before that are asked whether they have another 1s beg game which is non-lucky. if they want, and only if they want, their time is replaced with their best non-lucky game
7) allow 1 3bv boards and count all instant wins and lucky beg records as 1.00 RT.
by Tommy:
5')rounded to 0.40 or so (or a 0.31, for that matter :))
by KamilSaper:
8) Create historical ranking (old planet-minesweeper style), with beg+int+exp scores sum.
by arjadre:
9) Keep beginner in the world ranking (for show), but don't use it to computer players' ranks.
by thefinerminer:
7')use the "1 click on purpose" rule

principle:
by qqwref:
1."Historical" times on int and exp can be legitimately beaten
by KamilSaper:
2.World ranking would be based mostly on skill (Int and Exp times sum)
by thefinerminer:
3.players realised they could beat their integer times - this encouraged them to replace pictures with videos.
by arjadre:
4.the best scores should go to the best players
----------------------------------------------------
Now you pick up the 2) because the cheating method you point out. I'd like to express what I prefer directly, that is,

4') set the 3bv limit higher, like 10 or even 11, so that completely lucky games are very rare. You can rule out cheaters, too.

The main reason:
1. Mo proved 5 random clicks in BEG possible, the only other game which involved even higher clicks is gacord's 8bv game (0.82s).
2. at least 50% 3bv must be solved on purpose.
3. Fast beg time made on 10bv or more need extremely skill. I have noticed Damien, Pavel, Kamil have sub2 rt games on 10+ 3bv.
4. You can keep the same rule with 3bvs ranking so it's easy to explain to new sweepers.

gacord and I are studying all 1s on saolei.net, there're a lot of cheaters and I am tired to see all these through (approaching 1000 games!), so it's will be good to ban the fast games :P BTW I have marked "Specially Mention" on Mo's 0.95s on 10 in my list months ago, since I keep the same view with Tommy: viewtopic.php?f=12&t=160.
Jason Kriegel's 1.39s on 10 has 2 clicks on purpose and should be save.

possibly con of my proposal: how to deal with old exit sweepers? ------rounded to 3.00 :D , or max{10/(beg 3bvs record), 3.00} like Jason's example.(see the principle part)
-----------------------------------------------
Maybe another topic: there're pages of beg cheaters on Minesweeper Classic ranking, so my 1.71 ranked #1956 when int 18.22 ranked #135 and exp 68.59 ranked #290.
pauv
 
Posts: 78
Joined: Wed Nov 20, 2013 6:06 am

Re: [PROPOSAL]Remove Beginner From Time Rankings

Postby EWQMinesweeper » Sat Mar 21, 2015 12:35 am

Thank you for your input, pauv, especially for compiling these two lists.

Let me add a 5th and 6th principle: 5. Prevent cheating or make it as difficult as possible. 6. Arbitrary rules should be avoided and in unavoidable handled with great care.

Principle 1, 3 and 4 are pretty much obvious and don't have any influence on discussions.
Principle 2 is also obvious, since beginner times make up only a fraction of the total sum. It could be argued whether this fraction is so small that it makes the level neglible for time rankings.

Increasing the 3BV limit could solve the problem of macros but as mentioned in 6. it won't be easy to find the right lower limit for 3BV. 3 and 4 seem would already increase the average required effort to fake a solve with macros by a great factor, but we would have to further discuss how much effort needed to fake a game would be enough to deter a cheater. would a cheater spend 500 hours trying to fake a replay? 5000? a fully year?


You suggest 10 or 11 3BV. This would exclude a substantial amount of boards (~10%). If we implement such a high limit, the clones would need to be updated or else much of the time playing beginner would be wasted on boards uneligible for records. You also mention 2 games as examples. If I am not mistaken the non-lucky requirement in the current rules already exclude them from rankings.

I do not understand your reason 2). Is the 1 intentional click requirement not enough?

