Luck is overpowered, random-based events must be normalized

Describe your experience with the latest version of FreeOrion to help us improve it.
Forum rules
Always mention the exact version of FreeOrion you are testing.

When reporting an issue regarding the AI, if possible provide the relevant AI log file and a save game file that demonstrates the issue.
Message
Author
User avatar
afwbkbc
Pupating Mass
Posts: 96
Joined: Tue Mar 14, 2017 10:47 am

Re: Luck is overpowered, random-based events must be normali

#16 Post by afwbkbc » Thu Mar 30, 2017 10:39 am

So.... you're saying when we're born, somebody rolls some dice, and that's the amount of luck we're given for our life times?
Something like that. Exact mechanism isn't known but empirical evidence suggests of its existence.
I was a part time professional poker player for a number of years
Same here. But poker is not good example because influence of skill and amount of control is much higher than in freeorion, if compared to amount and influence of random uncontrollable events.
But they always reverted to the mean in the long run
Well it's highly possible that you are 'average luck' person. So you get 'average' results in long run. My luck is definetely worse than average - in poker I have to fold tons of times, and when I collected hands power statistics I noticed that I get good hands less often than others (though I encountered some players with equally bad luck, most of players got better hands more often). Even a triple is rare for me. Full house - maybe once in 100 games. I don't remember getting Royal Flush ever in my life. Ever. For most other players triples are more common. Royal Flushes are rare, but happen (and I get "WTF???" feeling every time I see it). In poker, however, you can still compensate lack of luck with good strategy and play adequately against luckier players, even in long run. In freeorion, you generally can't.
but in the long term, it will always, ALWAYS, average out
No. In my case long term leads to 'below average' result. For most people it's average. For some, it's above average.
The 'long run' is the problem, not temporal lucky or unlucky streaks.
That means that you can make your own "luck"
No, this is called "strategy". By luck I mean "positive or negative outcome of events which I don't have control over". For example, map generation in freeorion. Can I develop strategy to influence it? No I can't.
With a big enough data set, which by the way not one player of this game has, you'll see that every player on average was given identical starting conditions.
Well I'm pretty sure we are close to thousand games played now, and we already got used for one player (L29Ah) had ridiculously good starting conditions in majority of games. His empire's graph skyrockets nearly every game, and every time we find out that he got scylior with comp moon, muursh, kobuntura and tons of planetary bonuses early on. Definetely better than average starting conditions, most of the times.

User avatar
MatGB
Creative Contributor
Posts: 3310
Joined: Fri Jun 28, 2013 11:45 pm

Re: Luck is overpowered, random-based events must be normali

#17 Post by MatGB » Thu Mar 30, 2017 11:22 am

1) scouting
lucky: on first turn you send scout in random direction and find perfect planets and natives
unlucky: in first turn you send scout in random direction and find deep space with Sentry that destroys scout
this is meant to be impossible, so either report the bug or knock off the hyperbole because it degrades your argument.

The game is still in Alpha and nowhere near balanced properly, but it's also nowhere near as bad as some professionally released games I've played.

Try learning better strategies, it seems to work for your opponents.
Mat Bowles

Any code or patches in anything posted here is released under the CC and GPL licences in use for the FO project.

User avatar
afwbkbc
Pupating Mass
Posts: 96
Joined: Tue Mar 14, 2017 10:47 am

Re: Luck is overpowered, random-based events must be normali

#18 Post by afwbkbc » Thu Mar 30, 2017 12:23 pm

MatGB : you seem to not understand my point entirely.
"You must have better strategy" - I agree completely. The problem is that lucky players can get away with worse strategy and unlucky players must have better strategy for them to have equal odds. And what if some player is both lucky and has perfect strategy? Unlucky players will simply have no chance. Maybe in 1 game of 5. But generally they will keep losing without understanding why.
And the most concerning issue is that luckier players can sacrifice some safety to speed up empire growth and advancement (i.e. build colonies and research growth techs instead of building more scouts and military and researching scanners and planetary defences). Unlucky players will get owned if they do the same. So lucky players have better strategies available to them than unlucky players. That alone moves odds of winning the game strongly into their favor.
Please try to understand.

