Battle Mechanics

[robin74]

I figured I would let you answer it first Robin before I chimed in and you didn't let me down.

So didn't I? What I wrote isn't quite the same as what you propose (in your #4).
I tried to run some tests in a controlled environment. But I would need to sit for something like 40 hours taking notes before I would be confident of my observations. But right now, however insufficient data I have, it seems that it's most likely the size, and not the strength. But then again, my sample is so small that I might be wrong after all.
Cross shots - I just don't know. I saw it too often to just call them a bug But I wouldn't put the probability of it happening higher than 2-3%, and most likely it's smaller than 1%. Maybe it gets higher under some special circumstances I haven't tested. Personally, I very much doubt it.

I would even go so far as to propose cross shots will only happen in games that have witnessed the merge bug or whatever the merge bug has morphed into now that the same person can be king of multiple civs at the same time. You will know if all manner of strange occurrances are happening in the game already.

That's an interesting hypothesis I must say. Not that it's easy to check. You can never be quite sure that the merge bug did not happen in a given game.

[robin74]

As long as I ramble on about my tests
It seems to me that the odds of getting an attack start at about 2:1, even when you have just the minimal strength advantage. I would run battles for more than an hour keeping the strength difference at just 0.5-1, and the stronger side - even though just barely stronger - would take roughly 2/3 of all shots.
But then again, an hour or two of testing, that's not really such a big sample to be sure of anything.

[ShuShu62]

I believe the stack that shoots is a function of power and weather. I believe the stack that gets hit is a function of the number of units. That is why you can shield units behind a large stack of fortified units. But you can influence what shot you take by altering stances.

My guess is you run your tests with large armies. My experience comes from small one's. Troop losses from round to round are a higher percentage of total stroop strength in my sample sets.

At this point, I believe I have enough forum witnesses wathc my battles to back up my claims that my battles follow a consistent pattern: Trade shots till I flank or permanently shield heroic troops, then rarely get hit again. This is because once I stop changing stances, the momentum from the troop losses continues to tip the scales in my favor. You big stack folks take more hits in a single battle than I expect to take in an entire night... and my nights are busy... hehehe

It is interesting though, I come to the same 2:1 point from the oppositte perspective. I will stance dance to trade shots until I am outnumbered by 2:1. at that point, I agree there is some step function at work that makes it far more likely the other side will get a second shot.

[ShuShu62]

P.S. ... I also have witnesses. I spend more time asking teamates to stop dancing than to start it.

[robin74]

Actually, my most rigorous tests were done with small armies, of about 10-15 units in each, of various sizes, strengths and stances. I'd rather do it with a small army - I want each round to be done in exactly the same circumstances so that I don't have a mixture of various factors and can simply control for as few things as possible, so I need to rebuy any loss every round - with large armies losses would be large and I would soon run out of hammers.
Maybe I'll try to do some more battles next week to test specifically for size vs strength.

[ShuShu62]

... so I need to rebuy any loss every round -.....

Are you running both sides of the battle or just 1? If you are replacing losses on both sides, that would completely explain why you are not observing any momentum effects.

If you are replacing one side only on small armies, you should be seeing more momentum effects than I am.

[robin74]

Are you running both sides of the battle or just 1?

I did tests with a friend and we controlled both sides, replacing losses on both. I need to start with basics - given circumstances, what the probability of being selected depends on. It's easier for me to see when the circumstances don't change.
So what momentum effects would I want to observe exactly?

[ShuShu62]

I did tests with a friend and we controlled both sides, replacing losses on both...So what momentum effects would I want to observe exactly?

I feel much better now that we understand why you have considered me crazy all this time.

Actually, you are the reason momentum exists as a concept. It was your suggestion that memory of who took the prior shot did not exist (as your testing would clearly illustrate) so perhaps there was some other reason. Since I trust your testing rigor, but know that something alters attacks from striaght probabilities, I went off in search of possible explanations for my observations, and Momentum is what I came up with.


