Statistics dialog "Inflicted" and "Taken" unclear

Brainstorm ideas of possible additions to the game. Read this before posting!

Moderator: Forum Moderators

Forum rules
Before posting a new idea, you must read the following:
User avatar
revolting_peasant
Posts: 229
Joined: May 29th, 2012, 5:45 pm

Statistics dialog "Inflicted" and "Taken" unclear

Post by revolting_peasant »

I'm looking at the Statistics dialog - the one with the top table whose records are Recruits, Recalls, Advancements, Losses, Kills.

The bottom of the dialog, Damage, has Inflicated and Taken. Now, I would assume this would be in damage points, and I do see an integral value. but I also a lower value after a slash, and a percentage - both of which I don't understand. Example: "5739 / 5629.8 +1%"

I'm not asking you explain this to me, but I want to claim that - considering myself not very dumb - I'm thinking that a typical user would not understand what those numbers mean. I therefore suggest either a tooltip or splitting up those figures into different lines. Or some other visual indication of what they mean.
User avatar
Pentarctagon
Project Manager
Posts: 5531
Joined: March 22nd, 2009, 10:50 pm
Location: Earth (occasionally)

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by Pentarctagon »

That's not just a slash, it's the division symbol: 5739 / 5629.8 = ~101%

A tooltip saying it's "<actual damage> / <expected damage>" wouldn't be a bad idea though.
99 little bugs in the code, 99 little bugs
take one down, patch it around
-2,147,483,648 little bugs in the code
User avatar
revolting_peasant
Posts: 229
Joined: May 29th, 2012, 5:45 pm

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by revolting_peasant »

Ah, so it's expected in the probabilistic sense?

Well, for anything that's beyond a short scenario, it's safe to assume this will be very close to 100%, so it's not interesting to even display this (Law of Large Numbers and such).
User avatar
Pentarctagon
Project Manager
Posts: 5531
Joined: March 22nd, 2009, 10:50 pm
Location: Earth (occasionally)

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by Pentarctagon »

revolting_peasant wrote:Ah, so it's expected in the probabilistic sense?

Well, for anything that's beyond a short scenario, it's safe to assume this will be very close to 100%, so it's not interesting to even display this (Law of Large Numbers and such).
Probably, for longer scenarios even a difference of a few percent can actually be a lot.
99 little bugs in the code, 99 little bugs
take one down, patch it around
-2,147,483,648 little bugs in the code
User avatar
revolting_peasant
Posts: 229
Joined: May 29th, 2012, 5:45 pm

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by revolting_peasant »

Pentarctagon wrote:Probably, for longer scenarios even a difference of a few percent can actually be a lot.
But it won't ever happen after a few campaign scenarios. Much more likely for you to win the lottery.

Or - if it does happen, Wesnoth has a bug somewhere...
User avatar
Pentarctagon
Project Manager
Posts: 5531
Joined: March 22nd, 2009, 10:50 pm
Location: Earth (occasionally)

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by Pentarctagon »

revolting_peasant wrote:
Pentarctagon wrote:Probably, for longer scenarios even a difference of a few percent can actually be a lot.
But it won't ever happen after a few campaign scenarios. Much more likely for you to win the lottery.

Or - if it does happen, Wesnoth has a bug somewhere...
True; I just don't like saying "never" when random chance is involved :P
99 little bugs in the code, 99 little bugs
take one down, patch it around
-2,147,483,648 little bugs in the code
User avatar
Elder2
Posts: 405
Joined: July 11th, 2015, 2:13 pm

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by Elder2 »

It can consistently happen if you save and load a few times during the campaign scenarios, you do not need to even make like 10 save and loads during one scenario, fewer is enough, and the deviations you get are statistically significant.
Tad_Carlucci
Inactive Developer
Posts: 503
Joined: April 24th, 2016, 4:18 pm

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by Tad_Carlucci »

I've not done the maths so I can't speak to the significance.

What I see is a calculation with rounding which should be 'good enough' given a fair die.

