Why do I have such bad luck?

[Jerrymander]

... or more to the point, from "square feet" to "cubic feet"?

I get it now. I wasn't initially thinking of exponents. I still don't think it crosses a magnitude boundary, since this is the same as randomly picking the lottery numbers, which you yourself said did not cross a magnitude boundary.

 

[RD-BH]

Ah OK you are referring to order of magnitude as log10 of a number, I see.

I'm not familiar with that as applied to statistical events though (and I did 2 years of stats during my maths degree). It is several standard deviations away from the mean however.

I'd be surprised if you lost 12 attackers against 1 defender at 62% odds though. Are you sure you didn't try attacking once 12 times and reloading inbetween (tedious though that sounds)? If you don't check "new random seed on reload" you always get the same result.

I'm not sure if CivIV uses the C++ rand() function or the Python Mersenne Twister method. rand() can be fairly clumpy, I'm not really a fan of it myself.

LOL, I was being generous with my math.
... The odds as stated by Buffy 3.19.003 were 100%.
... There were 20 attackers, and all 20 lost without a single attack.
... My assumption is a first strike calculation is missing from the advanced odds calculations.
... This has happened in more than one game (stack sizes vary).

I'm not familiar with that as applied to statistical events ...

Nor am I.
I'm not analyzing the outcome; I'm analyzing the equation predicting the outcome.
In certain situations, it crosses several magnitude boundaries.
Usually (not always), this indicates a problem somewhere in the calculation.

BtW:
I despise RNGs in general, because of their predisposition to crossing magnitude boundaries.
I have written and analyzed too many RNGs through the years to trust any of them.
I favor outcome stacks (like drawing lots),
... but the overhead is sometimes prohibitive,
... and you still need to shuffle with a RNG,
... which leads to its own set of harmonic problems.
I saw one RNG (reportedly used for security) based on lava lamps ... funny choice.
I was part of a team (decades ago) that analyzed lottery systems to make them less predictable.

 

[ParadigmShifter]

Mersenne Twister is apparently very high quality RNG (used by online gambling sites). rand() often shows clumps when plotted in 2 or 3 dimensions.

Mathematica uses a 1D cellular automata algorithm for its large integer random numbers but I heard that it doesn't perform well on chi-squared tests for randomness.

 

[clif9710]

What impresses me about CIV4 is the way it reproduces real life. Sometimes I think a kid might get a better education by spending time with CIV4 and an experienced guide than spending an equal amount of time in school...math, history, relations with others, it's all there in simplified form, though complex enough!

The odds in battles are one of the great teaching experiences. You learn quickly how much gambling has to do with emotions in the face of odds that just don't give a damn for you! You can get elated or depressed and do crazy things based on, essentially, a roll of the dice. Then, that craziness can lead to victory (I'm great!) or defeat (I can't play worth a damn). Imagine losing real troops and weapons and lives.

But best of all, CIV4 teaches opportunity cost. You can do one thing or another but you can't do both. All things are interrelated and several different victories are possible. Go for one type or another? Change plans in midstream or stick with it? You have many choices with each turn and the beauty of it all is that in the end you can look back and see where you went wrong in the face of a pretty clever AI - and play it again. The AI, and reality, never complain - they are always ready to take you on. The question is - what do YOU do?

And, by the way, the BUG/BULL mods with all the expanded stats in many well-laid-out screens makes a huge difference in understanding what is going on. For someone who is a learning addict, this game is tops.

So don't let the battle stats get to you. There is no sure thing. It IS possible to lose when everything says you should win, or win when it looks like loss is inevitable. Heck, aren't about 90% of movies about this? The real victory in this game, as in life, is learning about yourself.

 

[RD-BH]

There are many variations, they are still predictable.
"Better than the alternative" does not always mean sufficient.
Of course, we're only talking Civ IV here ...
... Python Mersenne Twister (sufficient)
... If used:
... ... The RNG isn't the source of misinformation I've seen.
... ... The displayed odds are suspect.

The Game is still The Game.
I'm not bothered by losing a stack of units to a RNG (okay, a little bothered).
I am bothered by being told the odds are X% when they're not.
Civ IV is a series of interesting decisions.
If I make a decision based on false information, then my plans are frustrated.