1) If you repeat your testing but alter stances (one side fortified/one side heroic) every round. you will find shot predictability to be 100% rather than the 75% a straight strength ratio would predict.

That means there is another variable, which you have been carefully controlling out. I call that variable... momentum. The test for momentum is a bit difficult.

Calibrate the amount of damage a single hit deals.
2.a) set one side up with that advantage.
repeat your basic test restocking both sides to starting level to verify the shot probability is straight strength probability
2.b) stock both sides to equal strength...
do not restock after first shot
restock after second shot
verify the odds of the side taking a second consecutive shot is greater than the odds found in step 2a. That difference is the Momentum difference (i.e. due to the relative change in strength between rounds)

Note --- I believe momentum becomes more predominant the larger the relative swing becomes (hence the reason dancing is 100% predictable rather than 90% predictable). So you might want to conduct 2a and 2b with all troops heroic, and all stacks so that they are not eliminated by a single hit. This is also why defending fortified troops can actually result in more shots than standard stance ones ---i.e. less momentum impact from round to round)


In the wiki, this corresponds to the concept of 'Ferocity of Attack'

[robin74]

1) If you repeat your testing but alter stances (one side fortified/one side heroic) every round. you will find shot predictability to be 100% rather than the 75% a straight strength ratio would predict.

I don't believe it's a straight strength ratio. Gamma does. I think it's some function of relative strengths, but not a linear function. I wouldn't expect 75% predictability.

Here's what I would expect with the regime you propose (ie equal strength armies, but one in fortified and the other in heroic, making them 4:1 strength, restocking every round): predictability at about 90% - and I would expect it do be exactly the same if we didn't change stances (but if we did restocking); stronger side is simply much more likely to take a shot. I don't expect see 100% predictability. With a long enough run I would expect to find no evidence of hypothesis that there is any correlation between the shots other than coming from the fact that one side is stronger (ie I would expect to see exactly the same pattern whether I change stances every round or don't change them at all - the chance a stronger side will attack shouldn't in my opinion depend on who took the previous shot). I understand that's where we disagree? You would expect full predictability under one scenario but not the other?

That means there is another variable, which you have been carefully controlling out. I call that variable... momentum.

Well, no. That may mean that the probability function is not a straight strength ratio.

to verify the shot probability is straight strength probability

But I don't believe that in the first place How am I to verify that if I don't think it's true?

Tell me this -
i) If I took armies of the same strength (accounting for weather and flanking and doing restocking) and did the dancing - which would essentially mean changing the relative strengths from 4:1 to 1:4 back and forth - then you believe this scenario to be fully predictable and would _always_ expect the stronger side to take a shot, right?
ii) You believe that the pattern would be different if it was always one side heroic to the other fortified (given restocking of losses), right?
iii) What do you think would happen if armies were not the same (normal) strength? If - say - one was twice stronger, so dancing would switch between 2:1 and 1:8? Would it be fully predictable too?

[ShuShu62]

.. I understand that's where we disagree? You would expect full predictability under one scenario but not the other?

Perfect statement.

.. But I don't believe that in the first place How am I to verify that if I don't think it's true?...

'Straight' strength ratio was a poor choice of words on my part as straight implies linear, when I really meant 'single variable' function F(x). You do believe it is a single variable just not a linear function. I believe it is F(x, y).

.. i) If I took armies of the same strength (accounting for weather and flanking and doing restocking) and did the dancing - which would essentially mean changing the relative strengths from 4:1 to 1:4 back and forth - then you believe this scenario to be fully predictable and would _always_ expect the stronger side to take a shot, right?
ii) You believe that the pattern would be different if it was always one side heroic to the other fortified (given restocking of losses), right?
iii) What do you think would happen if armies were not the same (normal) strength? If - say - one was twice stronger, so dancing would switch between 2:1 and 1:8? Would it be fully predictable too?