Reason 3: Increasing the 3BV limit would mean that most beginner records would no longer meet the rules. For most sweepers there is no way to determine their best time on a board that meets the increased 3BV limit. Even though it would increase the challenge to get a good record, such a high 3BV limit would have a too drastic impact on the rankings. Or do you have a nice suggestion for how to deal with cases where no suitable beginner record can be found?

Reason 4: Do they need to be consistent? Last time i checked 3BV/s rankings required a sum of at least 10. Thus players on that list are not totally new.

I also really don't like your proposal of simply rounding up retired sweepers' times to an arbitrary value. There has to be something better.


Also, could you please give some reasons why you think we should actually kepp beginner for time rankings?
„Das perlt jetzt aber richtig über, ma sagn. Mach ma' noch'n Bier! Wie heißt das? Biddä! Bidddää! Biddddäää! Reiner Weltladen!“
EWQMinesweeper
 
Posts: 410
Joined: Sun Nov 30, 2008 11:50 pm

Re: [PROPOSAL]Remove Beginner From Time Rankings

Postby pauv » Sat Mar 21, 2015 9:03 am

EWQMinesweeper wrote:Increasing the 3BV limit could solve the problem of macros but as mentioned in 6. it won't be easy to find the right lower limit for 3BV. 3 and 4 seem would already increase the average required effort to fake a solve with macros by a great factor, but we would have to further discuss how much effort needed to fake a game would be enough to deter a cheater. would a cheater spend 500 hours trying to fake a replay? 5000? a fully year?

You suggest 10 or 11 3BV. This would exclude a substantial amount of boards (~10%). If we implement such a high limit, the clones would need to be updated or else much of the time playing beginner would be wasted on boards uneligible for records. You also mention 2 games as examples. If I am not mistaken the non-lucky requirement in the current rules already exclude them from rankings.

I do not understand your reason 2). Is the 1 intentional click requirement not enough?

OK I will do some further explanation. reason 1+ reason 2 is why I suggest at least 10 for a 3bv limit, it's easy calculation(5/50%=10) ;) So it's not arbitrary like your 6) principle. Skill must be more than luck, so 50%(10 as limit) or even higher. I think it's not likely many persons will do the same like Mo(I mean the 0.95s), but if occasionally happened we can have a vote :)
I mentioned the 3bv=8 game to show the ONLY out of bound (3bv=5) instance, that's not likely to happen twice. So the number 5 is still reasonable.
This rule covers(be stronger than) the "1 click on purpose" rule, and you have proved "1 click on purpose" not enough.
A game need at least 2 click on purpose is easy to observe the mouse cursor trace. So no cheater will survive :lol:
No update needed. There're some players want low 3bv games.

EWQMinesweeper wrote:Reason 3: Increasing the 3BV limit would mean that most beginner records would no longer meet the rules. For most sweepers there is no way to determine their best time on a board that meets the increased 3BV limit. Even though it would increase the challenge to get a good record, such a high 3BV limit would have a too drastic impact on the rankings. Or do you have a nice suggestion for how to deal with cases where no suitable beginner record can be found?

Believe me, this rule is strong enough and there won't be new mess, so there will be no increasing topic. Challenge? Yes it will take beg more fun! Remember why people don't care beg time is because of no skill needing. People who is now not aware of msinfo won't argue, for who still care their score...why not take several minutes playing just a little beg thing? Follow the principle3(influence here :P ), it's really easy to beat. Since the most share the same beg score(3.00) now, the ranking will appear the same like your proposal will do, and it's even less drastic than your 2).

EWQMinesweeper wrote:Reason 4: Do they need to be consistent? Last time i checked 3BV/s rankings required a sum of at least 10. Thus players on that list are not totally new.

They don't need (as a must), but it seems better. Int & Exp cases are consistent. There will be new players join msinfo. And if saolei.net shares its database(in the future) with msinfo, we need to explain the rules, so consistency is needed in that case.

