Do you attack up or down the stack?

[JammerUno]

Strongest to weakest gives the best results. To minimize losses you have to minimize average losing odds. Since the strongest defender will be chosen, pairing anything else than the strongest attacker with that defender doesn't minimize losses.

Elite Units go last. At least for aggressive warring you need them to become effective stack defenders.

Defending units are rarely offensive units. Formation/shock/cover don't help attacking units.

 

[karadoc]

I think YMMV may apply here.

If my elite unit has 97% success odds, and my rookie has just 60%, then I would risk the elite.

If it's a suicide job, the rookie goes first.

If the odds are good all around, then I prioritise units close to reaching the next level of experience. This has the advantage that you could potentially promote them to save healing time on the next turn.

I don't know what "YMMV" is, but this basically describes what I do. Also, I try to pay a bit of attention to what the defenders. For example, if the top defender is a crossbow then I'll probably attack with a horse first rather than a swordsman, even if the swordsman gets better odds. That way I negate their bonus vs melee.

 

[Solon70]

I don't know what "YMMV" is, but this basically describes what I do. Also, I try to pay a bit of attention to what the defenders. For example, if the top defender is a crossbow then I'll probably attack with a horse first rather than a swordsman, even if the swordsman gets better odds. That way I negate their bonus vs melee.

I don't really understand the purpose of taking worse odds just to "negate the bonus." The bonus is already part of the odds. Maybe this could have been phrased better.

 

[karadoc]

Suppose I have a horse archer and a swordsman for attack, and his city has a longbow and a crossbow. Suppose that the promotions are such that his crossbow is the stronger defender against either of my guys and my swordsman is the stronger attacker (for example, I have city raider on the swordsman and he has city defence on his crossbow - no other promotions). Ok.. so as I said, the swordsman has the better odds, but if I attack with the horse archer first then that will get the crossbow out of the way so that when my swordsman does attack it will be against the longbow. That way the crossbow misses out on his bonus vs melee - because he fought the horse archer instead of the swordsman. The end result is that although my odds on the first fight were worse (compared to using the swordsman first), the odds for the two fights together are better because of how the bonuses play out.

 

[DaveMcW]

his city has a longbow and a crossbow. Suppose that the promotions are such that his crossbow is the stronger defender against either of my guys

This never happens in a real game.

 

[trc]

Suppose I have a horse archer and a swordsman for attack, and his city has a longbow and a crossbow.

If this is real situation, then you played the war wrong.

Successful waring based on playing advantages, either in sheer force, or by tech advantage. Fortified LB is where I generally stop until new siege, unless fiercely pumping units, which normally isn't the case at this date.

Recent save on this? I'd wager that a look at the power graph suggests a defensive posture..

 

[slobberinbear]

Let's use a realistic example. You are Phalanx rushing with Alex against Gandhi, whose city is defended with two archers and an axe.

Let's say you built the Phalanxes from a barracks and have a promotion to give each unit.

Without a promotion on your Phalanx, let's say the archer is the chosen "best defender." So you give one Phalanx the Cover promotion. The "best defender" then changes to the axeman. You make a note of the odds. You then give the City Raider I promo to a different Phalanx. You draw the archer but at better odds than the unpromoted unit. You then attack with the Combat I / CR I Phalanx. Chances are you will have to do this again with a second Phalanx to hurt/kill the second archer. Eventually you will eliminate one class of defender or weaken them so severely that one class of unit is forced to defend ... then you promote your Phalanx to match up against the known defender.

There are those that simply would give the CR I promo to every Phalanx. But cover / shock / formation give 25% bonuses, 5% more than CR I. That 5% is meaningful. I will be the first to admit that CR has its uses (especially when you can stack on CR II and III), but you if you are searching for the best odds with fresh troops you can use the promotions to get a slight edge.

Once I have a stack of veteran units, I will use newbies as cannon fodder unless the veterans have good (say 70% or better) odds.

Probably the thing I need to work on more is preserving my catapults. I tend to suicide a lot of cats for collateral damage, and end up with overkill on the melee units. This is fine for the first conquest but tends to slow down the overall war effort while the surviving cats heal and replacements arrive. I should probably use my melee units more aggressively after the walls are reduced and the defenders have taken 20% collateral or so.

 

[karadoc]

This never happens in a real game.

If this is real situation, then you played the war wrong.

Successful waring based on playing advantages, either in sheer force, or by tech advantage. Fortified LB is where I generally stop until new siege, unless fiercely pumping units, which normally isn't the case at this date.

Recent save on this? I'd wager that a look at the power graph suggests a defensive posture..

