bad great generals

[andrewlt]

The problem mostly isn't that the RNG is inaccurate or whatever. The problem is that the game encourages you to use your great general last in combat, which is the opposite of reality until the modern ages. Just about every great general lead from the front or near it. Right now, the game is basically asking you to have 10 macemen attack your enemy's 5 or so longbowmen first and only use your great general against that last one with probably 0.5/6 strength remaining. Attaching them to a unit seems vastly underpowered for what you have to do to get them. It's only useful to use the first one for a heroic epic/west point rush, then use subsequent generals as a super specialist.

 

[drkodos]

The problem mostly isn't that the RNG is inaccurate or whatever. The problem is that the game encourages you to use your great general last in combat, which is the opposite of reality until the modern ages. Just about every great general lead from the front or near it. Right now, the game is basically asking you to have 10 macemen attack your enemy's 5 or so longbowmen first and only use your great general against that last one with probably 0.5/6 strength remaining. Attaching them to a unit seems vastly underpowered for what you have to do to get them. It's only useful to use the first one for a heroic epic/west point rush, then use subsequent generals as a super specialist.

Excellent points, particularly about the way real world combat has seen large changes in how leaders interact on the Battlefield, but I have had many victories when using them as the initial attacker to take out the strongest unit in the targeted city. When properly promoted and attached to units such as Macemen with max City Raider & Combat Promotions, they pimp smack the best city defenders (after a nice softening by cats, cannons or artillery as the period would dictate), and then make the job easier for the lessor units to be successful in their follow up attacks.

Weird segue: If a coin is flipped 50 times and lands heads up each time, there is still a fifty percent chance it will land heads up on the very next flip. Thus, I agree with those posters arguing that statiscal analysis is mostly misunderstood by many people and that human brains tend to improperly weight personal experiences in relation to mathematical probabilities.

Even if something has a 1 percent chance of occuring, and since each incident is unrelated to other incidents (subsequent outcomes are not affected by previous outcomes), an incident could theoretically occur many times in sequence and not be anomolous, regardless of the odds that such sequences could happen.

We are here aren't we? What were the chances that a bagfull of 1,000 paper clips would form a completely united and linked chain simply by shaking the bag?


It happens.

 

[magfo]

Humans just can't understand odds. That's how it is. If you've got 95% to win, then you exepect to win every time. If you lose you are pissed. Still, you OUGHT to lose 1 in 20. But every time you do, you're pissed.

I've played a lot of online poker, THERE you see this discussion multiplied by a million times. Odds are hard to intuitively understand.

Also, Lady Luck is not a fair lady. Some get more, some get less. In Civ4 it's not really a matter of life and death though. If you get *THAT* annoyed, then just reload and replay.

 

[kcbrett5]

Weird segue: If a coin is flipped 50 times and lands heads up each time, there is still a fifty percent chance it will land heads up on the very next flip. Thus, I agree with those posters arguing that statiscal analysis is mostly misunderstood by many people and that human brains tend to improperly weight personal experiences in relation to mathematical probabilities.

Even if something has a 1 percent chance of occuring, and since each incident is unrelated to other incidents (subsequent outcomes are not affected by previous outcomes), an incident could theoretically occur many times in sequence and not be anomolous, regardless of the odds that such sequences could happen.

We are here aren't we? What were the chances that a bagfull of 1,000 paper clips would form a completely united and linked chain simply by shaking the bag?


It happens.

There are so many problems with this logic I don't know where to begin.
1st-if you land on heads 50 times in a row I am betting my life savings on that 51st flip being tails. You cannot view it independently like you suggest. True, the odds of the 51st flip are still 50/50. But the odds of hitting 51 in a row are astronomical. This is why you can win at roulette betting only black/red, high/low, and odd/even. If you dont believe me, try a simple experiment. Flip a coin continuously until you get 5 heads or tails in a row. Then record what the 6th flip is. Repeat this as often as you want until you have a large enough sample. Then see what all of your 6th flips were and I guarantee you they will not be 50/50. Try it. You will see.

