PDA

View Full Version : how is this possible


derek
02-13-2014, 11:39 PM
I have recently completed 3 4x7 puzzles in what i thought was a very fast time ie each one took around about 5 mins. After pressing submit solution i found that the record time in each case is less than one minute for each puzzle!!!!! How is this possible it takes at least two minutes to sort through the clues. Any suggestions???

gypsywyz
02-14-2014, 12:21 AM
I also just completed a 4x7 and the record time was 51 seconds. I just don't believe that is possible. I am calling foul and challenging the admin of this website to prove to me that it can be done that quickly. Bogus scores like that skew the bell curve and set unfair standards.:mad:

admin
02-14-2014, 08:01 AM
I've reset about 400 "record times" just now that were past the threshold of what I would consider to be a reasonably achievable solve time, and have put some safeguards in place this morning that should prevent these from registering as legitimate record times in the future.

Thanks!

Alethea
02-14-2014, 09:25 AM
I'm so glad you've done this! I didn't feel it was for me to say what other people are capable of achieving as far as record times go, but under a minute for a 4x7 seems improbable to me as well.

gypsywyz
02-14-2014, 06:29 PM
Thanks, Admin! Your efforts to keep this a fair and equitable game are much appreciated. As a member of the NPL, I love a great puzzle challenge and am more encouraged now :)

Anthony33
02-15-2014, 09:07 PM
What's the NPL?

Anachronismatic
02-15-2014, 09:35 PM
I'd be interested in knowing, from those who have written scripts to solve these puzzles quickly (not that much of a challenge for a programmer), what's the point of "solving" hundreds of puzzles a month like this? Those of us who are interested in fastest times find it rather obvious when someone suddenly jumps from an average of 600 seconds to below 100 seconds. Definitely busted. The rest will never recognize your "expertise" either way.

To admin, I would suggest random encryptions in the display of named attributes and common words in clue descriptions, invisible to display but not to cut-and-paste, that would tire out the script kiddies. Or perhaps generating the clues as an image rather than as text.

Hopefully, the new time limits are reasonable. The old limit was 100 seconds, and that didn't work well. Wherever they're set, I don't think it's the right solution to the problem.

mixt
02-19-2014, 06:51 PM
I'd be interested in knowing, from those who have written scripts to solve these puzzles quickly (not that much of a challenge for a programmer), what's the point of "solving" hundreds of puzzles a month like this? Those of us who are interested in fastest times find it rather obvious when someone suddenly jumps from an average of 600 seconds to below 100 seconds. Definitely busted. The rest will never recognize your "expertise" either way.

To admin, I would suggest random encryptions in the display of named attributes and common words in clue descriptions, invisible to display but not to cut-and-paste, that would tire out the script kiddies. Or perhaps generating the clues as an image rather than as text.

Hopefully, the new time limits are reasonable. The old limit was 100 seconds, and that didn't work well. Wherever they're set, I don't think it's the right solution to the problem.

Personally I don't like inflating the leaderboards with programs I write for anything. I opt out whenever possible and throw in purposeful ineffeciency when I can't (such as here). That does inflate the bottom of the board and will afftect things like average times, but people tend not to care about feeling higher on the scale than they actually are (plus there are always more people pulling it the other way). I do it just because I like programming and like to find ways to make computers do things that are challenging or I don't like to do (I've made so many scrpits to do math homework for me).

Though my voice on the subject is somewhat void because I've never actually scripted these puzzles and many don't take my care to preserve leadrboards.

One option that I've seen work somewhat well is to make people opt in to the leader board. It has the draw back of removing a lot of valid data, but it also removes false data from the people that simply don't think about the leaderboard and allows for more direct action against people that are actively damaging it. Or you could even go as far as to make a "coder's leage" or some such alternate board where anything goes and people can compare script's results to each other. But such a board quickly becomes useless.

Alternatively, at the cost of more processing, you could replace the static bell-curve images with an actual graph so people can see the anomalies themselves and remove them mentally.

BrightSky9
02-20-2014, 03:43 PM
What's the NPL?

http://en.wikipedia.org/wiki/National_Puzzlers%27_League
via Google