PDA

View Full Version : Possible Problem or Issue - 31 Sec Solve time?


mcgalas
10-09-2008, 07:07 PM
I just initialized a new puzzle and saw the record for it is 31 seconds (derek). I am guessing this is a bug so figured I would share.

I've solved a puzzle in something like 120 seconds... but there really is no valid possible way someone could do it in 31 seconds, is there?

The Puzzle in Question has the following facts before the puzzle:
Wedding Years = 1921, 1941, 1944, 1971, 1974
First Names = Alyssa, Hayley, Linda, Meredith, Nicolas
Wood, Boxwood, Cedar, Chestnut, Maple, Oak
Pets = Dog, Frog, Hamster, Parrot, Pirhana

Hope that is enough info.

Amanda
10-10-2008, 08:41 AM
I found another record time of 38 seconds (derek) --no, we're not picking on you:

Wedding Years
1921, 1941, 1943, 1973, 1989

First Names
Alexa, Arianna, Claudia, Georgia, Ximena

Cars
Honda, Jaguar, Kia, Oldsmobile, Volkswagen

Last Names
Alston, Conner, Duncan, French, Houston

mikiboop
10-12-2008, 08:35 PM
The answers require at least 30 mouse clicks: two for each of the correct answers. That is one click per second if you only mark the positive answers. Personally, I do the puzzles just for fun. I know that there are patterns to the clues and answers, but that would spoil my fun and defeat the purpose of puzzles.

sarah.paggen
12-02-2008, 04:14 AM
I think I can give you an answer on this. I almost always rate the puzzles and twice I have been unable to rate a puzzle that I just finished because I rated it before. I'm guessing that's because I played it before! I know from reading the posts that some people make separate spreadsheets when playing. My guess is that they keep their old answers and compared the columns on the first page to previously completed puzzles and all they had to do was fill in the green marks to be done!

I know some people might think this is cheating, but until all the bugs are worked out of the site, I call it pretty smart!

bowl231
12-02-2008, 04:39 PM
I have seen a few which were below a valid amount
the latest was 29 seconds (hazeleyes262000)
this is below a 1/2 minute which is pretty much impossible
these are either a big bug or cheating.

Honestly anything under at least one minute but in my opinion
under two minutes (120 seconds) is not really valid.
I am hoping a bug but there are way too many people out there
that need to feel special so they do cheat All those should be
deleted as the Record

I really love this game and try to ignore the scores

sarah.paggen
12-03-2008, 06:01 AM
I guess I just figure eventually we won't get the same puzzle a second time. Then you'd have to play puzzles either as a guest or under another login to see the same one twice.

That's cheating, but someone will always cheat to try to make themselves look better, and they will almost always be caught by someone. The competition is fun, but I think just playing is more fun!

brandx
12-04-2008, 10:26 AM
sarah is right! people are keeping a record and replaying them if the puzzles appear again. must be leading very fruitful and productive life.

Laura
12-07-2008, 05:13 AM
Well, I just solved one that I had earlier today and it still took me almost 10 minutes. :o

I've also been astounded by the super quick solves because I find solving these is a heck of a lot slower than solving cryptograms. Of course, I have walked away to do something else and then forgot I was in the middle of a puzzle, only to come back the next morning and find it still sitting there. However, after 2 months of doing these my average has come down from about 30 minutes to usually about 10.

But I'm pretty impressed by some of the scores here. Even though it's hard to believe the ones under 1 minute.

mcgalas
12-12-2008, 09:32 PM
Frankly, I don't care that much if people are cheating (er, looking up the answers) because its just their time they are wasting... but we might as well make this as good and fair as possibles since it is so popular.

One possible solution to prevent cheating is to not give out the options and categories to start? That way if they do still want to look it up, they must take the time to do so during their actual game (which I'm sure they can still do and still get fairly fast times, but at least it would help slow them down).

I would be curious to see how many people actually read through them and think having them really helps before you start vs spending 5 sec and looking at the grid (which then makes you remember where they are too)?

Just some thoughts to try and help improve the site. Thanks.

flutefish
12-17-2008, 02:43 AM
I will say that I have solved a puzzle in less than 2 minutes (120 sec) before, and under 200 seconds very often. And I have never used a spreadsheet or saved my answers. Don't quite know how Derek gets his fast times, but under 2 minutes is doable.

darsinewman
12-17-2008, 05:20 PM
I agree that under 2 minutes is doable, but under one minute seems to be impossible without cheating. Do people really feel better about themselves for getting good scores while cheating? Oh, well.:rolleyes:

Keli42
02-24-2009, 04:26 PM
I am assuming this is still a bug or hasn't been fixed?

Congratulations!
You solved the puzzle correctly in 4 minutes, 52 seconds. That's worth 439 points!

