As mentioned some time ago in the general Nonosweeper thread, I'd prefer to have some additional rankings which are less sensitive against "lucky punches" (i.e. a record time which is a lot faster than the "normal" times of a player).

I suggest again a ranking of Top10-times. The ranking time for each level would be the sum of the 10 fastest games a player has achieved in that level. Those 10 times can be easily seen together by looking at the "Best times" table available in the Pro Version (see picture on the left of http://www.nonosweeper.com/pro.html).

For missing times (if anyone has completed less than 10 games in a level), a penalty time is counted. I propose the last listed time of a player, with an additional penalty of 10% of that time.

An example: My genius times are 216, 299, 322, 367, 607. So, my penalty times 6 to 10 would be 607 + 607·10% = 607 + 61 (always rounded up to next integer!) = 668. My ranking sum for Genius level would be 216+299+322+367+607+668·5=5151

If that is too difficult to calculate, we can take 999 as time for all missing ones.

Btw, I'm quite sure that, finally, no one will really care about the calculation of the penalty as to get in direction to the top of the rankings, you will have to achieve ten decent times regardless what the penalty algorithm is.

I know that at least Roland Seibt likes the idea of a ranking of Top10 times.

What does the rest think of it? Any comments or ideas?

Hi, as a new member of the forum I can now repeat what Stephan said. I like the idea of the Top-10or5-ranking very much. I like the calculation idea of the last stat. But considering for genius I think it'd be better to let people only join in general when 5 v 10 are finished. Thinking myself back a few months I was so happy completing genius even in 999secs. So it doesn't have to be a disadvantage taking that as penalty time.... Me switching now to other topic. Happy Nonosweeping.

I still prefer 10 over 5. lloyd, the 5 scores uploaded are the 5 best scores uploaded which is not necessarily the same the the 5 best scores overall.

I suggested 10, because in the Best Times window of the Pro version, there are listed 10 times. It might be easy to calculate the sum of those times and then create an url to upload the sum (maybe implemented in version 1.4?)

About the penalty: I uploaded every personal best as soon as possible to get a better rank in the ranking. I would have done this even if I had a worse average. Yariv, I think, there won't normally be a player who has finished only one single Genius game, but that one in 204. It's always a process of practising. That's another reason why to take Best Times and not Best Uploaded Times. A player who plays fast won't look at the timer until he has finished - and then, the game is saved regardless whether he likes having it in his Top10.

Maybe, we should wait on some more opinions...

OK, the problem is that the best times list stored locally can be edited by hand - so it can't be used in the rankings because people will cheat. Even if I change this, you would have to lose all the scores achieved up to now. 

At present, you can avoid putting your time into the list by closing the name entry screen, but anyway, I think players WILL start checking the timer before clicking the last square if there is a chance that the score could move them down the rankings, and I don't want to create an incentive not to finish games!

So I think we have to make it a matter of choice which scores are uploaded and used in the rankings - therefore it would be possible to have only one very good score in the database.
Either way, you must agree that the 216 / 204,220 situation is a problem.

You said that "I uploaded every personal best as soon as possible to get a better rank in the ranking. I would have done this even if I had a worse average" - this is reasonable (notwithstanding the above problems) because PBs will always move you up, but the question arises when you get a good score that isn't a PB - it shouldn't move you DOWN the rankings unless you have only a few scores AND it's substantially worse than your average.

@Jon: I never upload at once. I save the URL (there is an option to "Get URL") and then upload it when I'm connected to the internet next time.

Just a minor point, but if a score fails to upload due to internet problems, you can still grab the URL retrospectively from lastURL.txt and try again later. (from v1.3)

Good idea for the penalty, I think, but I'm not sure what you mean with "available scores". Uploaded ones?

I'd still prefer the Top10 idea, with the scores listed in the best times window. So, everyone can prove by screenshot if necessary.

It took a major rewrite (because it can't be done with a single SQL query any more), but I've finally done the rankings! Am I brilliant or what?!
It takes too long to query all the data each time you view the rankings, so I've had to construct a table which holds the averages and is (or should be) updated when someone uploads a new score - I'm too knackered to test it properly tonight, so please let me know if it isn't updating or doing what it's supposed to! Ta 

Whoops, I managed to screw up the upload script last night, sorry... should be OK now.

I don't quite see the point of a weighted best times list, although if there's a big demand I'll consider it.

You said that "not a single player is listed"... this worries me because they should all be there, although I had a similar problem for a while where the server couldn't make temporary tables. If that problem happens again, I'll have to investigate.
Can you see all the scores now, at least?

So far as I can see, the "problem" with the weighted rankings is that you are not at the top of them - have you considered that Roland deserves to be top with the current scores? What's more, if you upload more scores equal or close to your current bests you can go to the top anyway. In fact, I think the ranking system is working rather well, although I'm open to suggestions for more tweaking. 8)

Yes, I'm sorry about the fact that the scores for which there was previously no benefit in uploading are now used in the rankings
Unfortunately there's not a great deal that can be done about it except to play some more and upload more good scores! Anyway, I expect most people will be in the same situation...

It's supposed to update automatically, but I haven't tested that part yet so thanks for letting me know!
I'll try and fix it tonight (I'm at work at the moment)

Good point about the level-naming: I agree that a "beginner" level shouldn't really be included in the rankings at all, but as you may have guessed I deliberately kept the names the same as Minesweeper, so I blame that!

Notwithstanding the ranking positions, I think there's little doubt that the best Nonogram solvers are the best players on Genius and Expert, but I deliberately wanted to make it so that there are other ways of moving up the rankings, to give more variety in the play. I don't see much point in unweighted rankings, not because they don't show who is the best player (they probably do), but because they will almost always be the same as the expert rankings, so you might as well use those.
(Having thought about this, I can now see your point in wanting averaged level rankings, so I'll look into doing that when I have time.)

Anyway, (as with the Best Times listing where I put newer scores first,) I think it's fair to say that I've probably sacrificed a small bit of accuracy/purity in order to increase the enjoyment/interest levels and keep people coming back to the game.

Oh, the trials of being a game writer...

Yariv and all sweepers,
What about put players with the same scores in the same ranking position?

for ex. gero and I could be both second in intermediate..

do you understand my english and what I mean?
Luca

Normally, if two players have the same total, the player with the faster Expert time will be placed higher - I hadn't considered the possibility that two players would have the same time on every level.

Because of the limitations of SQL it's quite difficult to change the displayed ranking position - I'll do it eventually but for the moment I think we might as well agree that players with equal scores are ranked equally, even if their numbers in the lists don't show that!