Arguing that one can skew things be using save-and-reload is using a decidedly not-fair die. The entire goal of save-and-reload is to maximize damage inflicted and minimize damage taken. So, of course, the results are skewed.
I forked real life and now I'm getting merge conflicts.
User avatar
Elder2
Posts: 405
Joined: July 11th, 2015, 2:13 pm

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by Elder2 »

I get the impression that if you save and reload 90% of the time you will get positive luck and 10% of the time negative luck, about that, and even if you get negative luck later it will usually turn into positive. I think it has to be some flaw in the RNG or the effect of loading on the RNG
User avatar
Pentarctagon
Project Manager
Posts: 5531
Joined: March 22nd, 2009, 10:50 pm
Location: Earth (occasionally)

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by Pentarctagon »

Saving and reloading affecting the RNG in that significant of a way would be a pretty big bug I would think. Without a lot of evidence, I'm more inclined to think that you're getting better results after saving and reloading simply because you are creating additional opportunities for that better result to occur.

For example - If you flip a coin wanting heads, get tails, then flip the coin a second time and get heads, the fact that you got heads the second time isn't really evidence that there's something wrong with the coin.
99 little bugs in the code, 99 little bugs
take one down, patch it around
-2,147,483,648 little bugs in the code
User avatar
Elder2
Posts: 405
Joined: July 11th, 2015, 2:13 pm

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by Elder2 »

What would you consider "a lot of evidence"? Is it even possible to actually provide true evidence on this case?
Tad_Carlucci
Inactive Developer
Posts: 503
Joined: April 24th, 2016, 4:18 pm

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by Tad_Carlucci »

You have a coin with both sides showing heads when you save and reload.

You can't use your all-heads results to claim a fair coin, with both heads and tails, used in normal play has a problem.

The lot-of-evidence phrase was just being polite. The RNG is well-known. People better at it than you or I have been looking at it for decades.
I forked real life and now I'm getting merge conflicts.
User avatar
Poison
Posts: 171
Joined: August 13th, 2017, 4:54 pm

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by Poison »

How is this expected to work may I ask? What does expected dmg really means? Does it count terrain difference? I'm looking a scenario I haven't save-loaded at all, it run (at least visually) "normally", lost half of my army, and I was +4% / -9% (so huge difference)??? What does that mean? That I got extremely lucky?
User avatar
Elder2
Posts: 405
Joined: July 11th, 2015, 2:13 pm

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by Elder2 »

This sounds more like dismissing the problem to me, but if they indeed have been then I have nothing more to say about this.
User avatar
Inky
Forum Moderator
Posts: 527
Joined: September 22nd, 2014, 1:02 am
Location: USA

Re: Statistics dialog "Inflicted" and "Taken" unclear

Post by Inky »

ElderofZion wrote:I get the impression that if you save and reload 90% of the time you will get positive luck and 10% of the time negative luck, about that, and even if you get negative luck later it will usually turn into positive. I think it has to be some flaw in the RNG or the effect of loading on the RNG
When you say you got "90%" positive luck after reloading, did you actually pull up the statistics and check? Because of course you'd seem luckier after a save load, because you only save load when getting exceptionally bad luck! And when checking the statistics dialog, were you looking at the actual damage taken / received and not just the percentages? (For example, if you got +5% inflicted / +3% taken maybe it seems you got positive luck at first glance, but actually maybe you didn't deal much damage that turn and took a lot of damage so the +3% taken is the more important statistic and you actually got bad RNG.)
ElderofZion wrote:What would you consider "a lot of evidence"? Is it even possible to actually provide true evidence on this case?
Evidence would consist of giving a precise definition of "positive luck," (a logical example would be if (amount of extra damage inflicted) - (amount extra taken) >0 ), and then reloading a bunch of times and recording how many times the resulting turn meets that definition of positive luck.

Personally, I think the fact that no one else has mentioned this despite playing wesnoth for a long time is "a lot of evidence" that things are working correctly.
Post Reply