FAQ detail: which units are the best for a fight ?

[Lo2b]

Hello !

In the help section of the game, the FAQ states as follows:

"Alliance Contribution:
Each alliance member (up to the contribution limit) will bring the 4 best units available to a fight."

But it doesn't explain in full details which are the best units available for a fight. This has an influence in reducing the cost in $ of a fight, and therefore in the choice of purchasing units. Let's take an example.

Let's say I am at level 20, with an alliance size above a 100. I can therefore bring 400 units into combat. I own 410 fighting units, so 10 units won't enter the fight. The 410 fighting units are composed as follows.
390 units have an individual attack value of 2 and above.
10 units are light gunners (attack value: 1 / defense value : 0 / $: 600 each)
10 units are engineers (attack value: 1 / defense value : 4 / $: 4000 each)

The 390 units above 2 are the best and will fight. Out of the 20 remaining with the same attack value, only 10 will enter the fight.

My question is: which of them ? The light gunners or the engineers ? How is the answer calculated ?

Can some developpers answer this question please ? Or maybe some guru around already know about it ?

Thanks in advance, happy fighting,

Lo2b
Modern war ID: 931-508-075

[Mudphud]

In my experience, it depends. The best way to check is to actually attack someone and then check the battle report to see which units were actually used. The units with the lowest attack scores are listed at the bottom of your unit list, just before your losses. Hope this helps.

In past fights, sometimes the algorithm has chosen the gunners and other times the engineers, but never a mix of both.

[Lo2b]

Thanks Mudphud. I made the same observations as you did. But the point is how to predict it, upon which you don't buy the same units.

[nighthunter]

My experience is during attack, light gunners will be brought to the field last, after scouts and engineers, which kinda suck cause they have the same attack score and die equally fast except light gunners are much cheaper. On the other hand, light gunner doesn't contribute to defense so you still have to buy scouts/engineers/medics.

[Agent Orange]

I think it may depend on your level and the number of units you can bring into a fight. In the lower levels where you are limited I would want the most powerful since your attack average will be higher with a few low value shields just in case.

In the high levels you probably want a mix, I've been stalking up on ambulances and spectres to beef up my defense. I don't buy gold or cash so rely on cashflow and valor to buy units. I like the Super Hornets and Stealth Frigates but they are quite vulnerable to loss contrary to what the game says.

The other thing I notice is in the lower levels you attack/defense losses aren't to high if anything but up past 50-60 I find my attack losses were so high that I only attack or raid for the heck of it as it costs more than I gain. As with the rest of the game balancing attack and defense seems to be the trick.

[Wildfire]

I like the Super Hornets and Stealth Frigates but they are quite vulnerable to loss contrary to what the game says.

I find that with all valor units, I seem to loose low casualty ones at an equivalent rate to medium paid for units. I also seem to be dropping frigates at an alarming rate at the minute which is surprising given their very low status, but statistically it's still too small a sample to really mean anything other than bad random luck.

[osubuckeye]

in my experience the engineers will be picked first. This also happens with jet fighters (3/1) vs humvees (3/3).

Given the cost of the units being compared, i think it sucks beyond words that the game is "too dumb" to send an engineer to attack something rather than a gunner. Seriously WTF?! lol

[nighthunter]

Sending Engineers first make sense because they are "strong against ground", but for fighter and Humvee or scout and gunner doesn't

[frenda]

The "attacks" are totally random, so you don't know if you're attacking someone with heavy sea or ground or whatever, unless you want to spend time evaluating his assets. It's probably just a straight-forward comparison of attack and def numbers based on units and level and allies, with a roll of the dice thrown in to make it more random. I think we're giving the game a lot more credit than it deserves in this case.

[Dover]

I think we're giving the game a lot more credit than it deserves in this case.

I love that line and totally agree with you.

[FearlessHero7]

How do you reset your modern war game plz help

[frenda]

Thanks, Dover. I can't imagine the developers of the game taking into account relative values of atk/def of each entity and comparing it to each and every entity of the attacked units. I do that for a living, so I know how hard and tedious it is.

[osubuckeye]

I don't think nighthawk was giving credit to the game. And for my part, I am definitely NOT giving this game too much credit :P

But at least nighthawk's comment explains why, even though it is totally senseless in my mind, the ai would choose engineers over gunners. I don't want it to, but I guess if they have an edge vs whatever ground unit the other army may be fielding, then it makes sense to the ai.

so just to clarify, we're not of the mindset that the ai sits and analyzes the enemy army to figure out if engineers makes more sense than gunners. But that it just makes that decision based on their attack being the same but engineers having edge vs any supposed ground enemy units.

Which still sucks if you're in that spot.

But what are we even talking about? Who can't afford a ton of rangers? spend $2k and just know you're going to lose that $2k every fight. You really shouldn't be picking fights were you're not expecting at least a $3k return. I don't even do it unless I get more than $5k

[Agent Orange]

Thanks, Dover. I can't imagine the developers of the game taking into account relative values of atk/def of each entity and comparing it to each and every entity of the attacked units. I do that for a living, so I know how hard and tedious it is.

They don't for choosing rivals so what are the odds it's done for attacking and defending.

[azy]

Couldn't agree more. It's not like the "picking" algorithm was carefully thought out by a PhD in math. Rather, some guy coded a function that returns a unit or four out from your army while applying some weight to the selection and a healthy dose of calls to random(). The reason the logic is not explained is that nobody truly knows. And I bet the statistical sampling done by the folks in this thread is more thorough than the dev's.