Random Factor in Combat

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

    • Random Factor in Combat

      Precursor: This is going to be a very long post so I will break it up into the research and rant sections.

      I am trying to wrap up a 22 player game, and have had several battles over the last few days with the #2 player. I am making this thread both to air some frustration over and hopefully get some legitmate research started on the combate formula, particularly in determining the randomness factor's range.

      RANT SECTION

      While I understand some randomness must be involved, here are some of the things that have occurred in my battles during this end phase. I will list my units first with level in (x), and indicate [A] for attacker and [D] for defender, and any notes where confusion may occur as to if there is a defender at all. This will be followed by the result, and my issue with the result. Please note each of these is a single attack phase.

      1) 12 Interceptors (5) and 19 tac bombers (5) [A] vs 6 commandos [A] - Commandos were marching towards a province, thus the [A]. Result: 3 commandos killed vs 2 Interceptors AND two bombers.

      Issue: Commandos have an anti-air ability of 1, so they had a total anti air capability of 6. I was not sure of health totals before the fight, so it's possible both my interceptors and bombers were low enough that one of each could be lost. To lose a second of EACH though? That's a minimum damage of 45 when the base was SIX!

      2) 3 Medium Tanks (max) [A] vs 1 Medium Tank (2) [D] - My medium tanks were merged at low health for a total of 52 and change HP. Result: His MT took 5 damage, my MTs took 26 damage.

      Issue: Despite my MT stack's damage after SDBE should have been in the neighborhood of 20, I have seen enough attacks do 10-30% damage that this did not surprise me. What DID surprise me was the lone defending tank with a defense of 5 vs tanks did over 500% damage!

      3) 11 Interceptor (max) [A] vs 5 Interceptor (4) [A] - His Interceptors were in patrol formation, which I would assume warrants an [A]. They have the same attack/defense rating, so it shouldn't matter anyway. Result: 0 Interceptors of his killed, SEVEN Interceptors of mine were lost.

      Issue: Do I really even have to type this one out? My Interceptors had a strength rating of 75+ while he had a strength rating of around 40. I somehow fail to do less than 1/3 damage for what has seemed like the billionth time in this game, while my opponent somehow does another absurdly high damage percentage of over 300%.

      Again, I absolutely understand and acknowledge randomness must be involved in the battles, but to the extent I'm seeing (and to the frequency as well), I'm starting to think that attackers are being overly handicapped while defenders are being overly favored, especially in battles where the unit has the same attack rating as defense, this makes no sense, and it definitely does not make sense with the ratios I'm seeing (that commando battle was at minimum a 750% damage value).

      RESEARCH SECTION

      What I would like to do is start gathering data to try to come up with some semblance of a combat formula and help determine the random factor windows. If anyone would be kind enough to post individual battle phases ( which occur every hour, or 15 minutes if involving planes in patrol mode) with the following information:

      Your Unit Information - type, count, level, attacker/defender status, starting HP, ending HP, SDBE, morale.
      Their Unit Information - same as your info list.
      Battle Conditions - Terrain, Stack status (moving, patrol, stationary), fort level if applicable, anything else I may have missed that would apply.

      I will start compiling data gathered every other day or so when I get the chance, and let you know once I feel I have enough results to give a ballpark range on the formula.

      Thanks in advance to anyone who has gotten this far and is still willing to help me out with this endeavor.

    • Well, I have definitly not done any consistent testing or research on this so the following is mostly guesswork...

      First of all, in your first example I'm not surprised at all... You totally neglected the statebased efficiency by overstacking enormously, I would guess you were working on a 50-60% efficiency. Further you don't mention the terrain. Commandoes in mountains for instance would have anti-air of 2, double what you assume... A commando have 25HP while a ceptor has 20... In other words, your basic ceptor would have an attack value of less than 1 (due to state based efficiency) while the commando would have a value of anything from 1 to 2 depending on the terrain. In other words, the commando was MORE likely to make the damage roll than each ceptor. When it comes to the bombers I'm guessing this is where you got damage done to them as here the odds vere definitly in your favor even in mountains.

