Conquer Club

Had a quick look and hadn't noticed anyone raising this previously.

I have a slight issue with the up/down percentage figure included with the dice stats. Long term, everyone's % is zero. I understand that over a longer period it's much harder to maintain good or bad (despite many player's assertions to the contrary) luck, but the % figure means nothing. I think it would be better to use a value based on the standard deviation (positive or negative), or placing it in a percentile (different to %) of say between 1 and 100, thereby giving a more accurate idea of luck.

At present the % figure is the same if you area 0-2, or 0-202, yet there's clearly a huge difference in terms of luck between the two. I understand where the figure comes from, the attacker in a 3v2 has about a 55% chance of winning each battle, so the result comes from taking the current percentage (0% in both the above) minus the expected (55%), would be -55%. Whereas a standard deviation or percentile would give a truer reflection of the status.

Be good to hear what others think.

I like it, but not everyone knows what std dev is.

I also wouldn't mind a little standard deviation. Would that be difficult to code?

But, agent is right, there are probably many who would not understand it.

chapcrap wrote:I also wouldn't mind a little standard deviation. Would that be difficult to code?

But, agent is right, there are probably many who would not understand it.

i'm not sure how hard it might be re coding, i'm barely computer literate. If standard deviation might not be understood, a percentile score (eg out of a hundred) could give a simpler representation. eg. a score of 96 would mean you're in the tops 4% of luck, 50 being the even. In the previous example of 0-2 and 0-202, the first might represent about 25, whereas the second would be 0.0001, or lower. But again, I've no idea on the complexity of coding that.

Problem:

Why the users want Dice Stats: conversation fodder, curiosity, another thing to do on CC
Why the admin made it: To get people to shut the hell up about how dice aren't random or they ALWAYS get worse dice.

I have a feeling that the dice stats aren't going to see much change.

I created an excel random dice roll generator “=INT(RAND()/(1/6))+1” and ran a list of 100 3 vs 2 rolls. I actually ran this is groups totaling 750 times.

When I considered each 100 turns separately, I had results from 40% to 64% in favor of the attacker with an average of 53.9%. When I grouped the rolls in groups of 5, essentially 500 rolls in each larger group, the larger results were from 49.3% to 57.6% with the same average of 53.9%.

What this basically says is that if you take 75,000 rolls of 3 vs 2 you should win 53.9% of the time. If you cut that back to just 100 rolls, you could lose 60% of the time or win 64% of the time.

Random is random except when you are doing your thing thousands and thousands of times.

So, yes, the attacker does have the advantage, but don’t count on it!

agentcom wrote:Problem:

Why the users want Dice Stats: conversation fodder, curiosity, another thing to do on CC
Why the admin made it: To get people to shut the hell up about how dice aren't random or they ALWAYS get worse dice.

I have a feeling that the dice stats aren't going to see much change.

Indeed, the dice stats were created to stop people complaining that they have bad dice. But most of them keep doing so...
The only thing that would be interesting would have been to have a number or a bar that shows the luck in each specific game. This would stop people from claiming they lost on dice and prove those cases that this is the case indeed.

Hmmmm .... that would be kind of nice. And it would actually be useful. I know that I would use it personally and many clans would. Did I or a clanmate or my team lose because of bad dice or bad strategy? That would be useful information.

Quite a bit different from this suggestion though ... probably should be in a separate thread.

I don't even think that overall game luck matters quite as much as luck on specific attacks, like breaking bonuses, etc. Although, I wouldn't mind seeing the luck for each game.

chapcrap wrote:I don't even think that overall game luck matters quite as much as luck on specific attacks, like breaking bonuses, etc. Although, I wouldn't mind seeing the luck for each game.

Shhhhh....
Some people stopped complaining about the dice when the stats were created. Do not offer them new arguments they would never have thought about by themselves. By tomorrow there will be tens of people complaining that they have bad dice at the critical moments of the game. Please delete your post

agentcom wrote:Problem:

Why the users want Dice Stats: conversation fodder, curiosity, another thing to do on CC
Why the admin made it: To get people to shut the hell up about how dice aren't random or they ALWAYS get worse dice.

I have a feeling that the dice stats aren't going to see much change.

fair point. no reason why the percentile (or standard dev) suggestion couldn't be used in conjunction with the dice stats. I like the options also suggested by others, re stats for the game. My point is the current dice percentage is not a good indicator of luck in a specific sense, but can give a general impression at best.

peacemakers wrote:

agentcom wrote:Problem:

Why the users want Dice Stats: conversation fodder, curiosity, another thing to do on CC
Why the admin made it: To get people to shut the hell up about how dice aren't random or they ALWAYS get worse dice.

I have a feeling that the dice stats aren't going to see much change.

fair point. no reason why the percentile (or standard dev) suggestion couldn't be used in conjunction with the dice stats. I like the options also suggested by others, re stats for the game. My point is the current dice percentage is not a good indicator of luck in a specific sense, but can give a general impression at best.

I like the Std Dev, too. Just think that it would be useless for most users. Perhaps there could be a toggle/option. The default would be percentages, but you could change it to std dev if you wanted