User avatar
Krikkitone
Creative Contributor
Posts: 1416
Joined: Sat Sep 13, 2003 6:52 pm

Re: Luck is overpowered, random-based events must be normali

#19 Post by Krikkitone » Thu Mar 30, 2017 3:22 pm

There is one issue where afwbkbc has a point. (perception rather than his metaphysics)

In a two player game (human or AI), each player has a 50% chance of getting "lucky" in that game.

However in a 4 player game, each player has a 25% chance of being the "luckiest".

Since you are playing for first place, you are always playing "against" the luckiest Other player.

So in a 4 player game, there is a 75% you will play against someone luckier than you.

(in evenly matched 2 player games you win-lose 50-50, in evenly matched 4 player games you win lose 25-75.. which makes it seem not evenly matched)

defaultuser
Vacuum Dragon
Posts: 570
Joined: Wed Aug 26, 2015 6:15 pm

Re: Luck is overpowered, random-based events must be normali

#20 Post by defaultuser » Thu Mar 30, 2017 10:14 pm

I wouldn't disagree that there can be a game balance problem where a player that gets a really great start is nearly unstoppable as long as they are reasonably skilled. I don't know what you do with that other than the more uniform starts as mentioned. The different layouts are one of the things I like about FO.

But that doesn't seem to be what is under discussion. You've stated without evidence that some people are inherently more lucky in some fashion to the point of being able to control things down to how a computer program operates. I don't know what that would be, some sort of psychic power? Because the physical laws of the universe as we currently understand them don't work that way.

User avatar
Vezzra
Release Manager, Design
Posts: 4674
Joined: Wed Nov 16, 2011 12:56 pm
Location: Sol III

Re: Luck is overpowered, random-based events must be normali

#21 Post by Vezzra » Fri Mar 31, 2017 8:00 am

I thought that the Dwarf Fortress Community is the most bizarre out there, but this discussion here certainly is a worthy match for them.

This is really... dwarfy. :lol:

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 12279
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

Re: Luck is overpowered, random-based events must be normali

#22 Post by Geoff the Medio » Fri Mar 31, 2017 8:54 am

Symmetric (rotational / reflected) galaxy layouts and start positions can probably be implemented as an option.

User avatar
afwbkbc
Pupating Mass
Posts: 96
Joined: Tue Mar 14, 2017 10:47 am

Re: Luck is overpowered, random-based events must be normali

#23 Post by afwbkbc » Fri Mar 31, 2017 4:52 pm

defaultuser wrote:You've stated without evidence that some people are inherently more lucky in some fashion to the point of being able to control things down to how a computer program operates. I don't know what that would be, some sort of psychic power? Because the physical laws of the universe as we currently understand them don't work that way.
I can't provide scientific evidence of that.
However, nobody can provide scientific evidence of the opposite (lucky/unlucky events are balanced in long run) either.
I have own empirical evidence that some people are just lucky than others, so they are better in activities that are influenced by chances.
Psychic power? No.
Actually, I don't know how it works and there doesn't seem to be a way to know.
There is no reason to discuss how 'luck' works, it's definetely off topic.
What we can do is add balancer programmatically which will keep us safe from any unfairness that results from somebody being luckier than others.
Symmetric (rotational / reflected) galaxy layouts and start positions can probably be implemented as an option.
No. As I said, reducing randomness will make game more boring. Randomized events are great, they bring diversity into gameplay. They just need to be balanced so everybody will be equally lucky overall.