Why do I have such bad luck?
You don't.
You're being told you will be luckier than you can be.
I want to understand why.
... Short of that epiphany,
... ... Pay attention to successful/unsuccessful outcomes.
... ... You need to learn what works.
... ... I think it was TheMeInTeam who said if you play enough you'll get a "feel" for it.

 

[ParadigmShifter]

Take it up with the guy who does the Advanced Odds Calculator then (is it PieceofMind?)

 

[Tristan_C]

Unless you siege every city so completely that your raiders are merely a cleanup crew, you'll lose them all the time. Even 75-90 percent odds really isn't so great over a long period. Suicide catapults may be a shame, but they're the best medicine for our emotions. Throw a green catapult against a city, and give up on it, and once in a while you're in for a treat when it survives.

 

[PieceOfMind]

Take it up with the guy who does the Advanced Odds Calculator then (is it PieceofMind?)

Agreed. Take it up with me.

The mod is called Advanced Combat Odds and I can assure you all it is correct, including its dealing with first strikes, with a couple of nuances...

The first one being that as opposed to the vanilla Civ odds calculator, in ACO percentages are rounded. This means 100.0% as reported could in fact be between 99.95% and 100.0%. ACO will actually tell you explicitly if you have exactly 0 chance of losing a combat but that is a very very rare circumstance (like attacking severely wounded scouts after nukes with modern armor) so I doubt anyone has even seen it yet.

The second nuance: barbarian odds can be misleading depending on whether barbarian "secret combat mechanics" are set to hidden or not. The default behaviour in ACO is to reveal these but it's possible for a modder to hide them again.


(my comments in blue...)

A question or two or three then:

Q: When I lose a stack of troops against a lone defender with 62+% odds, and the defender does not sustain any damage (less than 8% chance according to extended odds in Buffy 3.19.003) ...
... Is there an error in the odds as reported? No
... Hidden data in the calculation? No, except perhaps as mentioned above
... Did the RNG roll the right number 12 times in a row? Can you be more clear about the meaning of this question? What is the "right number"?

Example:
12 vs 1 @ 62% odds of victory
62% is in range 7/12 .. 8/12 victories
Magnitude boundary is in range 2/12 .. 3/12 attacker victories
0/12 victories is below the boundary

Defender is expected to win 4/12 .. 5/12 times
From attacker PoV,
... defender victories could reach as high as 10/12
From defender PoV, (if permitted to defend against all 12 attacks)
... defender victories could fall as low as 2/12
12/12 defender victories crosses a magnitude boundary
Something is out of order ...

It isn't impossible for this to happen, but when it happens repeatedly, it starts to feel "cooked".

Please post a screenshot of the odds when hovering over the unit to attack. Alternatively, post some screenshots of the combat log after the 12 combats you are reporting.

I find it extremely unlikely you'd lost 12 combats in a row against a single defender with 62% odds, which is why I want to see proof.
Note that as someone else pointed out, reloading and retrying the combat does not count if you are playing under the option "no new random seed on reload".

LOL, I was being generous with my math.
... The odds as stated by Buffy 3.19.003 were 100%. Reminder: 100.0% means 99.95% or greater, but not necessary exactly 100%.
... There were 20 attackers, and all 20 lost without a single attack.
... My assumption is a first strike calculation is missing from the advanced odds calculations. First strike calculations are correct
... This has happened in more than one game (stack sizes vary).


Nor am I.
I'm not analyzing the outcome; I'm analyzing the equation predicting the outcome.
In certain situations, it crosses several magnitude boundaries.
Usually (not always), this indicates a problem somewhere in the calculation.

BtW:
I despise RNGs in general, because of their predisposition to crossing magnitude boundaries. What are these "magnitude boundaries"? EDIT...Ok I see what you mean by magnitude boundaries now. I think you could rephrase the problem as - floating point operations (as done by a machine) are more likely to produce errors as you reach more extreme orders of magnitude. That said, I am fairly confident the floating point calculations done in ACO are stable - that is, they are not overly susceptible to things like catastrophic cancellation and the like. The calculations are kinda like Taylor series expansions - as you add to a number, smaller and smaller numbers, the last very small numbers might have less and less effect on the overall sum, but the overall sum is very accurate in its order of magnitude. e.g. 100+10+1+0.1+0.01+... might get truncated as 111.11 (this is a simplification - floating point operations are much more accurate than that) but that is good enough for many uses.
I have written and analyzed too many RNGs through the years to trust any of them.
I favor outcome stacks (like drawing lots),
... but the overhead is sometimes prohibitive,
... and you still need to shuffle with a RNG,
... which leads to its own set of harmonic problems.
I saw one RNG (reportedly used for security) based on lava lamps ... funny choice.
I was part of a team (decades ago) that analyzed lottery systems to make them less predictable.

