Ratings

[quarterthru]

I don't know if anyone else has this problem, but I get really tired of seeing my ratings get destroyed by one down vote.  Is there anything we can do so that maybe we treat scoring like they do in the Olympics (for example) where the highest and lowest scores get removed from the tally so that one bad actor can't kill a puzzle's rating.  I've seen it happen to some great puzzles.  I think this is also a good idea for difficulty just to get a better average.[url=https://logic-masters.de/Dateien/index.][/url]

Ich weiß nicht, ob jemand anderes dieses Problem hat, aber ich habe es wirklich satt zu sehen, dass meine Bewertungen durch eine Abwärtsstimme zerstört werden. Können wir irgendetwas tun, damit wir die Wertung vielleicht so behandeln wie bei den Olympischen Spielen (zum Beispiel), bei denen die höchsten und niedrigsten Punktzahlen aus der Liste entfernt werden, sodass ein schlechter Schauspieler die Bewertung eines Puzzles nicht töten kann? Ich habe gesehen, wie es einigen großartigen Rätseln passiert ist.  Ich denke, dies ist auch eine gute Idee für Schwierigkeiten, nur um einen besseren Durchschnitt zu erhalten.

[Hausigel]

Not sure what you mean when you say the one vote "destroys"/"kills" your rating. Apart from that, same response as in the other thread: Authors are not entitled to receive good ratings. People have the right to not like a puzzle, and to express their opinion by voting accordingly.

[quarterthru]

Not sure what you mean when you say the one vote "destroys" / "kills" your rating. Apart from that, same response as in the other thread: Authors are not entitled to receive good ratings. People have the right to not like a puzzle, and to express their opinion by voting accordingly.

You have a good point. Everyone is entitled to their opinion and to express that opinion with their vote. And certainly no author is entitled to a good rating, but they are entitled to a proper representation of the scores they have received overall.  I'm simply suggesting that a minor change to the averaging system be made to get a better average closer to the mean.  Right now the way it is calculated, where you can give a score from 1-5, the score gets corrupted with one "out of normal" rating.  (ex. If by the 10th vote, a puzzle gets 6 5s and 4 4s the score comes out as 90% Now take that same with a 1 thrown in. 6 5s, 3 4s, one 1... 83%.)  By taking out the highest and lowest marks you get a better representation of the general rating.  You can also take note of this anomaly in the difficulty ratings.  Just because one solver finds something easy it skew the results for the actual difficulty for other solvers.  Maybe I'm just whining here, but I've seen these lone 1 ratings in my stats for a long time and it is not representative of the majority and should not be factored as such.

[Realshaggy]

I could make some of them 2 :-)

[Hausigel]

By taking out the highest and lowest marks you get a better representation of the general rating.

I disagree. As I see it, the best representation is the one which includes all the votes, not just some. Everything else is arbitrary.

[Joe Average]

Indeed, best representation is all the votes... at the times they were made. People's tastes change and puzzle solving skills evolve over time. A puzzle that I would have considered hard two years ago, may be a medium now, because I have more practise.

Additionally, it's not only the solvers who change, but sometimes also the puzzles. Maybe someone released their puzzle with very poor graphics, making it hard to spot some detail or distinguish between colors, especially when you print it out as a b&w sheet. That's something that can affect the beauty rating... and if the author re-uploads a better image, the people who solve it afterwards have a different experience than the first solvers.

For example, I recall a specific incident involving a slightly a bigger puzzle with me being the first solver a day or two after it was released. It was a long road to get there, as I had to do the puzzle about ten times, always hitting some dead-ends. I checked and double-checked and eventually messaged the author, asking if there could possibly be a mistake. Turned out there was not just one, but a whole lot of four mistakes in the puzzle, but those were discovered and fixed only one at a time. After every fix, the author assured me that the puzzle was now correct...

As it was I had a rather bad, frustrating experience, and it definitely showed in the rating I gave, despite it being a good puzzle after all the bugs had been fixed. I'd say my difficulty rating is accurate and comparable to those of other solvers, but the beauty one - who knows. For me presentation is part of the deal. It affects my "beauty" rating, as do originality, eureka moments along the solving path or an intresting frame story around the puzzle. And respect is another factor. I do respect the work the author puts in, but at the same time I want them to respect the time and energy I spend while working on their creation. All it would have taken them was to have someone test-solve it before release, or at the very least take it down for maintainance after the first signs of there being something wrong.

Well, much time has passed since, but for someone to not count my bad rating because it happens to be the lowest in the bunch for that puzzle just don't seem right. I had a valid reason to rate that puzzle the way I did.

[glum_hippo]

I am also not advocating for changes, but I am still glad quarterthru brought up this rather annoying matter. I have a handful of puzzles with decent ratings where a single 1/5 can have a really destructive effect even if it's just one of 10 solvers.
[Don't get me wrong, 1/5 is sometimes certainly justified, including on my own puzzles...]
But the only thing to do is shrug, hope that people will eventually grow up, and just make your next puzzle.

[PjotrV]

Hi Quarterthru,

I feel you! I got something similar going on. I posted it in an earlier topic as well. Nothing will change unfortunately.
I even had proof who it was, but all I got was the blame put on me.

New-joiners with Sudoku puzzles are apparently not welcomed by some people and you don't get support from the others. The only remedy that is working is not posting 3 star or lower difficulty, because those will not be rated by this specific person.

Kind regards,

PjotrV

[Semax]

Is there anything we can do

Yes, put "Sudoku" into the title, if it's a Sudoku variant.

In general: make the title (and tags and labels) as accurate as possible, such that people not liking this kind of puzzle won't solve it.