conquer and hold, or retake?

    This site uses cookies. By continuing to browse this site, you are agreeing to our Cookie Policy.

    • conquer and hold, or retake?

      Ok so i was thought that you take a city, you hold it (with 1, 2, 3 units, depending on personal style) and then you go away with the smoke.

      I'm in a game now where noobs and abandoners are the vast majority. I changed coalition 3 times, and ended-up with a couple of guys who expand very, very fast. Their style is not 'once the smoke goes away you move your units out', but it's 'take city and move on'. If and when rogue takes a city, you use air to clear the rogue unit, send an inf, retake and go away again. Immediately.

      I'm puzzled. It might be because he's close to his homeland, so the period and chances of insurrection are lower. But does this strategy make any sense?

      Thank you. Ciao. s
    • Insurgency isn't a sure thing, and after the first day it's fairly improbable. Expected value analysis might justify moving on if your goal is rapid expansion, knowing that 4/5 times you'll get another city faster, whereas 1/5 times you'd be penalized with having to do twice the work. Overall that adds up to 4/5*0.5 +1/5*2=4/5, which is less than 5/5 which is the default of waiting. So doubling your speed 4/5 times at the cost of being twice as slow once out of 5 times when you lose a city to rogues does pay expected dividends. Naturally this calculation is quite simplistic, I'm just showing that this question is amenable to rigorous probabilistic analysis and the answers comes down to whether the risk of the cost of turning outweighs the chance of the reward of grabbing more cities faster.
    • user8394273 wrote:

      Overall that adds up to 4/5*0.5 +1/5*2=4/5, which is less than 5/5 which is the default of waiting.
      i don't get the mats for multiple reasons... example: i'm now Indonesia, and am conquering Europe. If i leave a city empty, the tooltip says that the insurrection %ge is 50. That changes this quite radically, no?

      On the opposite side, i think if you don't stop the speed-up factor is much higher. I'm currently capturing 19 cities with an inf unit each. Now, if all of those 19 were out capturing a city each, and moving on... (instead of waiting 2 perhaps 3 days in the same city)... but if then 10 of them need to go back, and then 5 of them need to go back, and then 2 of them need to go back... i'm lost with the formula, but the one above doesn't sound right :)
    • Ok, let's make a BIG simplifying assumption, that the probability of revolt is 50%, then 0% after the first day (it actually drops to like 20-30, but that makes calculations a bit more annoying)

      Strategy of full forward: Day 1 we take 20 cities, on day 2 we take 20 more but 10 of yesterday's revolt. We have 30 cities, 10 of which will not revolt. Day 3 we take 20 more cities, but 10 of day 3 revolt. We now have 40 cities, 20 of which will not revolt. Etc. Each day we take 20, and 10 revolt, therefore the rate of conquering is 10 cities per day with that strategy.


      Strategy of stay: Day 1 we take 20 cities, day 2 we take none, day 3 we take 20... Also rate of conquering 10 cities per day. Under these assumptions, we calculate that it is equal.

      Now we can extrapolate to a situation where P(revolt | day 1)=0.5 and P(revolt |day 2=0.2) and do the necessary calculations.

      Strategy of full forward: The yield of 20 cities is 8, because 10 revolt on day 1 and 2 revolt on day 2 after being conquered. So we get a ratio of 8 cities per day.

      Strategy of stay: We wait 3 days, so we conquer 20/3 =6.66666 per day.

      => Real life strategy of full forward is better than stay til smoke clears.

      What about stay for one day only?

      Then we have a base rate of 10 per day, and we are losing 2 per day too, so also 8. This is tied with full forward.

      Conclusion: On first day it doesn't matter what u, because 50% favours staying and leaving equally. After that, you SHOULD leave mathematically.
    • user8394273 wrote:

      Ok, let's make a BIG simplifying assumption, that the probability of revolt is 50%, then 0% after the first day (it actually drops to like 20-30, but that makes calculations a bit more annoying)

      Strategy of full forward: Day 1 we take 20 cities, on day 2 we take 20 more but 10 of yesterday's revolt. We have 30 cities, 10 of which will not revolt. Day 3 we take 20 more cities, but 10 of day 3 revolt. We now have 40 cities, 20 of which will not revolt. Etc. Each day we take 20, and 10 revolt, therefore the rate of conquering is 10 cities per day with that strategy.


      Strategy of stay: Day 1 we take 20 cities, day 2 we take none, day 3 we take 20... Also rate of conquering 10 cities per day. Under these assumptions, we calculate that it is equal.

      Now we can extrapolate to a situation where P(revolt | day 1)=0.5 and P(revolt |day 2=0.2) and do the necessary calculations.

      Strategy of full forward: The yield of 20 cities is 8, because 10 revolt on day 1 and 2 revolt on day 2 after being conquered. So we get a ratio of 8 cities per day.

      Strategy of stay: We wait 3 days, so we conquer 20/3 =6.66666 per day.

      => Real life strategy of full forward is better than stay til smoke clears.