It is worth mentioning, and I went into this in detail in this thread (Combat Odds and RNG), that the RNG and odds calculator are two completely separate things. It is all too common to see the mixing of the two functionalities in conversations. The odds calculator is completely deterministic - you can be suspicious of its correctness if you want but I will assure you it is correct up to rounding. The RNG on the other hand is pseudo-indeterministic meaning its results can vary and it's possible to see streaky or flukish behaviour. (I call it only pseudo-indeterministic because strictly speaking it's completely predictable but only if you are actually aware of the random seed and know the algorithm being used).

 

[RD-BH]

Agreed. Take it up with me.

The mod is called Advanced Combat Odds and I can assure you all it is correct, including its dealing with first strikes, with a couple of nuances...

The first one being that as opposed to the vanilla Civ odds calculator, in ACO percentages are rounded. This means 100.0% as reported could in fact be between 99.95% and 100.0%. ACO will actually tell you explicitly if you have exactly 0 chance of losing a combat but that is a very very rare circumstance (like attacking severely wounded scouts after nukes with modern armor) so I doubt anyone has even seen it yet.

The second nuance: barbarian odds can be misleading depending on whether barbarian "secret combat mechanics" are set to hidden or not. The default behaviour in ACO is to reveal these but it's possible for a modder to hide them again.

I accept your assurance the odds calculator isn't giving false information.
... 1) These anomalies are against barbarians
... ... (I'll read up on that nuance shortly).
... 2) The rounding is interesting to note,
... ... ... but it is not large enough to be significant
... ... ... (It could knock another notch off the 2/12 gap making it a coin flip**,
... ... ... ... but the actual gap was larger).

**coin flip is an idiom here and not a specific reference to 1:1 odds.

Please post a screenshot of the odds when hovering over the unit to attack. Alternatively, post some screenshots of the combat log after the 12 combats you are reporting.

No screenshots.
I remember the combat log was more informative than the odds display.

I find it extremely unlikely you'd lost 12 combats in a row against a single defender with 62% odds, which is why I want to see proof.
Note that as someone else pointed out, reloading and retrying the combat does not count if you are playing under the option "no new random seed on reload".

I don't reload games,
... except between sessions,
... or when I'm studying a new teching path.
In my defense,
... I don't normally waste a stack after a certain percentage of losses,
... but I wanted to understand where the RNG is a factor.
I didn't see any difference in the combat log for these attacks.
... They were data clones of each other.
... This is why I thought it might be a first strike issue.

It is worth mentioning, and I went into this in detail in this thread (Combat Odds and RNG), that the RNG and odds calculator are two completely separate things.

Thanks for the link. I will read it shortly.
RE:Combat Odds vs. RNG
... Q: Could there be a mathematical anomaly ...
... ... ?) between Civ IV's combat calculations and ACO
... ... ?) something new in combat calculations with 3.19
I know, I know ... screenshots.

Did the RNG roll the right number 12 times in a row? Can you be more clear about the meaning of this question? What is the "right number"?

Darn,
... I was hoping you would clear this up for me.
I'm not sure where the RNG comes into the equation,
... or even how many times a new number is generated.
Sometimes it feels like only one random number is ever associated with a whole stack of combats ... uh, oh.
... I'll have to get back to you on this.

... uh, I have a sickening feeling right now.
What RNG does Civ IV use? If they are using a stock routine in a dll on my machine (either WinXP or C/C++) then I may be the cause of my own problem.
Nah! That's just crazy ... end paranoid delusion.