      My basic understanding of how combat plays out is that it's more unit-based, not stack based. Each unit in your stack does one attack where your value (attack/defence based on territory according to state-based efficiency (health + overstacking)) is compared to your opponents and a chance to do damage is calculated. IF you are successfull in this roll, the other side will take damage. This damage does not necessarily have anything to do with your attack number but is rather a set damage value (this is pure guesswork, but fits in with what I've seen so far).

      In other words, your interceptors had a very low chance of actually inflicting damage and the damage they would inflict would probably be low as they're not supposed to be used against ground troops. On the other hand, a commando is a specialist unit witch by all accounts seem to have a pretty high damage output. Against your ceptors he would have a likely chance to do damage.

      Did this make any sense? Like I said, I'm far from certain about the mechanics myself, but this is what I've concluded through my own experiences...

    • Quasi-duck wrote:

      You should never use int for ground attack...
      Thank you for your very thorough and constructive feedback, Quasi. I paired my interceptors as a deterrent from him trying to hit my bombers with his own interceptors, which is evident that he owned based on the other scenarios. I'm aware that interceptors do very little damage to ground units.

      Do you have anything to say regarding the random factor in combat? Have you noticed any unusually high or low results after individual battle phases?

    • SpiritNOR wrote:

      Well, I have definitly not done any consistent testing or research on this so the following is mostly guesswork...

      First of all, in your first example I'm not surprised at all... You totally neglected the statebased efficiency by overstacking enormously, I would guess you were working on a 50-60% efficiency. Further you don't mention the terrain. Commandoes in mountains for instance would have anti-air of 2, double what you assume... A commando have 25HP while a ceptor has 20... In other words, your basic ceptor would have an attack value of less than 1 (due to state based efficiency) while the commando would have a value of anything from 1 to 2 depending on the terrain. In other words, the commando was MORE likely to make the damage roll than each ceptor. When it comes to the bombers I'm guessing this is where you got damage done to them as here the odds vere definitly in your favor even in mountains.
      I appreciate the effort, SpiritNOR. I'll respond to each portion as best as I can:

      With my first example, that was the only one of the three I gave where I was completely OK with my own troops' damage output, you'll notice in my Issue segment I never mention them at all. My issue there was strictly with the egregious amount of damage done by the commandos. I am well aware of SDBE and have found it best with planes to use larger stacks rather than splitting them because generally the loss of excess damage is mitigated by only taking one damage cycle instead of several if you split them up. As for the terrain, it was plains, so their defense was definitely a total of 6. As mentioned to quasi, the Interceptors were there as a counterattack deterrent and not meant to be my main damage dealer.



      SpiritNOR wrote:

      My basic understanding of how combat plays out is that it's more unit-based, not stack based. Each unit in your stack does one attack where your value (attack/defence based on territory according to state-based efficiency (health + overstacking)) is compared to your opponents and a chance to do damage is calculated. IF you are successfull in this roll, the other side will take damage. This damage does not necessarily have anything to do with your attack number but is rather a set damage value (this is pure guesswork, but fits in with what I've seen so far).



      SpiritNOR wrote:

      In other words, your interceptors had a very low chance of actually inflicting damage and the damage they would inflict would probably be low as they're not supposed to be used against ground troops. On the other hand, a commando is a specialist unit witch by all accounts seem to have a pretty high damage output. Against your ceptors he would have a likely chance to do damage.


      I'm not sure that the unit based damage would make much of a difference since you'd be multiplying that damage roll by the number of units in the stack (since your guess already accounts for SDBE). If anything, if each unit got their own roll, I would expect far more battles with stacks involved where the damage each round is closer to the expected value as opposed to right now where you can have such ridiculously high variables.

      Regarding what you said about damage dealt not necessarily being about strength and being more based on the unit itself (commandos being specialty units so would be likely to damage interceptors in particular), I disagree - what would be the point of having strength values vs other units in the first place? Also, Commandos can be built on day 8 in the 22 person game, whereas leveled bombers are day 36/40. Lastly, you say Interceptors are not meant for ground unit attacks (which is correct, per their strength level) - the same could be said for commandos defense value against air assault, so I'm not sure I understand where you think commandos would be likely to do so much damage.