      What about stay for one day only?

      Then we have a base rate of 10 per day, and we are losing 2 per day too, so also 8. This is tied with full forward.

      Conclusion: On first day it doesn't matter what u, because 50% favours staying and leaving equally. After that, you SHOULD leave mathematically.
      Improve your math skills before making big statements on a forum...
      "but that makes calculations annoying so let's just remove it"
      "Le patriotisme, c'est aimer son pays. Le nationalisme, c'est détester celui des autres."-Charles De Gaulle, Leader of Free France in World War 2.
      English: "Patriotism is to love your country. Nationalism is hating that of others."
    • If you continue reading the post I then reintroduce that variable, and produce a final outcome which takes into consideration that fact. My initial computation was merely a demonstration of the method. If you have read any serious math papers you will see they start by isolating the crux of the problem and then reintroducing the meaningful, but methodologically trivial complications. If you have an understanding of expected value under conditional probability that I do not, and reach another conclusion, I invite you to share it with us.
    • user8394273 wrote:

      If you continue reading the post I then reintroduce that variable, and produce a final outcome which takes into consideration that fact. My initial computation was merely a demonstration of the method. If you have read any serious math papers you will see they start by isolating the crux of the problem and then reintroducing the meaningful, but methodologically trivial complications. If you have an understanding of expected value under conditional probability that I do not, and reach another conclusion, I invite you to share it with us.
      Uh, no, sorry too late, you exceeded ewac 30 second attention span before you reintroduced it. I'm going to have to call a fowl.


      "PHEASANT!!!"
      *** The Creator of Zombie Farming ***
      The KING of CoN News!!!
      The "Get off my lawn!" cranky CoN Forums Poster - not affiliated with Dorado in any way


      "Death comes to us all. Shall I deal you in?" - DoD
    • Dealer of Death wrote:

      user8394273 wrote:

      Blah blah blah variables reintroduced after writing a mathematically "accurate" essay and percentages...
      Uh, no, sorry too late, you exceeded ewac 30 second attention span before you reintroduced it. I'm going to have to call a fowl.[/b][/b]

      "PHEASANT!!!"
      "Le patriotisme, c'est aimer son pays. Le nationalisme, c'est détester celui des autres."-Charles De Gaulle, Leader of Free France in World War 2.
      English: "Patriotism is to love your country. Nationalism is hating that of others."
    • IMO you don't really have to capture a city if you can't hold it, only time where it really makes a difference is when it comes to capturing enemy homeland/annexed cities so they can't recruit shit, otherwise you can simply ignore them and just capture the surrounding provinces cuz having to go back and kill insurgents and then move in is rather annoying
      I am The Baseline for opinions
    • There is a trick to it, You can observe and find the time of rebel spawning on that specific day. if you lucky, you could skip a day without need to guard it or no need to guard the city at all.

      It work like this, You notice that rebel spawn on occupied cities else where in the world at the perfectly same time at 9.00, now any city capture after 9.00 won't spawn rebel at least until next day, which will increase moral and even lower the rebel chance.

      How ever the "guard until smoke go aways" isn't a bad a advise, it is perfect for newbie to not complicate things too much.
      This post was made by Leader of the Church of ROAD
    • Hello; this is my first post back since my tour de force magnum opus hiatus, so I'll give you my perspective from my last three best games then. As the other posters have explained, the window of opportunity for rogue appearance is only at midnight, providing the player with a significant advantage against them. There are two options from here. One can either use units to hold it, or abandon the city and come back to it, with timing providing additional strategic variation within these choices. It seems the teammates you had focused on rapid expansion with the second method, but I contend that the former approach is superior, while the latter is automatically disincentivized by a calculatable (although not necessarily problematic) margin.

      If you capture a city, leave it, allow rogues to spawn, and keep coming back with air and infantry, you're just wasting your time, hp, and resources compared to delaying capture by sitting outside the city and then capturing after midnight or garrisoning it right before midnight. You end up having to pull units back to retake the city, having any structures that might have been beneficial taking higher damage instead, losing production during that time, and may have to worry about needing several missions to clear the city. In contrast, capturing a city and garrisoning it (2-3 Inf early-game, 1 mid-late with airpower) lets you collect that production, keep structures with less damage, send units on once the city is cleared without needing to pull any back from the front, and you can easily tell the strength of the rogues. have been better off. The main exception depends on the rate of insurgency, which might make it reasonable to leave a city with a small chance of revolting.

      Finally, the probabilistic expected value analysis applied by user8394273 was partially flawed because it failed to fully address the effects of recalling units, the ability to simultaneously push forward while holding some units in reserve, and the benefits of holding a city over losing and then retaking it. In my implementations of a hybrid rapid expansion strategy, I advanced rapidly while retaining almost every city I captured. There is flexibility to occasionally post a unit on the frontline for show of force and impact of presence before recalling it, but the former approach is unilaterally superior when this capacity is accounted for. Hope that helps!