I despise RNGs in general, because of their predisposition to crossing magnitude boundaries. What are these "magnitude boundaries"? EDIT...Ok I see what you mean by magnitude boundaries now. I think you could rephrase the problem as - floating point operations (as done by a machine) are more likely to produce errors as you reach more extreme orders of magnitude. That said, I am fairly confident the floating point calculations done in ACO are stable - that is, they are not overly susceptible to things like catastrophic cancellation and the like. The calculations are kinda like Taylor series expansions - as you add to a number, smaller and smaller numbers, the last very small numbers might have less and less effect on the overall sum, but the overall sum is very accurate in its order of magnitude. e.g. 100+10+1+0.1+0.01+... might get truncated as 111.11 (this is a simplification - floating point operations are much more accurate than that) but that is good enough for many uses.

Excellent summation!
... I knew there were better words out there.
I'm too old to learn new math,
... thanks for looking past the words,
... ... and seeing my meaning
... (Okay, Taylor series aren't new math ... he's been dead a long time).
... Another example:
... ... Infinite sum (Euler?) of cosine and sine
... ... ... Stop calculating at the wrong magnitude and values can vary significantly.
However
... (isn't there always a however with me ),
... I'm not talking about the FPC.
... The shift in magnitude is ...
... ... 1) between the reported odds and the observed odds
... ... 2) in very specific combat situations.
... ... 3) frequent enough I noticed.
... ... 4) not frequent enough I was impelled to write a bug report.

RE: magnitude boundaries (orders of magnitude, scale, etc)
... They aren't always Log10,
... ... the base depends on the precision expected from the equation/model.
... They don't prove errors.
... They suggest the possibility of error.
... Combat percentages (0..100) lend themselves to Log10.

RE: RNG
With assurance the combat odds are accurate,
... this only leaves the RNG.
There is no upside to pursuing this issue.
... If my paranoid delusion turns out to be real,
... ... I've done this to myself.
... If Civ IV is using its own routines,
... ... then I have to live with the occassional anomaly.

TODO:
... 1) Read PieceOfMind's article.
... ... I'm sure I've read it before,
... ... ... though not with my current questions in mind.
... 2) Grab a screenshot the next time I see the anomaly.
... ... This may be awhile,
... ... ... I am currently playing Deity games with no barbs and an extra AI.
... ... ... ... (Barbs are too random on Deity,
... ... ... ... ... they might as well be goody huts)
... ... ... IIRC, the last time I noticed this was on immortal.
... ... ... Who knows,
... ... ... ... maybe I'll get a bee in my bonnet and stage an attack or two.
... 3) Reconsider using so many ellipses (...) for text formatting.
... 4) Have FUN 8)

 

[RD-BH]

Linear Congreuential Generator?

You're joking, right?

That isn't a RNG it's a dinosaur.
The problem isn't one of period,
... it is one of distribution,
... and programmer error in its correct usage.
... you can't even generate random numbers if you misapply the equation.
... Okay, they can't be misusing it significantly or there would be a whole lot more anomalies than I've seen.

The only advantages are its footprint (it is very small) and speed.
... Why would that even matter on a modern PC/MAC?
(note: some register shift tricks/methods are small and fast and have much better distributions)
Whoever suggested this as a good choice for a game was only thinking about size and speed.
... or they worked on old console games.
... or didn't have time to research alternatives.
... or are old and set in their ways.

I have to read more in PieceOfMind's article,
... maybe they changed the RNG since release.

 

[PieceOfMind]

Nope, the RNG is still the good ol' LCG and has been probably in every version of every civ game to date, and probably will be for Civ5 too.

In its defense, the LCG is not that bad. It suffers mainly from correlation issues which make it particularly unsuitable for gambling uses (like lotteries or online poker) and simulation purposes. For games it's more than adequate and I'd be highly suspicious of claims from people who reckon they could tell the difference if the game used a stronger RNG like the Mersenne twister. That is something where I would challenge them to prove their detection of the difference (between the two RNGs) is statistically significant.

It sounds like you'd be interested to know that every time a die-roll happens in Civ4 combat it is a 1000-sided die, with the numbers 0 to 999.
This is why combat odds for individual rounds of combat are exact and always accurate to the one decimal place.

Regarding your mention of the Taylor expansions for sine and cosine, they are a bit different to the situation in the Civ4 odds calculator because they involve alternating addition and subtraction. In Civ4, it's basically always sums of terms from a (positive) binomial distribution so they're always positive. The only subtractions that take place are very late in the piece and would definitely never cause any problems.