This Puzzle's Statistics
Success Rate: 57.8%
Average Time: 710.1 seconds
Record Time: 28 seconds (derek)
Your Time: 292 seconds

I don't know how's it's possible to solve a 5x5 puzzle in 28 seconds. Even if it's printed out first. Maybe if there were a way to pause and then the screen dissapear so you can't cheat? Seems to kill anyone else's hopes for beating the record (not that my 4.5 mins is going to do that!)

TAPSfan
02-26-2009, 02:46 PM
I'm glad to see that I'm not the only person wondering about the extremely fast solve times. I can't remember the puzzle, or the player, but someone apparently solved a 4x5 grid puzzle in 14 seconds!! That seems highly unlikely!! You can't even read through the clues in 14 seconds!! :confused:

Macabre Sunset
02-28-2009, 11:01 PM
Yeah, someone has probably reverse-engineered the hash function used to check solution submissions.

I doubt anyone has written down solutions and could access them in less than a minute.

I'm on the verge of breaking three minutes myself. Not sure if two is possible for me. I click slowly because of the bug where it erases your form if you click in the middle of a graphic using Firefox.

bowl231
03-01-2009, 03:05 AM
No you are not alone wondering how this can happen. There are a at least 3
"cheaters" and I think any scores under 120 sec should be questioned and thrown
out totally. This one below should not even be credited
33 seconds (Dreamscape)

I still can't see how someone can finish a 5X5 in less than 2 minutes (120) anything
under 160 seconds is debateable to me on how they can read and answer clues.
Which was why I asked when the timer clock is actually activated and requested a
a clock be added to the puzzle so I could see how long it was taking me.

Another thing was to add to the puzzle completion statistics the number of
plays which were use to compute the average times, the success rate , max points
possible and average points. Just to see how many actually did that puzzle

I love the game and am very glad it is available

admin
03-03-2009, 12:22 AM
Hi everyone -

I'm working on some scripts that will help figure out how some of these super-fast scores are being generated. At this point I can't say for sure whether its a bug or someone purposely gaming the system, but I'm looking into it.

In the meantime, I'll be refactoring game statistics to take out of consideration any "suspect" games.

Thanks for playing!

admin
03-03-2009, 02:37 AM
Hi again -

Well, I've been crunching some numbers and have made some conclusions about this issue. There are definitely "bogus" times being counted as valid which should not be, and these will be removed from the system.

The tricky part is trying to determine what is a "valid" score and what isn't. I've run a few models based on half-a-million solved puzzles, plotting the total solve time against how frequently they occur in the system, and I came up with the following lop-sided bell-curve:

4

You'll notice there's a rather misshapen "bump" that proceeds an otherwise normal-looking bell-curve, starting at 11 seconds, peaking at around 32 seconds, and then dropping back to near-zero at 50-60 seconds. From 60-65 seconds onwards the graph looks like a normal bell-curve.

My conclusion from this is that all scores below 50-60 seconds are bogus, and I will be treating these as such when we refactor our statistics. If there is anyone here who can legitimately solve a puzzle in under 50-60 seconds, please contact me. Otherwise, I will stick with my assumption and discard these scores.

As far as how the bogus times are being achieved... well, I won't put that information here, but suffice it to say that after crunching some data over the past few hours an obvious trend emerged and I am 99.9% sure I know how its been done. There are measures that can help reduce the effectiveness of this tactic, and I will be employing them over the next several weeks.

Anyway, thanks to everyone for bringing this to our attention. While it is next-to-impossible to absolutely prevent cheating on an online game of any sort, I do want to do whatever I can to minimize its impact, so that folks can compete here and know that they're getting a fair shake in the standings.

Cheers -

Stephen

darsinewman
03-03-2009, 11:50 AM
Stephen - I think your 50-60 second rule is good. I have solved several puzzles in under 120 seconds (especially the new 4x4s) and have never cheated. I consider myself to be a very fast solver, but it boggles my mind how any puzzle could legitimately be solved in less than 1 minute. I know there are some out there who are faster than I am, but it seems like even clicking the answers would take at least that long without even reading clues. Thanks for your efforts!

bowl231
03-03-2009, 12:10 PM
Stephen
Thank you for all your consideration, hard work and tremendous generous efforts. You do a great job and these games are great. Thanks for the great explanation and consideration.
Wow-After seeing over a million games being played - proves how wonderful this place is
Thanks again - have a great day

TAPSfan
03-03-2009, 02:33 PM
Stephen,

I would like to join the other players in thanking you for your time and efforts involving this matter. :)

Macabre Sunset
03-03-2009, 05:18 PM
In theory, I think I could get to around 80 seconds on the 4x4s and maybe 120 on the 4x5s. On puzzles where the lower grids aren't necessary and I instantly identify the right order to address the clues.

I still have to learn when filling red Xs isn't necessary, though. It's possible that's the key to lower scores.