2nd-Humans are not a bag of paperclips. Contrary to biblical genesis, we were not conjured out of thin air, but evolved over a period of millions of years from a single cell organism. Of course paperclips cannot assemble themselves, they are not alive. DNA however, is extremely good at assembling itself and copying itself.

 

[Older than Dirt]

If you take a course in molecular biology you may be surprised at the complexity of a "simple" cell. If macro-evolution were possible (i.e., several basic laws of physics are not operative and it were possible for information to generate itself and the problem of irreducible complexity is overlooked) the probability of such simple cell forming by random process is larger than the number of atoms in the presumed size of the universe. This is just silliness and shows how irrational and illogical the post modern era is.

 

[merlinme]

kcbrett5: are you for real? If so, can I play poker with you? The number of people that have gone bust assuming that something that happens several times consecutively cannot happen again is very, very high.

Quite apart from the mathematics, your logic seems very strange. Would you have bet your lifesavings on the coinflip being tails on the 50th flip (as opposed to the 51st flip)? If so, you just lost your lifesavings.

I can actually see a stronger case for assuming the pattern will continue (because the coinflip is being fixed) then for assuming the pattern must end. But assuming the coinflip is genuinely random, it's 50-50, no matter how many times it came up heads. Very unlikely random events will eventually happen if you try for long enough.

 

[DrewBledsoe]

kcbrett5: are you for real? If so, can I play poker with you? The number of people that have gone bust assuming that something that happens several times consecutively cannot happen again is very, very high.

Quite apart from the mathematics, your logic seems very strange. Would you have bet your lifesavings on the coinflip being tails on the 50th flip (as opposed to the 51st flip)? If so, you just lost your lifesavings.

I can actually see a stronger case for assuming the pattern will continue (because the coinflip is being fixed) then for assuming the pattern must end. But assuming the coinflip is genuinely random, it's 50-50, no matter how many times it came up heads. Very unlikely random events will eventually happen if you try for long enough.

That reminds me of something, now what was it?.........ah yes, some monkeys , some typewriters and the complete works of Charles Dickens...."It was the best of times, It was the blurst of times"..you stupid monkey as Mr Burns said.......

 

[kcbrett5]

kcbrett5: are you for real? If so, can I play poker with you? The number of people that have gone bust assuming that something that happens several times consecutively cannot happen again is very, very high.

Quite apart from the mathematics, your logic seems very strange. Would you have bet your lifesavings on the coinflip being tails on the 50th flip (as opposed to the 51st flip)? If so, you just lost your lifesavings.

I can actually see a stronger case for assuming the pattern will continue (because the coinflip is being fixed) then for assuming the pattern must end. But assuming the coinflip is genuinely random, it's 50-50, no matter how many times it came up heads. Very unlikely random events will eventually happen if you try for long enough.

Oh I am quite serious and I assure you I am right. Yes I would have bet my money on the 50th flip being tails as well and I would have won that one too. But see, I wasn't presented with that option. In his fictitious world he had already managed to flip heads 50 times so my only choice was on the 51st. If he had said he flipped 49 heads than I would have bet on the 50th and won.

To steal from Mr. Bledsoe's analogy. If a million monkeys were to flip coins for the rest of their existence not one of them would flip 50 heads in a row.

If I am wrong, than the odds of flipping 50 straight heads would have to be 50/50. Do you really believe that to be the case?

Poker is different. It is not a black/white, right/wrong decision. If you dont believe me, try my experiment above. If you are really convinced you are right, you have nothing to lose.

 

[kcbrett5]

If you take a course in molecular biology you may be surprised at the complexity of a "simple" cell. If macro-evolution were possible (i.e., several basic laws of physics are not operative and it were possible for information to generate itself and the problem of irreducible complexity is overlooked) the probability of such simple cell forming by random process is larger than the number of atoms in the presumed size of the universe. This is just silliness and shows how irrational and illogical the post modern era is.

So evolution is impossible? Explain to me how viruses become immune to medicines that used to kill them? Explain why penicillin stops working if you take it too much? Explain why radiation causes mutations? On the single cell level, evolution happens every single day. It may be extremely improbable, but there are unimaginable #'s of cells in the world.