You are welcome to have a look at the algorithm in my mod, or the algorithm from vanilla, that is if you are happy reading a lot of C++ code.

Alternatively you can read the very brief derivation of the odds sums, linked to as a pdf article from the Combat Odds and RNG thread if I remember correctly.

Regarding "magnitude boundaries", I don't see how they have any relevance here.

By the way, I'm also willing to offer you the assurance that the LCG is working correctly as well. That is to say there isn't programmer error, at least not for the way the RNG is used in die-rolls for combat. I haven't checked for other RNG-calling features like GP births and so on.

 

[ParadigmShifter]

There's a website here: http://www.cacert.at/random/

That does standard random tests on 12MB+ of random data. Gnu C rand() and win32 rand() seem to do OK on the tests.

It doesn't mention what format it expects the data in though

 

[ruff_hi]

I did some (crude) testing on civ4's RNG - it passed my sniff test but I never did have a sensitive nose.

http://forums.civfanatics.com/showthread.php?t=339303

I will shortly be playing a SP game that will have lots and lots or fights in it. I am going to plot predicted outcome v actual outcome for that game and see how it stacks up. Back when I have finished the game.

 

[ParadigmShifter]

Yeah I read that yesterday when I saw the link to it in PieceOfMind's post about the odds calculator.

There's a standard suite of tests for statistical randomness, that's what the above website does to the data.

http://en.wikipedia.org/wiki/Diehard_tests

 

[PieceOfMind]

Thanks for the link PS. After following it a little, I found this (off topic!) gem:

http://en.wikipedia.org/wiki/Infinite_monkey_theorem

In 2003, lecturers and students from the University of Plymouth MediaLab Arts course used a £2,000 grant from the Arts Council to study the literary output of real monkeys. They left a computer keyboard in the enclosure of six Celebes Crested Macaques in Paignton Zoo in Devon in England for a month, with a radio link to broadcast the results on a website. One researcher, Mike Phillips, defended the expenditure as being cheaper than reality TV and still "very stimulating and fascinating viewing".[1]

Not only did the monkeys produce nothing but five pages[24] consisting largely of the letter S, the lead male began by bashing the keyboard with a stone, and the monkeys continued by urinating and defecating on it. The zoo's scientific officer remarked that the experiment had "little scientific value, except to show that the 'infinite monkey' theory is flawed". Phillips said that the artist-funded project was primarily performance art, and they had learned "an awful lot" from it. He concluded that monkeys "are not random generators. They're more complex than that. … They were quite interested in the screen, and they saw that when they typed a letter, something happened. There was a level of intention there."[1][25]

 

[ParadigmShifter]

I've read that article too

I notice that 6 << infinty though so the test is flawed.

 

[RD-BH]

Nope, the RNG is still the good ol' LCG and has been probably in every version of every civ game to date, and probably will be for Civ5 too.

At least they're using it correctly.

In its defense, the LCG is not that bad. It suffers mainly from correlation issues which make it particularly unsuitable for gambling uses (like lotteries or online poker) and simulation purposes. For games it's more than adequate and I'd be highly suspicious of claims from people who reckon they could tell the difference if the game used a stronger RNG like the Mersenne twister. That is something where I would challenge them to prove their detection of the difference (between the two RNGs) is statistically significant.

Civ IV combat outcomes aren't even that significant.
I disagree with you about people noticing.
People do notice shifts in scale on the low end.
Shifts from large to larger are mostly unnoticed.
Even used correctly,
... this RNG will be an annoyance to a percentage of players.
Last I checked,
... that is just a fact
... (human brain = pattern recognition),
... and it notices anomalies more than conformities.

It sounds like you'd be interested to know that every time a die-roll happens in Civ4 combat it is a 1000-sided die, with the numbers 0 to 999.
This is why combat odds for individual rounds of combat are exact and always accurate to the one decimal place.

It reduces some problems with the RNG.

Regarding your mention of the Taylor expansions for sine and cosine, they are a bit different to the situation in the Civ4 odds calculator because they involve alternating addition and subtraction. In Civ4, it's basically always sums of terms from a (positive) binomial distribution so they're always positive. The only subtractions that take place are very late in the piece and would definitely never cause any problems.