1) yes
2) yes
3) yes

Three gets more fun... I will step out on a limb and say yes even if the odds 32 times stronger so dancing went from 32:1 to 2:1...
although, I must say I haven't tested it because folks with a 32:1 advantage don't dance, and I don't hang around. But I have found that gong from X:1 to 2:1 will still get me the shot most of the time even though I am outmanned. (Note: in my world 32:1 to 2:1 increases my chances of getting a shot over 4:1 to 2:1).

[ShuShu62]

P.S. Yes I assumed you felt the function was linear as well. So I stand corrected on that point,

I am already wondering whether a single variable exponential step function could explain what I have observed (no).

But I do think a non-linear two variable function has more promise than the linear two variable assumption I have been operating on.

[robin74]

Perfect statement.

So what's your x and y? How these two scenarios would differ? Is taking the latest shot one of the variables?
My conjecture is F(x), where x is strength1/strength2 and F is non-linear and symmetric around 0.5 (possibly non-continuous and linear in each half).

As for the last scenario - if you say fully predictable, then tell me, what pattern should emerge? Let's stick to 2:1 v 1:8 case. Trading shots? What? Weaker side will take every fourth shot while heroic? I take it by "100% predictable" you mean a deterministic formula which allows you to tell exactly who will take the next shot, for sure. So how will it go?

[robin74]

although, I must say I haven't tested it because folks with a 32:1 advantage don't dance, and I don't hang around. But I have found that gong from X:1 to 2:1 will still get me the shot most of the time

Could you elaborate on this part? What do you mean "most of the time"? You get most shots at 2:1 disadvantage?

[Kinga]

Are you saying ShuShu that if my army is outgunned 2:1 with us both in the same stance, then I can swing the chances of me getting the shot by Fortifying and then dancing back to Heroic in order to come "storming back on to the field". And this sudden change from X:1 odds to 2:1 is seen by the game as a sudden increase of strength and somehow triggers the game to give you a shot?

[ShuShu62]

Kinga: exactly
Robin: Y used to be 'prior' shot, until you convinced me otherwise. Now Y is 'Momentum'


100% predictable meant 100% of the time. The reason I 'almost always' get the shot in the last scenario (not 100% predictable, so it is a 'Risky' tactic) rather than always get it, is because the odds are not all the way to 32:1 before i bring them to 2:1. If the other side is ungeneralled, and I can retreat my forces in between rounds, then I do always get the shot even though outnumbered by 2:1

[RobG]

I have been waiting for 2 weeks for my new alias from 2k to respond to this thread....I give up. So yes some of what I am responding to is weeks old now. It was ShuShu's observed combat mechanics copied below that I want to respond to:

1) The highest strength takes the first battle.
2) There is a bias against the side that took the prior shot taking another shot.
3) Units only shoot units in the same slots unless there are none... then they can hit anything
4) The odds of a unit shooting are prportional to the strength of the stack
5) The odds of being hit is proportionate to the size of the stack

Sorry ShuShu, I can confirm 1., 2., 3., and 4. are not full proof bets any more. I say any more, because it was about mid-Jan when I started noticing in many different battles of various sizes consistent irregularities (regularities then?). Most of the time old battle mechanics worked as usual, every battle would produce 1 or 2 rather random hits that made no sense at all. I think the designers wanted to add a degree of 'chaos theory' to battles to keep seasoned generals and/or favored armies on their toes.

And it isn't just me that has noticed this 'consistent irregularity'. Monkey Boy, as you refer to him in another thread, and I were both surprised at 2 hits that occurred during our battle you wrote about. Very irregular (but hugely beneficial to my side).