 

[Sparta]

Oh I am quite serious and I assure you I am right. Yes I would have bet my money on the 50th flip being tails as well and I would have won that one too. But see, I wasn't presented with that option. In his fictitious world he had already managed to flip heads 50 times so my only choice was on the 51st. If he had said he flipped 49 heads than I would have bet on the 50th and won.

To steal from Mr. Bledsoe's analogy. If a million monkeys were to flip coins for the rest of their existence not one of them would flip 50 heads in a row.

If I am wrong, than the odds of flipping 50 straight heads would have to be 50/50. Do you really believe that to be the case?

Poker is different. It is not a black/white, right/wrong decision. If you dont believe me, try my experiment above. If you are really convinced you are right, you have nothing to lose.

Okay, now I want in on this poker game, too.

Seriously, kcbrett5, your analysis is significantly flawed (no offense). Statistics just don't work that way. The odds of 50 heads would be .5^50, but the odds of heads on that next coin flip would still, and always, be 50% (very nearly; IIRC tails has a negligible advantage due to coin weighting). If your theory worked, why would any of us be wasting time online discussing anything when we could all be becoming rich at the casinos' expenses? And how have casinos continued to exist for this long in the first place with a prominently featured 'beatable' game (you don't suppose you're the first to come up with this theory, do you?)?

Martingale System

Originally, martingale referred to a class of betting strategies popular in 18th century France. The simplest of these strategies was designed for a game in which the gambler wins his stake if a coin comes up heads and loses it if the coin comes up tails. The strategy had the gambler double his bet after every loss, so that the first win would recover all previous losses plus win a profit equal to the original stake. Since a gambler with infinite wealth is guaranteed to eventually flip heads, the martingale betting strategy was seen as a sure thing by those who practiced it. Unfortunately, none of these practitioners in fact possessed infinite wealth, and the exponential growth of the bets would quickly bankrupt those foolish enough to use the martingale after even a moderately long run of bad luck.

drkodos and merlinme are right - given a large enough sample, any number of counterintuitive results is possible within it.

If you still don't belive me (/us), and value(life savings) > $10K, I will bring a coin out to Kansas City to demonstrate .

 

[gdgrimm]

Does anyone know if it stacks on units that already have a chance to withdraw? Would a chariot warlord or a trebuchet warlord have a really high chance of withdrawing?

Yes, it does. A Cavalry warloard with 90% withdrawal rate is pretty indestructable on the offensive.

 

[kcbrett5]

Okay, now I want in on this poker game, too.

Seriously, kcbrett5, your analysis is significantly flawed (no offense). Statistics just don't work that way. The odds of 50 heads would be .5^50, but the odds of heads on that next coin flip would still, and always, be 50% (very nearly; IIRC tails has a negligible advantage due to coin weighting). If your theory worked, why would any of us be wasting time online discussing anything when we could all be becoming rich at the casinos' expenses? And how have casinos continued to exist for this long in the first place with a prominently featured 'beatable' game (you don't suppose you're the first to come up with this theory, do you?)?


If you still don't belive me (/us), and value(life savings) > $10K, I will bring a coin out to Kansas City to demonstrate .

Silly man. Casinos know this fact as well. Why do you think every roulette table has a maximum bet? The only way to beat the table is to keep doubling your bet everytime you lose, until you win. So they instituted a maximum bet to prevent people from doing just that.

I am not the first person to come up with this idea, but I do seem to be the only person here who understands it. Seriously, start flipping the coins as I say and you will see. Bring your coin to my house and I will gladly take all of your money until you learn. And you didn't answer my question so try this. What are the odds of flipping 2 heads in a row? 25% right?
What are the odds of flipping 3 heads in a row? 12.5% right?
So the odds of flipping the 3rd coin are not 50/50, it is only 12.5%. The events are not independent.
I know the statistical math behind your argument, and I know it isn't wrong. I am just saying it isn't relevant to this sitution. It is a commonly misunderstood aspect of analysis. Seriously, try my experiment above and then come back and tell me im wrong if I am.

