#11
|
|||
|
|||
![]() bwt, I also assume that keeping a detailed history of each board wouldn't really take that much storage, but I doubt if they'd bother. (They obviously keep a list of found words, most words, best score, etc., and I assume that's associated with the internal board number. That can easily be used to decide what constitutes a new record.)
One thing that gives me hope is that registered users have their own user number, and some info about them can be obtained using that. (Once users reach the 51 game threshold & show up on any of the records boards (aside from longest/best word) you can see all their current stats just by clicking on their name. Even if they haven't reached that, if you know (or can find) their user number, you can use that to find current stats. I just wonder if there are other easily-accessible pages associated with board numbers. (I don't remember ever seeing one.) If too much info is available (especially if you can see your current board number), that would help cheating, if you're so inclined. |
#12
|
|||
|
|||
![]() Congrats bwt! When you're as good as you are, even gradual improvement is impressive, and it really adds up over time. Can't wait to see what
As for board information, I agree with Spike that it's really not so much bother or storage space if the site is keeping board-by-board information. If I were building such a thing by scratch as you suggested (a linked set of records), I agree it would be a nightmare to maintain. But the board and user IDs suggests the admin has done the appropriate thing and used tried-and-true database retrieval, where you store the data in keyed tables and retrieve it with queries. All you'd need to produce the statistics we've seen are two master tables: 1) Player information, keyed by player ID. No need to store every board; just the running tally of statistics, lifetime and monthly. The monthly statistics snapshots could be saved indefinitely to recalculate old records, or you could just save the record pages separately (I've done the latter myself, as all the old records are available back to 2009). There are about 100K player IDs, though only a few thousand are active now. Well under 1 KB of data per player, resulting in a totally manageable table of say a few dozen MB. 2) Per-board information, keyed by board ID. Only the current leaders and board overall statistics are necessary for the running "playerX beat playerY" sidebar, generated only when you update the table. Or you could store every player's statistics (words, score, best words) for each board. You could store this indefinitely, or clean up as boards are retired. Board IDs suggest there are fewer than 1 million boards, so information would be equally manageable. If it were me, I'd just store all the player information generated for every board. This would include additionally: * for each player, every word found, with timestamps. This is a lot more data, but it's just text from a modest corpus of words, so it can be compressed efficiently. The timestamps are already generated (they're in the html comments) and may be used internally to reinforce time limits / word entry order / etc. I find the timestamp a fascinating data source, and from my own play I've been dreaming up measures to analyze my performance (rate of word discovery as function of game time, clustering of word discoveries, etc.) Among other things, timestamps provide useful data to train an algorithm to detect cheaters, which is an interesting if not particularly trivial problem. Of course I'd hoard the data. As anyone probably guesses, I love data! |
#13
|
|||
|
|||
![]() Just broke 700 again (702) after having broken it for the first time last month. This in on a 5x5 game -- no way I'm good enough for that on a 4x4.
|
#14
|
|||
|
|||
![]() WTG bwt!
|
#15
|
|||
|
|||
![]() Awesome bwt! Looks like it might become a habit
![]() |
![]() |
«
Previous Thread
|
Next Thread
»
Thread Tools | |
Display Modes | |
|
|
All times are GMT -5. The time now is 11:38 AM.