      There are still two other scenarios with rather unusual results, not just the first one. Also, would you be willing to gather some data as I mention at the end of my OP? I'm very curious to see what the actual boundary ranges are for the random variable in the combat phase.

    • nemur, thanks for the heads up. I have been doing tests regarding this issue. Mainly militia vs militia and inf vs inf to measure the variance of damage per round. I will give you some hint. Variance is VERY high. The AVERAGE does not correspond to the expected average given attacker/defender values and SBDE. The per round damage is always integral, for example damage can go from 8.8 to 7.8, but not 7.7.

    • Javman, you're absolutely correct regarding two things:

      Poor mathematics in combat mechanics - I've noticed this in many of my latest battles. I believe this is a fairly recent development, and needs to be reviewed.

      Poor support and vague answers from other users - You and I share a thirst for knowledge regarding combat mechanics, however there's always the random user who throws in the question of why did you attack with that unit? or why did you have that many units stacked together? or even giving you random advice regarding life.

    • SpiritNOR wrote:

      Meaning the random factor is less for the attacking side?
      No, I mean for units where the defender has a strength advantage, like infantry, the average attacker power, given variance, is greater than expected. The average defender power, given variance, is less than expected. The defender still has an advantage, on average, but it isn't as great as expected based on stats.

    • Ahh random the very core of chaos.

      We talk about this before why large stacks have low effectiveness while smaller forces have high if not full effectiveness.
      I haven't play for a while so my explenation will be coming from past experience and general knowledge of it.
      1. Commandos have anti-air 1 but being on the "favorable" in the exchange, I remember back them we would argue the planes are grounded because the commandos are "attacking" and the planes are "defending". This would make a lot of sense in real life.
      2. The MT fight is definitely weird tanks have flat stats in both offence and defence, but in this example I presume the merge group is individually weaker, meaning the lone tank is picking off single targets since the group is already low to begin with. With low health comes low morale and effectiveness so attacking with such a group will yield poor results especially if used in the attack, while the lone wolf have a province bonus due to being the defender.
      3. This is the classic big number low power explenation, seriously I have no idea how this went down.

      Overall due to combat being random and chaotic to the outcomes and all, I recomend to bring "fresh" units and let those who fought rest a little. That's just me giving advice heck this stuff is way technical for me, I SAY POUND THEM WITH EXTREME ARTILLERY FIRE BEFORE GOING IN.

      Oh and yes defenders have the advantage, every unit have a stat for attacking and another for defending, most of the time when your attacking things will not go as expected so always fight at an advantage and don't fight fairly, fighting fairly is the last thing you will expect.
      "Victory needs no explenation, defeat allows none"
      -imperium thought of the day

    • SpiritNOR wrote:

      Aha... Interesting... I have actually wondered a few times why my infantry seem to be doing so much worse defending than I expected them to.
      Yep on average they don't do quite as well as expected, but regardless of that, if you fight 100 battles with 5 INF of same level on each side, in a certain percentage of those battles (maybe 10% or so), the defender will lose due to variance. Likewise, in some of them both will lose the last of their HP in the last round.

    • infantry have a more higher defense stat compare to infantry in the offence, not to mention province defense and terrain. when 5 infantry fight against another 5 on the defense generally the defenders should come out victorious but with severe damage, this is of course with the same level of tech but even then defending infantry are more better than attacking infantry. of course there's morale and effectiveness, the randomness would come from mixing units with variable levels of morale and effectiveness resulting into a lackluster unit.

      i also have not mentioned how effective a unit is against another, tanks last longer and hits harder than infantry for example, think of it as a weapon effectiveness triangle but with a lot of exemptions, infantry defending would lose against attacking tanks on their own.

      I personally use units with 100% conditions across the board to reduce the randomness, using whole units are better than those a lot of unknowns of unit effectiveness, variable morale and the like.
      "Victory needs no explenation, defeat allows none"
      -imperium thought of the day