Your own maringale paragraph confirms that I am right. It would take an extreme run of bad luck to lose all of your money. If the odds are always 50/50, then it would not take an extreme run of bad luck. You would lose all of your money half the time.

Of course the odds of each flip are 50/50. But the flips are not independent. And of course you can lose the 51st straight flip, its just that the odds of it happening (in a straight game) are astronomical.

 

[Pantastic]

1st-if you land on heads 50 times in a row I am betting my life savings on that 51st flip being tails. You cannot view it independently like you suggest. True, the odds of the 51st flip are still 50/50. But the odds of hitting 51 in a row are astronomical

Each coin flip is independant of the others; flipping heads 50 times in a row doesn't make the coin emit magical control rays to guarantee that the next flp is heads.

This is why you can win at roulette betting only black/red, high/low, and odd/even.

On average, the house always wins at roulette. You can get lucky and happen to come out ahead in a single game, but in the long run you'll lose. No roulette system actually works, except for the being the house, which makes billions but isn't what people usually mean.

If you dont believe me, try a simple experiment. Flip a coin continuously until you get 5 heads or tails in a row. Then record what the 6th flip is. Repeat this as often as you want until you have a large enough sample. Then see what all of your 6th flips were and I guarantee you they will not be 50/50. Try it. You will see.

I've done this before and you're wrong; your results will be approximately 50-50. Unless you do a lot of repeats it's not going to be exactly 50-50 since your sample size will be small, and the weight of the coin will introduce some small bias that way too, but there's going to be no difference based on what the previous coin flips were. Again, previous coin flips don't emit magical rays that affect future coin flips, each flip is independant of what you did beforehand.

 

[kcbrett5]

I've done this before and you're wrong; your results will be approximately 50-50. Unless you do a lot of repeats it's not going to be exactly 50-50 since your sample size will be small, and the weight of the coin will introduce some small bias that way too, but there's going to be no difference based on what the previous coin flips were. Again, previous coin flips don't emit magical rays that affect future coin flips, each flip is independant of what you did beforehand.

Somehow I doubt that you have previously done exactly what I am saying. Because I am tired of arguing this point with people not open minded enough to get it. Go on believing whatever you want.

I understand the argument you all are making, call it montecarlo fallacy or gambler's fallacy or whatever you want. I am just saying that the experiments don't support the math.

 

[Pantastic]

Silly man. Casinos know this fact as well. Why do you think every roulette table has a maximum bet? The only way to beat the table is to keep doubling your bet everytime you lose, until you win. So they instituted a maximum bet to prevent people from doing just that.

If you can use that system, you don't need to beat the table because it only works if you start off with infinite money. If you start off with only a finite amount of money, then your system is just a big gamble, it has a definate possability of failure because you can go broke and not be able to make the next doubled bet. Plus I don't know that I'd count it as 'beating the table' to get lucky early and win $4 early on.

 

[Pantastic]

Somehow I doubt that you have previously done exactly what I am saying. Because I am tired of arguing this point with people not open minded enough to get it. Go on believing whatever you want.

In the actual experiment back in the day, I think we only used 4 heads in a row instead of 5. Do you believe that the magical coin-flip-changing rays only start with 5 flips in a row?

I am just saying that the experiments don't support the math.

This would be big news, upsetting the science behind tons of modern physics, chemistry, and mathematics. Can you link us to a published version of this earth-shattering experiment? If not, you should really do this in a controlled environment and publish it, you'll be rich and famous in no time!

 

[PennHead]

What are the odds of flipping 2 heads in a row? 25% right?
What are the odds of flipping 3 heads in a row? 12.5% right?
So the odds of flipping the 3rd coin are not 50/50, it is only 12.5%. The events are not independent.

Let me see if I can explain why this is wrong.

For simplicity's sake, lets just consider two coin flips. The odds of getting heads twice in a row is indeed 25%. The reason for that is that there are four possible outcomes: heads/heads, heads/tails, tails/heads, and tails/tails.