I only used this as an example because the elemental values cross magnitude boundaries.
Very noticeable on the small end,
... not so much as the values get larger.
Note:
... I understand I am saying small in reference to the largest values.
... Think scale.

You are welcome to have a look at the algorithm in my mod, or the algorithm from vanilla, that is if you are happy reading a lot of C++ code.

Thank you, unnecessary. 8)

Alternatively you can read the very brief derivation of the odds sums, linked to as a pdf article from the Combat Odds and RNG thread if I remember correctly.

Forgive my bluntness:
... meaningless data.

Regarding "magnitude boundaries", I don't see how they have any relevance here.

What is wrong with magnitude boundaries?
They're lines on a logarithmic scale.
When observed data crosses those boundaries,
... you start verifying your equations.

By the way, I'm also willing to offer you the assurance that the LCG is working correctly as well. That is to say there isn't programmer error, at least not for the way the RNG is used in die-rolls for combat. I haven't checked for other RNG-calling features like GP births and so on.

LOL, correctly isn't reassuring.
I looked at your article (Combat Odds and RNG).
The range is maxed, the data sampled (0..999), what else can you do to it?

In this case it is sufficient,
... and completely misses the point.

I thought Firaxis understood player psychology.
... 1) You don't want a RNG.
... 2) If you use one,
... ... you don't want it to be as random as possible.
... 3) Players want predictable outcomes.
... ... ... When a random element frustrates a player's strategy,
... ... ... ... they view it as a nuisance not a nuance.
... ... Case in point:
... ... ... How many players turn off random events and goody huts?
... 4) Players have a memory,
... ... so should your RNG.
... ... ... Has the number been rolled more than once in the last X combats?
... ... ... ... Reroll!
... ... ... ... Avoid the psychological impact on the player.
... 5) Players will complain no matter what you do.
... ... Hmm, ... yep, that sounds like me.
... 6) Players, they wanna have FU-UN 8)
... ... Oh, Players, just wanna ...
... ... That's all they really want ...
... ... Some FUN 8)
... ... ... (or something to that effect).

I don't want to die, there are still things I want to hear myself say.

 

[PieceOfMind]

People do notice shifts in scale on the low end.
Shifts from large to larger are mostly unnoticed.

For starters, I think we're talking about different things. If you started fiddling with how combat relied on the RNG, sure, people would notice and it would probably be pretty obvious. I am arguing that people won't notice the difference based on which RNG you use. Changing how the RNG gets utilised or what you do after the RNG call is not the same as changing the RNG itself. This relates to further comments I make below...

Last I checked,
... that is just a fact
... (human brain = pattern recognition),
... and it notices anomalies more than conformities.

True, but that is basically exactly why our intuition fails us with RNGs because unpredictability is to be expected and we are more likely to recognise patterns falsely i.e. recognise patterns where there are none.

This is a wishy washy sort of topic to discuss though and it's hard to argue much from fact here.

I thought Firaxis understood player psychology.
... 1) You don't want a RNG.
... 2) If you use one,
... ... you don't want it to be as random as possible.

Again, perhaps we are talking about different things here but I disagree strongly with this suggestion. It is absolutely necessary IMO that in any application that uses a RNG the RNG be as random as possible. If you want to fudge the results so you don't get strings of combat losses in a row, you do that after calling the RNG.

Personally I'd be against such changes to game mechanics. Suppose for example you said that if you lose 4 combats in a row that were all 80% or greater odds, then you automatically win the next combat at 80% or greater. That might make the player feel better about odds and luck and whatever else you want to call it, but it's abusable because those who know how the combat mechanics work can take advantage of the fact. They can go into a combat knowing they have no chance of defeat. A good odds calculator in that instance (one that doesn't lie to the player) would have to report 100% odds of victory. IMO this is getting into fairly absurd territory.

Seriously, if people can't handle the occasional setback based on a "string of bad luck" from some random process or phenomenon, what does that tell you about that person's grasp of reality and approach to life?

I've no problem with the suggestion of reducing the number of game mechanics that require an RNG - in many places that's good for a strategy game and you'll find people like TMIT saying that all the time - but fudging with RNGs to make players feel better about the game is no-go.