My primary point is that old theories of battle mechanics are not full proof anymore. I believe the developers through in a degree of randomness (though very small) to the math algorithm they use to calculate battles/battle mechanics. (Notably since mid-Jan)

[RobG]

....developers threw in.....(too early to be responding to threads)

[ShuShu62]

I have been waiting for 2 weeks for my new alias from 2k to respond to this thread....I give up. So yes some of what I am responding to is weeks old now. It was ShuShu's observed combat mechanics copied below that I want to respond to:

1) The highest strength takes the first battle.
2) There is a bias against the side that took the prior shot taking another shot.
3) Units only shoot units in the same slots unless there are none... then they can hit anything
4) The odds of a unit shooting are prportional to the strength of the stack
5) The odds of being hit is proportionate to the size of the stack

Sorry ShuShu, I can confirm 1., 2., 3., and 4. are not full proof bets any more. I say any more, because it was about mid-Jan when I started noticing in many different battles of various sizes consistent irregularities (regularities then?). Most of the time old battle mechanics worked as usual, every battle would produce 1 or 2 rather random hits that made no sense at all. I think the designers wanted to add a degree of 'chaos theory' to battles to keep seasoned generals and/or favored armies on their toes.

And it isn't just me that has noticed this 'consistent irregularity'. Monkey Boy, as you refer to him in another thread, and I were both surprised at 2 hits that occurred during our battle you wrote about. Very irregular (but hugely beneficial to my side).

My primary point is that old theories of battle mechanics are not full proof anymore. I believe the developers through in a degree of randomness (though very small) to the math algorithm they use to calculate battles/battle mechanics. (Notably since mid-Jan)

My current understading as documented in Battle School is

1) is no longer predictable
2) is a function of momentum, not who took the prior shot
3) is still true except in games that exhibit other bugs as well (e.g. disappering stacks, merged civs, another civs barbarians...)
4) is proportional to stack strength AND weather (weather is a very recent observation)

20 is the only real contentious claim. I am waiting for Robin to post that he has confirmed I am right that the odds are 100% rather than 90% when dancing (I know he is testing it, or planning to )

3) remember proportional means proportionalso those 5 heroic fodder troops in beneficial weather woul already take every 10th shot even with that fat stack of 100 melee sitting there. But I agree that they take far more shots than that now however, sniping principles still work. You fortify that melee stack and your other stacks will take more shots.

[Monqi]

3) is still true except in games that exhibit other bugs as well (e.g. disappering stacks, merged civs, another civs barbarians...)

Here the 77 bowmen are shooting at the 150 horsemen. I haven't noticed any bugs in the game so far. The thing about this that is the same as every other cross shot I've seen (don't remember about bugs in those games) is that it is by far the largest stack on the field on the receiving end. This cost me 42 horsemen so I'm glad it's rare

[ShuShu62]

Can I just say how much I appreciate that screenshot with a wonderful array of non-dancing stances (although my daughter would point out that fodder archer should be heroic) ... oh, and the opponent melee should be fortified

I have not seen a cross shot for a couple months. Roughly the same time I have been playing with kids. So one of the other theories is that cross shots can happen when you have open slots. In other words, an exposed flank protected by a single fodder can sometimes yield a cross shot.

I agree that every cross shot I have ever seen or have ever heard about has been against a very large stack and there may be something based on the disproportion of the largest stack and the largets stack protecting an open flank.

[ShuShu62]

ooooooohhhhh.... hehehehe... I think I got it. I bet the shot selection algorythm is rounding, so it thinks there are nor ranged units when you actually have one....


hehehe... cool.

[robin74]

So ShuShu, I now officially don't think anymore you're crazy
I still don't understand how the battle mechanics works, but I'm now convinced that who gets to attack is not just a function of strength.

[ShuShu62]

So ShuShu, I now officially don't think anymore you're crazy
...

You have no idea how happy that makes this crazy person feel.


It is always really cool when theoritcal science and emperical science come to an aggreement. I do hope that means some experimantal observations are coming our way.

[robin74]

I need to run some more tests and do whatever statistical analysis I can.
Though frankly, I'm not sure I would be able to say anything more sensible than that it doesn't work the way I thought.

[ShuShu62]

...
Here the 77 bowmen are shooting at the 150 horsemen. I haven't noticed any bugs in the game so far. The thing about this that is the same as every other cross shot I've seen (don't remember about bugs in those games) is that it is by far the largest stack on the field on the receiving end. This cost me 42 horsemen so I'm glad it's rare

Yes. Someone accused me of subscribing to an old wives tale for recommeninding he keep all of his flanking troops heroic rather than dancing. It was his fight and his troops, so I did not argue... but that won't stop me from posting here....