I'd be interested in knowing how the people suspected of cheating are doing, now that the preview information is gone.

Laura
03-03-2009, 07:22 PM
Same here Stephen. I just solved a puzzle and wondered why the grid wasn't displayed prior to loading -- I've found it rather helpful because I create a table in Word to track my answers. However, I can see how displaying a grid beforehand can make cheating possible.

My times have gone down considerably now that I've been playing more often. I've even had some under 3 minutes. But under 20 seconds is just too astounding to be believed. You can't even read all the clues and fill in the grid that quickly.

I'd like to add my thanks to all the others -- you do a great job.

Laura

bowl231
03-07-2009, 12:40 PM
There are still some real low scores for 5x5
this is a "new" ID - Spirit Bear have not seen this person before
---------------------------------
This Puzzle's Statistics
Success Rate: 53%
Average Time: 656.6 seconds
Record Time: 34 seconds (spirit bear)
Your Time: 539 seconds

admin
03-07-2009, 05:46 PM
That will be the case for perhaps another week or two - we are gradually re-figuring statistics on all of our puzzles using the new criteria. To avoid heavy server loads while re-processing we are re-figuring 60 puzzles per hour. My guess it we will get through the whole of them in another 4-5 days or so, after which time the "false-records" should be all purged.

stongia
04-20-2009, 11:36 AM
Steps According to me for fast solution.
Once the puzzle come on the screen do print screen and come out of the puzzle screen it will stop the puzzle and put it in incomplete state.
From the print screen get the puzzle and solve it offline get ready with your answers and try to complete the puzzle and you only have to click correct entries.

And you will get decent fast time for the same.

fyre69
05-08-2009, 12:34 PM
I have seen patterns in the games. They are computer generated and only the lists of words change. So if someone has a great mind for patterns, or has written a program that can recognize the pattern after a few right answers it is possible to get those super fast solves. I do this for fun, there is no money involved, and is just for fun and bragging rights, I try to just have fun with the games and ignore the high scores. I will admit though, sometimes when I know I have solved it at a pretty fast time without using pattern recognition, it can be frustrating.

Happy gaming. Peace out,
Fyre69

bgates87
07-03-2009, 04:45 AM
I think that basically what is happening is that people have found a way to retrieve the clues and other information about their next puzzle without actually loading the puzzle. The puzzle you will play next is determined when you click "New Game," but the timer doesn't start until you click "Start Game," (I assume) so it's possible (I won't go into much detail here) that once you know the URL of the next puzzle you can use it to retrieve all of the information about the puzzle without starting the timer. You can then solve the puzzle at your leisure and once you have the solution you can allow the page to load. Then all you have to do is put in the answers which should take around the amount of time that most of these suspicious solves are taking.

The fundamental problem here is that the puzzles are pre-loaded. This is done as a convenience for the players so we can see information about the puzzle and so that when we're ready to start solving we don't have to wait as long. The pre-loading could be done away with, clicking "New Game" could throw us straight into a random puzzle, but the timer has to start when the page is loaded to avoid foil another much simpler method of cheating, and then we would have to wait for the entire puzzle to load while the timer was running which would probably add between 10-20 seconds to everyone's solve times.

Here's my solution to cheating: Do way with pre-loading the puzzles, just throw the player into a random puzzle as soon as they click "New Game." Start the timer as soon as they click "New Game" too. Now, to make up for the added time of waiting for the puzzle to load, have a second timer that starts when you click "New Game" and stops when the puzzle has fully loaded. When the player is done, subtract the value of the second timer from the total value of the first timer to get an accurate time.

At this point though I don't think that the cheating is bad enough that it's necessary to go this far.

Punkfunkmo
02-26-2010, 09:47 PM
I used to see the scores under 100 seconds and thought it was impossible or people cheating.

I have learned to do the puzzles using only the top rows and I have, on more than one occasion, completed puzzles in the area of 55-59 seconds.

There is always someone better out there, I'm going to assume that the 30 second scores are legit.:eek:

Punkfunkmo
02-26-2010, 09:55 PM
Okay, I didn't read all the posts....

