Why is RQP Time*(Time+1)/BBBV?

Want to calculate something or share your results?
Post Reply
sugarsugarsugar
Posts: 8
Joined: Thu Aug 24, 2017 1:16 pm

Why is RQP Time*(Time+1)/BBBV?

Post by sugarsugarsugar » Thu Jan 11, 2018 3:47 am

What does +1 term mean?
Why is not RQP sqrt(Time*(Time+1)/BBBV), (Time+0.5)/sqrt(BBBV) or Time/sqrt(BBBV)?
Indexes which have same dimension (or more strictly, linear relation or proportional) to time look more convenience for me.

pauv
Posts: 82
Joined: Wed Nov 20, 2013 6:06 am

Re: Why is RQP Time*(Time+1)/BBBV?

Post by pauv » Thu Jan 11, 2018 9:51 am

RQP is old-fashioned, you don't need to care about it. The following is history:

Old minesweeper clone sets time=real time(rt)+1 because it seems the timer on the original windows minesweeper works like that (if you solved the board in 4.04s(rt) then you will see 5s on the timer). But in fact rt is more proper and accurate to compare players' level with each other.

RQP=[old time]/[bbbv/s]=(rt+1)/(bbbv/rt)=rt*(rt+1)/bbbv, some people noticed that bbbv/s is generally higher on large board than small board, so they made this indicator to compare "real skill" of players, not only depend on time records. But this model is too rough.

sugarsugarsugar
Posts: 8
Joined: Thu Aug 24, 2017 1:16 pm

Re: Why is RQP Time*(Time+1)/BBBV?

Post by sugarsugarsugar » Thu Jan 11, 2018 10:21 am

Thank you for reply.
It is answer for my first question.
But second is not solved yet, because QG is proportional to Time^1.7 and STNB is like so (I don't know this index well).

Although deviating from the main topic, additional question. What is modern index instead of RQP?

pauv
Posts: 82
Joined: Wed Nov 20, 2013 6:06 am

Re: Why is RQP Time*(Time+1)/BBBV?

Post by pauv » Fri Jan 12, 2018 2:14 am

Time/bbbv, it's nonsense, since time/bbbv=1/(bbbv/time)=1/(bbbv/s), you will know nothing new about your game.

QG is instead. RQP contains time^2 so it's still a time-prior benchmark. Time^1.7 is to best eliminate the influence of 3bv on the ratio of time and bbbv/s (according to gacord who chatted with me yesterday)

Where did you know stnb? Are you from Japan? :? I don't know this either.

sugarsugarsugar
Posts: 8
Joined: Thu Aug 24, 2017 1:16 pm

Re: Why is RQP Time*(Time+1)/BBBV?

Post by sugarsugarsugar » Fri Jan 12, 2018 3:33 am

Sorry, I mean why QG is not Time/(BBBV^(1/1.7)). This is equal to QG^(1/1.7).

I'm Japanese.(I don't join ranking yet)
I heard stnb from yowapon (Mao Igarashi).
I understand STNB=Constant/QG, where constant term is 47.3 for Beg, 153.7 for Int, 435 for Exp.
I also heard this index can apply to lost game, but detail is indistinct.

pauv
Posts: 82
Joined: Wed Nov 20, 2013 6:06 am

Re: Why is RQP Time*(Time+1)/BBBV?

Post by pauv » Sun Jan 14, 2018 2:30 am

gacord said he would answer your question tomorrow(busy weekend), he is the author of QG and stnb ;)

the only thing I know about stnb is, stnb=Shi Ti Niu Bi=the level of your blast game

gacord
Posts: 86
Joined: Thu May 17, 2012 5:14 am

Re: Why is RQP Time*(Time+1)/BBBV?

Post by gacord » Mon Jan 15, 2018 3:33 am

QG|(TIMEESTREAL^1.7)/BBBV{3}
STNB|(87.420*(mode^2)-155.829*mode+115.708)/((TIMEESTREAL^1.7)/BBBV/((BBBVDONE/BBBV)^0.5)){2}

So STNB = cons/QG * CompleteRate^0.5

:D

sugarsugarsugar
Posts: 8
Joined: Thu Aug 24, 2017 1:16 pm

Re: Why is RQP Time*(Time+1)/BBBV?

Post by sugarsugarsugar » Mon Jan 22, 2018 7:41 am

Thank you for telling me the definition of STNB.

Return to main topic, I ask again why QG is not Time/(BBBV^(1/1.7)).

gacord
Posts: 86
Joined: Thu May 17, 2012 5:14 am

Re: Why is RQP Time*(Time+1)/BBBV?

Post by gacord » Fri Feb 09, 2018 7:39 am

I think time/(3bv^(1/1.7)) is not enough to describe the difference between 59s and 44s on certain 3bv.
When it is time^1.7/3bv the difference is 59^1.7/143/(44^1.7/143)=1.64656496253.
When it is time/(3bv^(1/1.7)) the difference is 59/(143^(1/1.7))/(44/(143^(1/1.7)))=1.34090909091.
In STNB I set 44 second on 143 bbbv completed as 100 point and 59 second will be 100/1.647=60.716point.
If QG=time/(3bv^(1/1.7)) and I also set 44 on 143 as 100 point and 59 second would be 100/1.341=74.571point.
At this system the 59s level player got much higher points than time^1.7 system.

sugarsugarsugar
Posts: 8
Joined: Thu Aug 24, 2017 1:16 pm

Re: Why is RQP Time*(Time+1)/BBBV?

Post by sugarsugarsugar » Sat Feb 10, 2018 4:29 pm

Is it a problem?
The difference between two definition is just a scale.
like this:
very good ~70 point~ good ~90 point~ ordinary ~110 point~ not good ~130 point~ bad for STNB based on time^1.7/3bv
very good ~80 point~ good ~95 point~ ordinary ~105 point~ not good ~120 point~ bad for STNB based on time/(3bv^(1/1.7))
(This is just example, it means 70 point for the former is not actually correspondence to 80 point for the later, and so on.)

Set time/(3bv^(1/1.7)) to QG, set STNB based on that, and I solve a board with 105 point.
Then, I can understand it is about 5% slower than standard in a moment. (Because when x<<1, 1/(1+x)~1-x)
Loss of compatibility with old data may be not nice, but I think this definition looks still useful.
If you think this is not convenience in particular, it's different sense of values.

gacord
Posts: 86
Joined: Thu May 17, 2012 5:14 am

Re: Why is RQP Time*(Time+1)/BBBV?

Post by gacord » Thu Mar 22, 2018 9:05 am

:D Definitely correct
It seems you are thinking a lot about this skill based index.
So I'll make a small AD here to welcome players over the world to chat with the most active minesweeper chatting group.
How to join? Send an e-mail to me : qxxqlls@163.com :D :D :D :D

Post Reply