I would liken his desire to perpetauate fortified losses, to avoid the rare cross shot as believing in the '... boogy man will get you'. Yes cross shots will happen, but you will lose far more troops worrying about the occassional bad luck than you would if you occassionally lose a big pot rather than fold every other hand. (Robin/Random... you still think I'm a dancing fool?)

Anyway. I am sticking with my cross shot = rounding error theory (i.e. the computer did not think you had an archer when calculating target), but I have a new theory to add...

I believe there is a new (uh... December?) ... pre-contact phase to every battle. Melee takes time to close and contact the enemy, so ranged units have a higher probability to take initial shots. Naval and Mobile are also more likely early. In the future, I will be inclined to dance a few rounds until the shots become a little more predictable. I already advise caution on the first round. I now advise it for the first few rounds.

I would hold off using advanced tactics until after the first 5 minutes of battle. (NOTE: this would explain and have prevented ALL of the unexplained losses I have suffered over the last month)

[random user]

Well I'll admit I'm not an expert at close battles. I prefer them to be lopsided. I will clarify my statement to say that if you hold a 2:1 or 3:1 lead over your opponent in strength while you are both wholly heroic, then I don't think you should dance in such a way that you lose the arrow.

If the battle is so incredibly close that every single troop matters, then I do believe that strategically changing your stances (though I think I would still prefer to go heroic/normal vs heroic/fortified in most cases) could be useful. I do still think that offering to trade shots with your opponent by each of you alternating between heroic and fortified is still foolish.

[ShuShu62]

So I had a conversation with the baby Nazgul...apparrently found in a basket and raised to be a Nazgul... you don't want to have to compete against a civ with the baby nazgul in it... for example 10% of the baby nazgul's battle bonus was better than what the kids could get in the minigame...

where was I?

Oh, so the baby nazgul said that mobile gets more shots than the other types. I disagreed and said that all areas have equal base probability and then probabilities are modified by combat power. So our MAA army (500 Maa's 150 snow pnies... P.S. the Snow ponies were my daughter's, and 250 of those MAA's were rasied by my kids in 24 hours in a slow game... Pretty cool) would pound the other side's melee troops.

Our Melee troops proceeded to prove me right by continuously pounding the rifles. But then the eldest Nazgul added 100 knights to the mix. The Melee troops were still substantially more powerful than mobile troops. There weere still twice as many MAAs than other troop types. But we pounded mobile troops almost exclusively from that point on. The Baby Nazgul was also right There was obviously a bias towards mobile troops taking shots in this battle.

I am thus prompted to suggest another theory... The battlefield you are on impacts which troop type is more likely to take a shot. Mongolia is clearly biased to horses(mobile). If I am right Japan/Britain should be biased to Naval. I am not sure whic ones will be biased to ranged and which to Melee... but my guess is that this explains a lot of the 'Why are those troops being hit' questions I have had in the past.

[random user]

ShuShu62, I've been thinking about how to best phrase a sentiment I've had without being rude or annoying, and I was reading an article today which summed it up much better than I could:

Science is ultimately about establishing cause and effect. It’s not about guessing. You come up with a hypothesis—force x causes observation y—and then you do your best to prove that it’s wrong. If you can’t, you tentatively accept the possibility that your hypothesis might be right.

IMO if you really want to get down to the nuts and bolts, you should be trying to disprove your hypotheses rather than figuring out how to make rules which fit the observations you see.

I hope that didn't come off as rude or annoying.

(The article, if you are curious, is this one: http://blogs.discovermagazine.com/cr...are-all-wrong/ and has nothing at all to do in relationship to this thread.)

[ShuShu62]

No offense taken.

Of course the quote you posted is absolutely wrong on many levels, as well as the assumption that all paths to knowlege follow the same course that is implicit in it.

Have I written about the 'Engineer, Physist and Mathemtician sit down to take a test' joke we used to tell back in school?

