PDA

View Full Version : anyone seen this before?


kat
08-23-2008, 10:57 PM
Success Rate: 60%
Average Time: 146 seconds
Record Time: 0 seconds ()
Your Time: 59 seconds

Just curious...maybe it has to do with the new puzzles that Stephen added into the database.:confused:

maradnu
08-23-2008, 11:02 PM
I got the same thing with another puzzle.

I've tried to attach the screen print in a Word doc.

bansaisequoia
08-24-2008, 10:25 PM
Did I split the uprights?

bansaisequoia
08-24-2008, 10:30 PM
Okay, this time I think I got it.

admin
08-24-2008, 10:34 PM
Hi folks -

This seems to be related to the new quotes I inserted a few weeks back. I'm looking into it and will post when I have a fix/explanation. Thanks for catching this!

- Stephen

kat
08-24-2008, 11:40 PM
I also thought this was a little strange..

Success Rate: 71.4%
Average Time: 55 seconds
Record Time: 54 seconds (1ARE)
Your Time: 31 seconds


Funny that the average time is only a second slower than the record time...:confused:

bansaisequoia
08-24-2008, 11:46 PM
kat, undoubtedly the cryptogram had a higher average time before your solution of it in 31 seconds. Once this happened, the average time was lowered a bit.

kat
08-25-2008, 12:39 AM
but it still seems awfully close to me...

kat
08-25-2008, 01:36 AM
I just happened upon this strange statistic as well....

Success Rate: 40%
Average Time: 117 seconds
Record Time: 117 seconds (jmorningside10)
Your Time: 34 seconds

I really don't think that one has anything to do with my score. I think something screwy is going on at this point. Weird.:confused:

Looking back, it just dawned on me that perhaps the newer cryptos that were added are just now starting to show the statistics. Seems logical to me once my pea brain made the connection. Duh. Pencil shavings again.

(some time later....)

Maybe NOT pencil shavings...check THIS one out.

Success Rate: 100%
Average Time: 124 seconds
Record Time: 149 seconds (1ARE)
Your Time: 61 seconds

admin
08-25-2008, 02:33 AM
Hi everyone -

This is happening only to new cryptograms added in the past few weeks. The short explanation is that the stats are skewed simply because they've not been played enough times for the statistics to really have any meaning.

The longer explanation is that the "average" time is actually a median time, meaning that we take all solve times, arrange them from shortest to longest, and then select the one in the absolute middle as the median time. This tends to be a much more meaningful statistic than a simple average of all values, since thanks to the "recent games" feature puzzles can often be solved 1 or 2 hours after they were begun, and these longer solve times tend to inflate bare averages.

So, if a puzzle was solved 5 times, and the times were,

31, 56, 58, 78, 1029

... the median time would be 58. (An actual "average" would be 1252 seconds, which you can see is not indicative at all of how most people solved it).

Once we get more statistics in on these new puzzles, these "wacky" stats will wrangle themselves into place.

Also, we don't give record times for unregistered players, though we do include their times when calculating median scores. So it is entirely possible for a record time to be higher than the median time, especially now when so few games have been played.

Again, this issue will resolve itself over time. In the meantime though I may simply just not show statistics until a puzzle has been player 75 or 100 times.

Thanks everyone for catching this - I hope my explanation made sense.

- Stephen

flutefish
08-25-2008, 02:55 AM
Here's another one:

Success Rate: 40%
Average Time: 104 seconds
Record Time: 0 seconds ()
Your Time: 34 seconds

Now, while I understand Stephen's explanation, that doesn't apply to this quote. I solved the same exact quote earlier this evening, so there should have been a record time, since I am a registered member.

admin
08-25-2008, 02:59 AM
Statistics are not recompiled automatically - the server would completely overload if we did that. Instead we cycle a script once every minute that recompiles the statistics for one puzzle at a time. We process 1,440 puzzle statistics a day, so it generally takes about 3 weeks before a puzzles stats are reconfigured.

Again, its only an issue for new quotes. I now regret making statistics visible at all for them, its causing too much confusion.

kat
08-25-2008, 02:59 AM
Thank you for the explanation, I appreciate it. You are on top of things, as always. :)

flutefish
08-25-2008, 03:03 AM
That makes sense then. Thanks.

I like the stats though - I like knowing how others have fared on the same quotes. :o

admin
08-25-2008, 03:10 AM
Hi everyone -

I've gone ahead and remove stats display for quotes that simply haven't been played enough times yet for stats to make any appreciable sense. Apologies for the confusion this has caused!

I'm glad I've learned the lesson though, because I've got several thousand more quotes waiting to be entered into the system over the next several months. I believe I will space these out a bit more than I had intended to previously, so that players won't be completely inundated with puzzles that have no statistics showing.