Logic Puzzle Forums (https://www.logic-puzzles.org/forum/index.php)
-   Comments, Bugs and Feature Requests (https://www.logic-puzzles.org/forum/forumdisplay.php?f=3)
-   -   "My Scores" Statistics (https://www.logic-puzzles.org/forum/showthread.php?t=1081)

 abbishop 07-03-2017 03:32 PM

"My Scores" Statistics

It would be nice to see some additional data on the "My Scores" tab:

1) # of puzzles for which you own the current record time
2) Median and/or Mode time

I imagine all this data exists. Why not make it available?

Part of the reason I'd like to see the median or mode time is because early on after finding this site, I didn't realize that pausing the game doesn't pause the timer. A couple 81,000 second puzzles later, I have a semi-permanent blemish on my record. The median would be a much better indicator of typical performance.

I agree that it would be really cool to see how many records you have. I'm assuming that my number of records is constantly decreasing, as people continue to beat my best times.

It would also be cool if the system somehow acknowledged when you "lock" the record. Each puzzle size has a limit for how fast it will accept a time (30s for 3x4, 50s for 4x4, 70s for 5x4, etc.), and if you get something below that time, it will just say your time is the minimum (30s, for example). The first person to get the minimum has "locked" that record, because the best that anyone else can get would be a tie.

So, I would think it would be cool to see something like "you have the record on 70 puzzles. Of those, 58 are at risk, 12 are locked".

-----------------------------------------------------------------------------

As far as showing median times and mode times, I think the median at least would be useful. Actually, that's a huge understatement. The "average" time is completely useless, and should be abolished, and only the median time should be shown.

The average time is useless because if I am solving a puzzle, and messed up early, and I hit the "eliminate errors" button, it may find and remove 12 errors, penalizing me 120 seconds for each error. If most of my times are under 60 seconds, then I have to decide whether to finish solving this puzzle and submit it, ruining my average, or if I should never finish it, ruining my completion percentage. I shouldn't have to make that choice.

Honestly, if you are going to show an average, it should be an average for puzzles that were submitted without penalties. That would actually be useful.

 abbishop 07-21-2017 03:01 PM

Also, interesting about records being locked. Had no idea that was possible. Of my records, I'm almost completely sure 0 of them are locked.

 Sarge 07-23-2017 04:54 AM

Also agree - would love to see this info! I have much more fun trying to get the record time on a puzzle than competing for who can earn the most points in a month, which is heavily tied to how much time you can spend playing.

Quote:
 Originally Posted by abbishop (Post 4315) Thanks for replying, uigrad. Glad I'm not on an island on this. Also, interesting about records being locked. Had no idea that was possible. Of my records, I'm almost completely sure 0 of them are locked.

 jeffnn 10-03-2017 12:29 PM

Great suggestions

Definitely would appreciate more stats.

 judesalles 01-24-2018 08:20 PM

Yep, agree, the median is a better measurement than average.
But in a symmetrical Bell curve, the median is the mean (average).

 abbishop 03-02-2018 04:02 PM

Quote:
 Originally Posted by judesalles (Post 4614) Yep, agree, the median is a better measurement than average. But in a symmetrical Bell curve, the median is the mean (average).
Sure, but the symmetrical bell curve is irrelevant here, since a player's results are unlikely to follow one. This is especially true if they have games they'd had to step away from and didn't complete until hours later, or perhaps the next day. A few extreme outliers will wreck your mean.

 All times are GMT. The time now is 05:21 AM.