Anyway...
Mathemeticians would choke on that quote because they attempt to prove theoretically (logic) rather than do their best to disprove empiraclly.

Physicists would choke on that quote too as theory is all about creating guesses that explain life, followed by focussed testing. My favorite story illustrating this procees is the 'glass half full' guess that led to the discovery of anti-matter. I can tell that story too if folks are curious because it was my senior thesis until I realized that I was too stupid to be a real physicist.

It is the engineers who feel the true path to knowlege is to grind it out with tons of emprical data and focus on disproving rather than proving.

all paths interact constructively together. Much like you, Robin and myself.

Think Tycho Brahe and Johannes Kepler. Neither would be famous today without the other.

Ultimately, this thread is a great thread for folks to make up their own minds. But my 'guesses' work very well... emperically speaking... hehehe

Ultimately, I agree with that article, and it is unfair to take that quote out of context. The article would say this thread is good science, not bad. What that article would say is that Robin published a falacious theory. Yes he had the numbers to back up his observations, but his tests assumed the conclusion (i.e. assumption... single variable function). Thus he missed that it is a multivariate function. However, Robin is practicing good science, because as soon as I presented a way to test his assumption (and my opposite one) he tested it and published that his initial assumption is wrong.

The best kept secret in the world is that public science is more about convincing people that you are right than it is about actually being right. The true scientists are less... uh... public. The difference between what that article is attacking as poor science and what I am doing is subtle, but profound.

What your article is attacking as bad science is 'The battlefield determines who takes the shot and here are the numbers to back it up'...

What I am doing is good science 'The numbers prove --- not disprove BTW --- that neither strength nor battle category are the only determining factor for who gets the shot. I am proposing battlefield as an explanation'

This is where Occam's Razor comes in. Whether my battlefield guess is right or wrong, it is now the explanation for why the Knights made the difference between hitting melee or hitting mobile because it is better than either of the prior explanations. If there is a better explanation (e.g. the enhanced Baby Nazgul explanation: There is always a bias to mobile taking shots when there is sufficient strength) It should show up here pretty quickly. In fact the only reason I haven't jumped on it is I have also wondered 'Why are boats getting shots here?' and 'Why are ranged getting Shots here?' leaving me with the impression that there is something more involved than the simpler theory.

One last thought, what you are implying is actually counter to what the author's article is getting at. Bad science occurs when folks are worried about publishing scientific facts rather than scientific knowlege. Facts are static agendas, knowlege is a dynamic guess providing an opportunity to take another step forward.

[ShuShu62]

Africa also seems to show a bias for mobile units in the one battle I have watched to date which would favor the enhanced Baby Nazgul explanation: There is always a bias to mobile taking shots when there is sufficient strength.

ALso interesting in that battle was that planes only hit planes and tanks only hit tanks. Again, my sample set is very limited (perhaps 20 rounds in one battle) But that is now something else I need to keep an eye on.

Final question... does anybody think the recent spate of missing civBucks means the battle algorythms have been changed/rebalanced in any way?

[spiph13]

Can someone explain why 360 M@As (90% bonus) hits 106 Phalanx (110% bonus, in snow, with the castle) and only kills 3 troops

but next round 37 Phalanx (45% bonus) hits the the same 360 stack and kills 4 M@As (no weather change, no other troop change)

Both sides swapping stances from Fortify to Hero appropriately?

[ShuShu62]

Damage is inflicted based on the total battle strength on the field.

My guess is that the defender's heroic strength + your fortified strength is roughly 25% stronger than your heroic strength + their fortified strength.

Which also means you are outmanned and will ultimately lose this battle without additional inputs of some sort. Especially if you are fighting against TroyD, in which case, you are doomed to watch your stack of 360 MAAs turn to dust over the next 4+ hours. But don't worry, the carnage will slow down as time progresses, so if you are stubborn, you can make Troy wait at least 8 hours for that victory... but wait he will and lose you will... so retreat now, especially since today is his birthday.

[spiph13]

Does that include weather bonus? Because overall we were pretty evenly matched in every weather except Fog.