EWQMinesweeper wrote:I also really don't like your proposal of simply rounding up retired sweepers' times to an arbitrary value. There has to be something better.

In old days msinfo rounded retired sweepers' time to 1,00 RT. It's not so arbitrary, but we can do the same now. Sup10&sub2 is not so easy, so I consider 3.00 as a rough cap. To improve the raw cap, I suggest to count 10/(beg 3bvs record) for weak players(I'm talking about old records, new player don't need to know this) and mentioned Jason as an example. I mean, playing for time record is just the same with playing for highest 3bvs on every given 3bv (of the game), and vise versa.

EWQMinesweeper wrote:Also, could you please give some reasons why you think we should actually kepp beginner for time rankings?

1. It contains the fact of skill. And it can be divided from most of the fact of luck this way.
2. It has high score on the original microsoft minesweeper.
3. Every new player begins with beg level, they want to know what's the world class level on beg(see the first half of 9) proposal).
Although further investigation is needed, gacord tried to use this limit on the data of saolei.net for analysing, and the result perfectly meet the principle 4: the best scores should go to the best players.
pauv
 
Posts: 78
Joined: Wed Nov 20, 2013 6:06 am

Re: [PROPOSAL]Remove Beginner From Time Rankings

Postby gacord » Mon Mar 23, 2015 5:02 am

One new idea:
3BV >= 10 and Efficient clicks >= 10
This rule just cut down almost all lucky games.
gacord
 
Posts: 70
Joined: Thu May 17, 2012 5:14 am

Re: [PROPOSAL]Remove Beginner From Time Rankings

Postby gacord » Mon Mar 23, 2015 5:07 am

There are many fake beg scores.
However most of them are below 10 3bv.Almost all of them have less than 10 efficient clicks.
gacord
 
Posts: 70
Joined: Thu May 17, 2012 5:14 am

Re: [PROPOSAL]Remove Beginner From Time Rankings

Postby Tommy » Mon Apr 06, 2015 10:12 pm

Well, as someone with (in the meantime) a non-lucky 0.51 who'd lose a rank that way... I'm with EWQMinesweeper, and more comfortably so now.

So: Kill beginner with fire. :twisted:

As for implementing it, I like arjadre's idea (keeping it on the ranking for show, but not using it to compute the ranking) most at the moment. There's nothing wrong with a little nostalgia if you don't let it impact how fair the ranking is.
Don't anthropomorphize computers - they don't like it.
User avatar
Tommy
 
Posts: 223
Joined: Mon Dec 01, 2008 9:22 pm
Location: Vienna

Re: [PROPOSAL]Remove Beginner From Time Rankings

Postby EWQMinesweeper » Thu Apr 09, 2015 1:52 pm

[quote="Tommy"arjadre's idea (keeping it on the ranking for show, but not using it to compute the ranking)[/quote]

this! so very much this!
„Das perlt jetzt aber richtig über, ma sagn. Mach ma' noch'n Bier! Wie heißt das? Biddä! Bidddää! Biddddäää! Reiner Weltladen!“
EWQMinesweeper
 
Posts: 410
Joined: Sun Nov 30, 2008 11:50 pm

Re: [PROPOSAL]Remove Beginner From Time Rankings

Postby pauv » Fri Apr 10, 2015 7:17 am

1. Fast score should be banned.
2. If there's 3bv limit, it should be 10.
3. Beg score should keep for show.
This is what I can conclude. I will stay fine if I drop only 1 place on the ranking like Tommy, LOL
----------------------------------------------------
But there would be another problem if arjadre's idea was accepted: what if new players upload the beg score for show? There will still be cheaters who bother the admins.
Every year tons of beg videos are poured into saolei.net :(
pauv
 
Posts: 78
Joined: Wed Nov 20, 2013 6:06 am


Return to Ideas

Who is online

Users browsing this forum: No registered users and 1 guest

cron