Rubbish. It happens all the damn time. The example I gave was very simple so that you'd see the point. Obviously there will usually be more than 2 attackers and two defenders. But the crux of the matter is that in pretty much any combat situation you will be up against mixed troops, and frequently the order of attackers will effect whether or not you are able to break the enemy stack.

Here's another example that you might like a little bit better. Suppose I have 14 war elephants and 4 macemen, and I want to kill an enemy stack which consists of 2 pikemen and 12 knights. If you go with the best odds attacker every time, you'll have maceman vs knight 4 times, and then you'll still have to face the pikemen with your war elephants when your macemen run out. It is much better to attack with the war elephants first, even though the odds are much worse, because once you get those pikemen out of the way then the remaining elephants will have much better odds vs the knights than the macemen would have had. You might think "oh, but in a _real_ game, you'd have siege units for collateral damage, and blah blah blah" -- it makes no damn difference. So maybe the enemy stack is all softened up by collateral damage, the same reasoning can still apply just with better odds for everyone.

 

[VoiceOfUnreason]

Here's another example that you might like a little bit better. Suppose I have 14 war elephants and 4 macemen, and I want to kill an enemy stack which consists of 2 pikemen and 12 knights. If you go with the best odds attacker every time, you'll have maceman vs knight 4 times, and then you'll still have to face the pikemen with your war elephants when your macemen run out. It is much better to attack with the war elephants first, even though the odds are much worse, because once you get those pikemen out of the way then the remaining elephants will have much better odds vs the knights than the macemen would have had. You might think "oh, but in a _real_ game, you'd have siege units for collateral damage, and blah blah blah" -- it makes no damn difference. So maybe the enemy stack is all softened up by collateral damage, the same reasoning can still apply just with better odds for everyone.

I completely agree that the principle is just as valid with flatlined defenders as it is with healthy ones (although I wouldn't be surprised to learn that in former case, the difference isn't worth the effort of doing the math).

To some degree, this mostly seems to be a reflection of the idea that the way to beat rock paper scissors is to play rock-rock-rock-rock... until such time as you can finally rip through the paper to beat upon the scissors.

In this particular example, the key is that since Elephant vs Pike encounter(s) are inevitable. By scheduling that suboptimal encounter early, everything else gets better (in particular, the final mop-up of the pikes is done with maces instead of elephants).

It also suggests to me that there may be a tricky balance in getting the promotions right on your suicide elephants - you want to go anti melee here, but not so much that you inadvertently draw a knight (unless, of course, you have sufficient promotions to clobber all of the knights first). Not sure that's a real concern in practice (at least not in this example, where the pikes will likely have a defensive bonus working for them).

There are some important ideas here that deserve to be pulled into the light and examined.

 

[Calouste]

Down the stack. That gives better win/loss ratios and win/loss ratios play an important part in the AI's perception of how the war is going and if they are willing to file for peace and how much they are willing to pay for that and if they are willing to capitulate. Well, unless you plan on annihilating your opponent completely of course.

 

[Bostock]

Attack with your 2nd best elite unit and down.

Your best elite unit needs to survive to unlock the Heroic Epic and West Point.

I thought West Point was, as the old programmers' joke goes, "considered harmful"?

 

[Bostock]

Oh, and down the stack personally, with quirky exceptions here and there. Overall I'd rather have lots of somewhat-promoted but not-dead units than a few elites fed by the corpses of green units.

 

[Tecibbar]

Unless I need to unlook Heoric, or if my unit maintainance is killing me, these are my rules.
1. Collatoral damage cats/cannons
2. Use flanking 2 mounted if they have the highest chance of survival.
3. When my city raider siege has over 20% chance of survival, use them.
4. When most defenders are at ~70% health, or my best attack have 95% chance, stop using siege.
5. Best units first, unless the next best have similar survival chance (<10% difference)
6. Use CRIII swords before maceman, if they have similiar chance (<10% difference)

When you use weekest units, they will just die from first strike without doing much damage.

 

[Udey1]

Strongest to weakest gives the best results. To minimize losses you have to minimize average losing odds. Since the strongest defender will be chosen, pairing anything else than the strongest attacker with that defender doesn't minimize losses.



Defending units are rarely offensive units. Formation/shock/cover don't help attacking units.

Didn't see this reply. Allow me to rebut: Formation/shock/cover aren't even in the equation for many later wars, so elite units promoted down the combat line are excellent defenders. That is more in line to what I was referring to. I do not count an axeman with shock as an elite unit, just a junky defender with no future.

 

[CivMcNut]

My vote would be for up the stack (most of the time). This is with the exception that if the odds are really bad for the rookies, I start out with the veterans. Most of the time if siege has properly softened up the defenders the odds will be favorable for low-unpromoted units. I hate loosing my top troops unless I know it was for a pivotal battle. I like having the chance to have rookies move up in experience fast.