I don't remember the exact totals.

[ShuShu62]

If they have gladiator you would be equally matched on normal stance, but they would inflict 40% more damage on the dancing rounds.

Himeji Castle or Militia would mean they would take fewer losses dispite equal damage being applied on the same rounds. (... looks over shoulder to see if Robin will correct him...)

Other than that, I think there had to be a strength difference.

[ShuShu62]

I got hit twice by naval units hitting my flanking snow ponies.

I verified that it was NOT a result of round off. (I replaced my naval fodder with a fortified stack and the second cross shot still occurred.).

The cross shots were also anomalous because they picked on the much smaller stack (25 vs 100). This game has shown no evidence of being buggy.

All theories seem outlandish at this point. Troy's theory that the computer holds a grudge and continues to hit the same unit is looking better all the time. I have a new outlandish theory. The stack getting hit was my dancing stack for momentum.

1) I get hit
2) I fortify my stacks
3) I set my smaller flanking stack to heroic to gain second shot
4) Instead, the heroic stack got hit

[Monqi]

here's my hypothesis for what it's worth:

step1: computer uses magic to decide which side shoots

step2: computer determines which section gets hit based on:
i) strength of opposing (shooting) sections
ii) total size of stacks in (target) sections
with i) being predominant

step3: computer determines which stack in the selected section gets hit based on stack size

step4: for the purposes of the graphical display computer determines which stack takes the shot, probably based on section being hit and strength of stacks on shooting side (but who cares, really, as the important factors - who gets hit and how hard - have already been determined).

this would mean cross shots don't really exist as anything other than a graphical display.

Predictions

1. cross shots are likely to hit the largest stack
2. chances of a cross shot are lower if your army is balanced
3. a strong section will take more hits from a balanced army

[ShuShu62]

...
step1: computer uses magic to decide which side shoots...

Agreed. or maybe just a straight random probability which means even largely out numbered armies sometimes get the first shot.

...
... ii) total size of stacks in (target) sections
with i) being predominant
...

That certainly would explain my last observation as my army was 150 snow pony 3 fodder (then 11 fodder). That theory would argue for a reasonably sized sturdy troop set to fortify, not to provide strength but to prevent hits to your flanking troops. I like this theory.

[ShuShu62]

I think I witnessed another factor last night.

I timed a second shot.
Waited for a weather change.
Had taken first shot while fortified
Set all my troops to heroic
And stole Himeji from the defenders, increasing the odds to 5-2 in our favor.

As soon as we built himeji, some one from the other side said 'Pop' and then 'too late'.
50 seconds later, I saw my heroic troops popped on the battle field.


I believe my opponents and I were playing with different clocks, and that they actually took the shot BEFORE we built himeji, and possibly before all my troops were heroic.

I was suffering horrible lag from maintaining a live minigame bonus across 8 concurrent battles.

I am now convinced that my tendency to do things 'at the last second' may actually account for some unpredictable behaviours if the shot was already resolved before my action although the screen indicated the shot had yet to occur.


It is something I am going to worry about now at least.

[StarShard]

The next round's attack should be calculated based on the numbers present in the current round. This would help prevent the "winners keep on winning" scenarios and make battles more balanced.

[ShuShu62]

8 hour battles aren't long enough for you?

Thank heavens for the winners keep winning or battles would never end. I think the old paradigm where the army with the greater strength pushed the bar whether they took the shot or not avoided that problem, but that is not where we are today.

I understand how the current battle structure allows the underdog to stay in the fight longer as long as they can keep getting shots rather than the old paradigm where the larger army would carry the field merely by being larger. At times, I like the difference as I am able to win battles much like the Army of Virgina by siezing initiative despite smaller gross numbers, but at other times it feels I have been dropped into the trench warfare of WWI. The problem becomes more pronounced the larger the armies become.

As the small teamsizes and spam have rendered the games to NBA affairs where only the 4th quarter matters (hollywood doubling already doubled eras), this massive army effect is amplified.