It is absolutely accurate to say that I only have a 25% chance of getting heads/heads before any coins have been flipped.

However, say I have already flipped once, and got heads. There is one flip remaining. Now there are only two possible outcomes: heads/heads, or heads/tails. I can't get tails/heads or tails/tails because I already got the first heads. Since there are only two possible outcomes, my odds are 50%. There is no factor that makes one of these outcomes more likely than the other.

Frankly, it's the commonly held belief that, assuming a perfectly random system, past events have an influence on future outcomes that helps keep casinos in business.

Back on the thread topic, I too have felt like I've lost more 90%+ odds battles than I should since buying Warlords. I haven't kept track though, so I can make no further claim other than it doesn't feel right. Enough so that I do plan to keep track of it in my next few games to see if the numbers bear out my gut feeling. I won't be surprised if it all works out according to the numbers, and that it's just my perception that's wrong. However, it also won't surprise me if there is indeed a bug at work. I'm not really qualified to say until I've actually made an attempt to do analysis on real numbers.

 

[Sparta]

Somehow I doubt that you have previously done exactly what I am saying. Because I am tired of arguing this point with people not open minded enough to get it. Go on believing whatever you want.

I understand the argument you all are making, call it montecarlo fallacy or gambler's fallacy or whatever you want. I am just saying that the experiments don't support the math.

I think we are just looking at it from differing perspectives. I'm saying that if you have already gotten three heads in a row, regardless of the odds of that specific scenario (which I agree would be 12.5%), you still have a 50/50 chance of heads on the next flip. Even though HHHH is a one-in-sixteen occurence, if the first three heads are already a given (/already occured), you still have a 50/50 shot for heads on the next flip.

To put it another way, and try to circumvent the divide in our logic, even though HHHH only has 6.25% odds, HHHT has only 6.25% odds as well, so either event is equally as likely. This isn't exactly the way our problem works, but it may help to shed some light on the difference in perspective.

Also, apologies to the thread starter for going so off-topic. FWIW, I think I have come to agree with what appears to be the somewhat common sentiment that great generals are just too rare and valuable to be risked as common soldiers. I think after the game I'm in now (had to try one as a warlord (and name him Brennus )), I'm going with the military instructor option in my future heroic epic city every time, if I already have a unit experienced enough to build heroic epic. West point is [seemingly typically] too distant and potentially unnecessary to worry about requirements by the time your first great general pops (IMHO).

 

[kcbrett5]

I give up arguing with you guys. Frankly I have better things to do with my time.

Back on topic,
I have also found that using Great Generals as warlords is a waste. The only way to keep them alive is to use them in battles they are guaranteed to win. Otherwise they are going to die eventually. In the long run, I think I will be better off getting +2 experience on every unit I build in my Heroic Epic city. Or I might try using 2 of them in a 2nd high production city to get the +50% build bonus to make a 2nd military city.

 

[drkodos]

Each coin flip is independant of the others; flipping heads 50 times in a row doesn't make the coin emit magical control rays to guarantee that the next flp is heads.



On average, the house always wins at roulette. You can get lucky and happen to come out ahead in a single game, but in the long run you'll lose. No roulette system actually works, except for the being the house, which makes billions but isn't what people usually mean.

There are two slots that are neither red nor black on a roulette wheel, and that is why the odds are always in the house favor even on red/black bets.

This person you are arguing against has no clue how probabilities actually work in the real world nor have the demonstrated an understanding of the concept that previous results do not influence future results when it comes to coin flips. With the exclusion that it could end up on its third edge, EVERY FLIP is always 50/50, even if it came up a million times in a row heads. Of course I know you agree, but I am hoping they may actually do some homework and eventually realize the facts of the matter.

Now, if they have never studied statistics, their ignorance is somewhat understandable, but if they have, they must have recieved an F in the course.

I give up arguing with you guys. Frankly I have better things to do with my time.

I suggest some learning would be a most appropriate use of said time as I agree that continuing to argue when you are incorrect makes little sense, unless you enjoy the pain, or are unable to suffer embarassment regardless of how foolish and untenable your position.