#what is up with castle defense?

1 messages · Page 1 of 1 (latest)

gentle tangle
#

Lost 9 attacking 7 my castle attacker 5 left over. Balanced blitz. Attacker lost 3. ?????Castle no longer have attack over wall. Dice rolls seem pretty screwy lately

glass pollen
#

It's pretty normal to have extreme outcomes with low numbers.

undone forge
#

I love that response. I also love the other common answer, “it’s pretty normal to have some extreme outcomes in battles over 70 troops as those battles are resolved by extrapolating from smaller battles.” So… to summarize…. It’s pretty normal to just have extreme results.

turbid quail
#

its not as simple as "its extreme with low numbers, its extreme with high numbers, its always extreme!" that is kinda how it ends up working, but it shouldnt.

from a math standpoint its expected that low troop counts should be high variance and you should expect some extreme outcomes. one good or bad TR roll can significantly change what becomes more likely to happen.

but for large stacks, going even on a cap roll is statistically impossible. you might see a god roll like that 1 in 1000 tries doing a 600v300cap but it happens like .... as often as you lose double, or go even, or go positive. its expected that you lose WAY more WAY more often with large stacks. but BB is broken.

BB is only guaranteeing the attack is a win, and says nothing about troop losses. troop losses really ought to follow TR outcomes, if anything

sour ivy
turbid quail
#

ok, maybe not "nothing." the takeaway for me is the distribution of what can happen really really really feels off.

for like a 100v100 attack where you could lose anything from 50 to 150 then you would expect to lose 100 most of the time, and see some 75 or 125 not too infrequently. but 50 or 150 would be INCREDIBLY rare. appears like the chances of any of those is almost equal though

#

aand of course for caps its even more apparent somethings super wrong because positive rolls should be practically impossible, yet you see them almost as often as losing 2x

#

its in the realm of believable to lose hard, say 8 troops, in a small 10v10 because it would just be 4 unlucky TR rolls under normal dice rules. so like ya, super unlucky, but fair enough. to lose 50 troops in a 100v100cap though. the same dice rules make that effectively impossible and yet rolls like that happen every game.

sour ivy
#

Yeah I agree. There's 36 possible dice combinations and the defender wins 21 of them. So in caps the defender should win roughly 58% of dice rolls. I'm sure there's a better way of doing that math but that should be a good approximation.

But put that aside, is it really a bad thing that the attacker loses less troops hitting a cap?

turbid quail
#

imo. yes. the whole POINT of a cap is to be a heavily defended point where the defender gets an advantage. that advantage should absolutely mirror the true statistical effects of the defender getting an extra dice.

#

for that matter, a normal 100v100 should also mirror the statistical attackers advantage better. theres some questionable non caps attacks too lol

turbid quail
#

very scientific visulization

image.png
sour ivy
#

Ok but what if you look at it like this...
The problem with caps is that if you play it like it was originally designed (basically TR) it is basically never worth killing someone if they just stack on cap. That's why caps never really gained popularity before, it was just a stalemate fiesta. But with the way it is now we get the best of both worlds. Ill explain with an example:

For this example let’s look at a 100 stack for the defender. If that stack is not on a capital it only takes 103 troops for the attacker to get a 100% roll on that stack. They will probably lose 75-80 troops.

If that 100 stack is on a capital that same 103 for the attacker has 0% odds of success. In fact it takes 171 troops for the attacker to get just 50% odds for the attack. It takes 200 to get a 100% roll. So you can be outnumbered almost 2:1 and still win if you’re on a capital. If that’s not an advantage for the defender idk what is.
Even if the attacker gets a really good roll they’re probably going to be ~even which is way better than you would expect for a roll on a stack off cap. And most of the time they will still lose 10-20% more than the defender. Sometimes up to 75% more than the defender.

So now we have a situation where it still takes a massive troop advantage in order to take a capital, but the player who has that troop advantage is able to hit the capital without being heavily punished and therefore suiciding their own game. Now caps games can be progressed.

turbid quail
#

well the explanation is correct id say, but the premise i dont agree on. i dont think it is ever worth attacking someones cap if they just stack on it infinitely, nor do i think it should be worth. the way to victory there is preventing them from getting cards and building up the troops to be able to kill them despite being on cap, or to force them off cap to hit troops if they want a card, right?

#

the game shouldnt be decided by crossing your fingers and throwing yourself at someones cap and hoping for the best

#

im alll for an attacker getting a really good roll but a good roll on a cap should be more like losing 25% to 50% more, not going even.

#

and that good roll should happen like 10% of the time compared to a 'normal' roll

#

whatever the losses actually end up being the god rolls for oeither attacker or defender shouldnt be happening every other cap hit

sour ivy
#

I think your logic is flawed. You want the roll to be even and repeatable but then a good roll only happens 10% of the time? It should either never or always happen in my opinion. The more troops you are dealing with the less deviation there should be in the troop losses.

I think anytime you have something where it is never "worth" hitting is bad for the game. It just inevitably produces stalemates, bot outs, and suicides.

turbid quail
#

no, id like to see the statistical expected outcome to be the most likely. and a smooth decreasing curve where chances of extreme outcomes are possible, but very unlikely. standard bell curve.

#

it feels like how it works currently, the distribution is basically a straight line where getting the expected outcome is just as likely as getting THE worst or best outcomes

real moat
#

OK….GETTING TIRED OF THE RIGGED DICE..! How can you lose a battle of 14 against 2 and lose all armies and they LOSE NONE??!! It has happened to me too many to count, and to be fair, I’ve seen it happen against other players as well. The probability of losing all to none with a overwhelming Army is statistically IMPOSSIBLE! I see others write-up’s regarding the same deal. Whoever is behind the coding needs to address this issue and explain this phenomenon that is pissing too many players. PLEASE LOOK INTO TO THIS AND EXPLAIN TO EVERYONE !