User avatar
Dilvish
AI Lead, Programmer
Posts: 4720
Joined: Sat Sep 22, 2012 6:25 pm

Re: Luck is overpowered, random-based events must be normali

#24 Post by Dilvish » Fri Mar 31, 2017 11:16 pm

Paying attention to this thread is reminding me of picking at a scab -- intellectually I am pretty sure I should just avoid it entirely, but it is somehow just so tempting to try to fix/improve this one little part of the issue, here or there.
afwbkbc wrote:However, nobody can provide scientific evidence of the opposite (lucky/unlucky events are balanced in long run) either.
You think no one has ever attempted to scientifically study luck? Or that they somehow generated no evidence whatsoever?

Here is just one of the many interesting scientific research articles showing why people might think they are lucky/unlucky even when they are not. Perhaps this is not sufficient to convince you, perhaps none of them are, but you cannot honestly claim there is no evidence.

Also, here is some advice from one luck researcher on how to improve your "luck"; it does not involve rigging the RNG in your favor and I think at least some of the advice could be helpful to you.
If I provided any code, scripts or other content here, it's released under GPL 2.0 and CC-BY-SA 3.0

Morlic
AI Contributor
Posts: 250
Joined: Tue Feb 17, 2015 11:54 am

Re: Luck is overpowered, random-based events must be normali

#25 Post by Morlic » Sat Apr 01, 2017 5:47 am

Come on guys, let's be fair.

It is perfectly expected that there are about 15 or so people in the world that have an "average luck" 6 standard deviations away from the mean. Dude just happens to play Freeorion.

I suggest hardcoding the name afwbkbc in the RNG for balance purposes.
If I provided any code, scripts or other content here, it's released under GPL 2.0 and CC-BY-SA 3.0

User avatar
Gargamel
Space Floater
Posts: 33
Joined: Tue Feb 22, 2011 4:54 am

Re: Luck is overpowered, random-based events must be normali

#26 Post by Gargamel » Sat Apr 01, 2017 1:36 pm

He's just one of those "pro players" us other players enjoyed having at the tables.

Let's just rename his account to "troll" and get rid of him.

User avatar
afwbkbc
Pupating Mass
Posts: 96
Joined: Tue Mar 14, 2017 10:47 am

Re: Luck is overpowered, random-based events must be normali

#27 Post by afwbkbc » Sat Apr 01, 2017 2:04 pm

Guys, please be serious.
I don't care why you believe in fairness of the world or if you should or should not believe it.
Freeorion needs to be fair for everyone, regardless of his beliefs.
Yes, there is no scientific evidence of "luck is average in long run for everybody", this topic can not be researched. 'Luck' occurs only when there's unknown variables influencing outcome. And if there are unknown variables - you can't research their influence, because they are unknown. And collecting statistics on 'long run' about yourself is not scientific - science is about objectively testable evidence.
So please stop providing same invalid argument over and over again. Luck is out of scope of science. It may be within scope of some religion or esoterics but let's not go on on that here.

Scara
Space Kraken
Posts: 140
Joined: Thu Apr 23, 2015 11:21 am

Re: Luck is overpowered, random-based events must be normali

#28 Post by Scara » Sat Apr 01, 2017 6:44 pm

Sorry I don't really go with the concept of a luck normalizer...
The leaver to take IMHO is the difficulty level. It might be a subsection of difficulty level if its possible to implement such a thing....
I like to choose a bit larger universe, mostly 500 with 6 players.
This way it is not that crowded and your having a bit more space for some extra buffer to survive, when bad things happen or even having a bad starting position.
Might be an interesting experience playing a bad luck game on purpose :wink:

User avatar
Vezzra
Release Manager, Design
Posts: 4674
Joined: Wed Nov 16, 2011 12:56 pm
Location: Sol III

Re: Luck is overpowered, random-based events must be normali

#29 Post by Vezzra » Sun Apr 02, 2017 11:57 am

Sigh... ok, I'll bite.

First, before this might get a bit out of hand, a few things.