... 3) Players want predictable outcomes.
... ... ... When a random element frustrates a player's strategy,
... ... ... ... they view it as a nuisance not a nuance.
... ... Case in point:
... ... ... How many players turn off random events and goody huts?

I think you'll find that in general the players that suggest turning off random events and goody huts don't do it because they find those things to be a nuisance while playing the game. On the contrary, they will usually say it adds some flavour or uniqueness to each game. However, they do get annoyed that pointless random elements (like random events and huts) that are introduced into the game make comparisons between single player games starting from the same save rather meaningless. This is an ongoing debate that has been discussed a lot.

So, players turning off goody huts or random events IMO is missing the point on this issue.

... 4) Players have a memory,
... ... so should your RNG.
... ... ... Has the number been rolled more than once in the last X combats?
... ... ... ... Reroll!
... ... ... ... Avoid the psychological impact on the player.

RNG should have no memory else it no longer be an RNG. This is a very simple point IMO and like before, I think we talk about different things. Change the way the RNG gets used - not how it functions.

... 5) Players will complain no matter what you do.
... ... Hmm, ... yep, that sounds like me.
... 6) Players, they wanna have FU-UN 8)
... ... Oh, Players, just wanna ...
... ... That's all they really want ...
... ... Some FUN 8)
... ... ... (or something to that effect).

To many people, having fun in a strategy game is not all about getting thing to go in your favour all the time. Most are mature enough to enjoy having setbacks and seeing how they recover from them.

Also, any suggestion of changing how the RNG works just to please the human player causes many other issues:
a) What happens in MP? All of a sudden players don't get that "cheaty" RNG favouring them anymore? Then they'll complain that the RNG in MP is bugged.
b) It's asymmetric for AIs and human players. By far the majority of people who play games like Civ want AIs to follow exactly the same rules as the human player and get all riled up about game mechanics where there is some difference between how the AI or human gets handled. Progress is made on the game when yet another isHuman() call gets removed from the game code.

 

[RD-BH]

For starters, I think we're talking about different things. If you started fiddling with how combat relied on the RNG, sure, people would notice and it would probably be pretty obvious. I am arguing that people won't notice the difference based on which RNG you use. Changing how the RNG gets utilised or what you do after the RNG call is not the same as changing the RNG itself. This relates to further comments I make below...

I agree, with the exception of scale changes.
This RNG is known for large scale changes in its random sequences.
These are being mostly mitigated by 0..999, choice of seed, and choice of modifiers.

True, but that is basically exactly why our intuition fails us with RNGs because unpredictability is to be expected and we are more likely to recognise patterns falsely i.e. recognise patterns where there are none.

I agree.
I'm not sure falsely is the correct word.
The brain is trying to make sense of observed data (ie working as intended).

This is a wishy washy sort of topic to discuss though and it's hard to argue much from fact here.

I don't disagree.

Again, perhaps we are talking about different things here but I disagree strongly with this suggestion. It is absolutely necessary IMO that in any application that uses a RNG the RNG be as random as possible. If you want to fudge the results so you don't get strings of combat losses in a row, you do that after calling the RNG.

RE: As random as possible
I would argue you do not want 1000000:1 random number sequences in a strategy game.

Personally I'd be against such changes to game mechanics. Suppose for example you said that if you lose 4 combats in a row that were all 80% or greater odds, then you automatically win the next combat at 80% or greater. That might make the player feel better about odds and luck and whatever else you want to call it, but it's abusable because those who know how the combat mechanics work can take advantage of the fact. They can go into a combat knowing they have no chance of defeat. A good odds calculator in that instance (one that doesn't lie to the player) would have to report 100% odds of victory. IMO this is getting into fairly absurd territory.

This is not what I am saying.
As random as possible permits wild changes in scale that disrupt gameplay instead of enhancing it.
I don't recommend changes in odds.
I do recommend removal of scale changes in random number sequences.
... 80% odds would still be 80%,
... but 80% odds would stop being artificially <1% and >99%
The RNG generates a number 0..999,
... why should it produce 20 zeroes in a row?
... ... (^^^an extreme example^^^)
... ... btw, 0..999 indicates they are aware of the problems with this RNG,
... ... and are trying to mitigate these inherent weaknesses

Seriously, if people can't handle the occasional setback based on a "string of bad luck" from some random process or phenomenon, what does that tell you about that person's grasp of reality and approach to life?