14 seconds, that's gotta be impossible

I have never cheated or used a spreadsheet on a logic puzzle....I really bums me out if all of my record times will be cleared :(

The record time cut off should be somewhere around 45 seconds at least. Why am I being penalized for being speedy???

moodymom
03-03-2010, 07:48 PM
I could never make myself cheat on these - it completely defeats the purpose. But I have on 2 occasions solved a puzzle in under 40 seconds. I was astounded both times, but sometimes the clues are just laid out in a way that makes perfect sense. For those puzzles, there's often a lot of extra clues with duplicate information. Sometimes the 4x4 puzzles only need 3-4 clues to solve, but they might give you 6-7. There's still lots of people faster than me out there.

pcuser
03-07-2010, 07:33 PM
I've just finished reading this post from start to finish. Is it even possible to write a program that can solve a puzzle automatically. If so, then these speeds make sense as computers are fast enough. However, I can't believe many people can write or even know how to write code for this purpose. As far as using a spreadsheet to solve them, I wondered if one could do so when I learned how to solve these puzzles on this site. However, even if you created a spreadsheet that could do it, it would still require time to input data from the clues to allow the spreadsheet to work, thus defeating the intent. When I considered a spreadsheet, I thought of it as an intellectual challenge to myself, not as a way to get a leg up in competition. I have found that my times and depth of understanding has dramatically improved as I do more puzzles. I also don't understand someone being perturbed about fast times as there will always be someone better out there and I do these puzzles for myself, not to impress anyone else. I guess that's just me however.
pcuser

bobcat1
04-01-2010, 12:01 AM
I haven't played to long but I have found some puzzles for which a number of answers work, but the site only accepts one. So there is a lack of info in the clues in some cases

Laura
11-02-2012, 10:05 PM
I know this issue was raised a few years ago, but I'm seeing quite a few average solve times that are absolutely astounding! It takes an average of 30 mouse clicks to solve a puzzle and for a very fast reader, it would take each question about 2 seconds to read.

I've solved a few in about 80 seconds (very few :(), but am wondering what the secret is for the consistent low scores.

DumperDimple
04-08-2013, 03:50 PM
I just solved a 4x4 puzzle in 59 seconds, my first ever under one minute (although I've come close before).

No cheating, just logic, and luck, because if the clues come in a convenient order it makes it quicker to solve.

Definitely doable.

Laura
04-08-2013, 10:41 PM
I wasn't talking about individual puzzles — since that post a few months ago I actually did one in 40 seconds, but that's because I had done the same puzzle less than an hour before. My solve time on some puzzles has lowered quite a bit and I'm not hooting anymore when I solve a puzzle under 80 seconds.

My question was about average solve times of 30 and 40 seconds. And individual game solves of 14 seconds. (Interestingly we haven't seen those since the new puzzles were introduced — they don't have such discernible patterns.)

DumperDimple
06-12-2013, 12:23 PM
Hi, I just solved a puzzle in a new personal record time of 46 seconds. Am I right in thinking that it won't be counted toward my stats because it's too fast? I solved the puzzle straight-ahead, no cheating of any kind, no pattern recognition... it was just a particularly easy 4x4 puzzle. It would be sad if there were a disincentive to solve puzzles in under a minute because they might not be counted. Just sayin.

admin
06-12-2013, 12:29 PM
Hi DumperDimple -

Your solve will still be counted toward your stats, both monthly and lifetime.

Thanks for playing,
Stephen

Nusa
06-12-2013, 11:41 PM
Hmm, it only took me a few minutes to figure out one way to do it, at least for a computer professional like myself. Input is easy, all you need is prior access to puzzles so they can be solved, which is also easy if you think about it.

Certainly you can detect the more obvious users, but I don't see any way to eliminate the possibility other than handing out unique puzzles every time.

gypsywyz
12-12-2013, 07:50 PM
It's not about access to the puzzles, it's about repeating patterns....I find that a lot when I get working a lot of puzzles in one day. After awhile I start seeing the same patterns in the grids.:cool:

bocci
12-14-2013, 08:38 AM
Confession Time!

I've never achieved a monthly trophy, but I had a simple system to achieve record times when I felt so inclined.

Before Logic Puzzles added heavily to their puzzle list, I worked 4x5's exclusively, and soon discovered that the solutions to puzzles containing a '5 person' clue could be predicted solely from the 'x' pattern resulting from that single clue. Of the thousands of 4x5 puzzles, there were probably only about three-hundred patterns, and it seemed like two-thirds of those contained one '5-person' clue. And those '5-person' clues always resulted in nine 'x's. Most of those made for 4 'x's in the uppermost 5x15 squares. I focused on those.

I printed up sheets of blank grids, onto which I would record the four upper deck 'x's from the '5-person' clue of a given puzzle, and then, after solving, I would also record the 15 answer circles. Repeating this until I had around 120 distinct patterns was enough so that I could simply 'look up' the answer to a large percentage of the 4x5's that I encountered. I arranged the patterns in a quasi-numerical order according to where the 'x's fell row/column-wise, so I could quickly locate a particular 'x' pattern/solution that matched a new puzzle, and simply plug in the answers. It was crude, but I achieved some blazingly fast 'solve' times. Several were in the thirty-something seconds range.

My main motivation was to equal, or eclipse, the crazy records I saw. After doing so, I was satisfied that they were achieved by methods essentially the same as mine, and I threw all my cheat sheets away. But it was fun thinking about how it so irked and confounded some players. There you go!