To be fair, afwbkbc is not a troll. A troll is deliberately disruptive, which afwbkbc isn't, he really believes the things he argues here. And while some good natured humor and poking some fun at each other is of course perfectly fine, please be extra careful not to take it too far, the line between harmless bantering and where things might get mean (or at least come across as such even if that hasn't been the intention at all) is crossed more easily than one might expect. Please don't get me wrong: I don't think any lines have been crossed yet, I just want to caution everybody before that happens, ok? :)
afwbkbc wrote:Guys, please be serious.
afwbkbc, you need to accept the fact that many of the ideas, thoughts and views you expressed here are, well, quite unusual, to put it mildly. And while of course everyone is entitled to their views/opinions/convictions you also need to be aware of how you, your opinions and statements are perceived by the others. Which means you maybe should consider backing down a bit and not pushing certain things too strongly if the feedback you get indicates that almost everyone else has a hard time taking you seriously on said certain things.

That will get you a long way when you want to be taken seriously.
Yes, there is no scientific evidence of "luck is average in long run for everybody", this topic can not be researched. 'Luck' occurs only when there's unknown variables influencing outcome. And if there are unknown variables - you can't research their influence, because they are unknown. And collecting statistics on 'long run' about yourself is not scientific - science is about objectively testable evidence.
So please stop providing same invalid argument over and over again. Luck is out of scope of science. It may be within scope of some religion or esoterics but let's not go on on that here.
Which basically utterly defeats your demand for a luck normalizer, don't you realize that? FO is a computer program, and computer programs are nothing else but applied math. Which means it's pure science in the end. It can only deal with things within that scope: known variables, things that can be researched (in whatever way), subjected to empiric evidence, etc.

How shall something that is essentially applied math account for something that is, according to your own words, completely out of the scope of science, any kind of research, cannot be grasped in any way by empirical means, is, by your definition, the ultimate "unknown factor"? If "luck" is what you claim it to be/works the way you claim it to work, it will cancel out each and any attempt at trying to "measure" it and then countering it by factoring in some numbers in the RNG which have been derived by these measurements. That would be impossible because of the very nature of "luck" like you describe it.

That said, even if we disregard the above completely and just look at what you want the luck normalizer to do, it cannot possibly work that way. The PRNG just spits out a pseudo-random sequence of numbers. If a particular number generated is "good" or "bad luck" (or neither!) entirely depends on the context in which that number is used. Calculating the positions of the star systems is a completely different matter than choosing starting positions, placing natives, specials, space monsters, chosing of targets by the combat engine, etc.

There is absolutely no way to evaluate how much "good" or "bad luck" a generated number is on the PRNG level. Even when completely disregarding that the PRNG has no clue at all how the number is further processed (does it become a coordinate of a star system, the index of a special to pick, the index of the starlane a space monster decides to continue it's travels on?), what is a "good" or a "bad luck" number is completely dependant (if that's determinable at all!) on context: a number might be "good luck" if it places a very valuable special near a players homeworld (although, if that special comes with a monster guardian, maybe not so much), but bad luck if it places a powerful space monster at the same location.

Furthermore there is the question: good or bad luck for whom? The number the RNG just returned might be very good for player A, but bad for player B. And it goes on just like that. Maybe you got the point by now.

Bottom line: a luck normalizer is completely impossible. If you don't believe me, well, be my guest, go ahead and try your hand at it. :wink:

User avatar
adrian_broher
Programmer
Posts: 1072
Joined: Fri Mar 01, 2013 9:52 am
Location: Germany

Re: Luck is overpowered, random-based events must be normali

#30 Post by adrian_broher » Sun Apr 02, 2017 12:28 pm

And while some good natured humor and poking some fun at each other is of course perfectly fine, please be extra careful not to take it too far, the line between harmless bantering and where things might get mean (or at least come across as such even if that hasn't been the intention at all) is crossed more easily than one might expect.
Are we at least allowed to make this a Adamant appreciation threada?
Resident code gremlin
Attached patches are released under GPL 2.0 or later.
Git author: Marcel Metz

Post Reply