Again, there is a significant difference between bad-luck, bad-luck squared, and bad-luck cubed.
The mere fact I observed extreme changes in scale does not mean this RNG is doing an insufficient job.
It does mean some people will notice even if they don't know what they are noticing.

I've no problem with the suggestion of reducing the number of game mechanics that require an RNG - in many places that's good for a strategy game and you'll find people like TMIT saying that all the time - but fudging with RNGs to make players feel better about the game is no-go.

Again, not what I am saying.
The only fudging I am speaking of is the removal of scale changes.
These scale changes are an inherent weakness of this (and other) RNG.
There is already fudging taking place ...
... 0..999 mitigates some of the effects of repeating sequences
... preselected seeds and modifiers predetermine the scale changes that will be present
... choice of RNG favors certain repeating sequences over others

I think you'll find that in general the players that suggest turning off random events and goody huts don't do it because they find those things to be a nuisance while playing the game. On the contrary, they will usually say it adds some flavour or uniqueness to each game. However, they do get annoyed that pointless random elements (like random events and huts) that are introduced into the game make comparisons between single player games starting from the same save rather meaningless. This is an ongoing debate that has been discussed a lot.

So, players turning off goody huts or random events IMO is missing the point on this issue.

Here you describe the point of my argument from my first post until now.
People notice when a random element has an inappropriate effect on gameplay.
A scale change in a random number sequence does not produce meaningful additions to a strategy game.
Non-Civ, extreme example:
Oh, look.
The AI just built a tank and took me out on turn two.
Wow, that was 1000000:1 odds.
Next game:
Oh, look.
The AI just built an Infantry and took me out on turn four.
Wow, that was 900000:1 odds.
... ad infinitum
Cool to observe and talk about?
Cool to plan a strategy around?
Such things are obvious scale changes and distract from gameplay.
The fact they are sampling 0..999 tells me they are aware of these problems and are already trying to mitigate them.

RNG should have no memory else it no longer be an RNG. This is a very simple point IMO and like before, I think we talk about different things. Change the way the RNG gets used - not how it functions.

This is Civ, not security or encryption.
This RNG already has a memory.
This RNG already has imposed limits.
This RNG is notorious for random number sequences containing large scale changes.
Again, 80% should be 80% not <1% or >99%.
My suggested correction to an already known problem does not change the odds.
It does limit the scale changes in sequences of random numbers.

To many people, having fun in a strategy game is not all about getting thing to go in your favour all the time. Most are mature enough to enjoy having setbacks and seeing how they recover from them.

Again, I am not saying this.

Also, any suggestion of changing how the RNG works just to please the human player causes many other issues:
a) What happens in MP? All of a sudden players don't get that "cheaty" RNG favouring them anymore? Then they'll complain that the RNG in MP is bugged.
b) It's asymmetric for AIs and human players. By far the majority of people who play games like Civ want AIs to follow exactly the same rules as the human player and get all riled up about game mechanics where there is some difference between how the AI or human gets handled. Progress is made on the game when yet another isHuman() call gets removed from the game code.

Again, I am not suggesting anything that changes the odds.
I am suggesting limits that prevent the RNG from artificially changing the odds.
This removes the psychological impact of noticing these shifts in scale.
This does not remove human desire to complain.

 

[jeffreyac]

Let me ask this:

You ask "... why should it produce 20 zeroes in a row?"

I ask, why not? Is producing 20 zeroes in a row not a possible, albeit unlikely, result from a random number generator?

I remember a Dilbert cartoon, where he tours accounting, and sees their 'random number generator', an imp that just says '9' over and over. Dilbert says "are you sure that's random?" and his guide answers "there's no way to tell..."

My point is random numbers being, by definition, random, will allow for some occasional wacky results. Personally, I welcome this as I view it as the introduction of uncertainty in my plans. I've been frustrated when a string of bad luck lays waste to my invasion plans as too much of my attackers get eaten in the first wave, but I've also rejoiced as a unit which probability says should not survive, instead triumphs in the face of incredible odds and saves my city.

So, maybe I'm not understanding the root of the debate here, but I think random means random, and I like it... am I looking at this too simplisticly?