The caluclus of exactly how Withdrawal (i.e. Flanking) works?

[ModernKnight]

The Civilopedia says Flanking 1 gives +10% chance, and Flanking 2 gives +20% chance, of Withdrawing during combat. And of course certain units have built-in withdrawal percents, like the Submarine.

It seems to me that the usefulness of Withdrawal depends heavily on how combat works "internally". How the dice are rolled, if you will. Longtime D&D or RPGers can relate:

• If only a few dice are cast, it's entirely possible to be killed in one particular dice roll, before your withdrawal capability can kick in.
• But at the other extreme, if there are very small dice cast, you will get e.g. exactly your rated 30% chance to withdraw - whenever the game decides is a decision point to withdraw.

So there are two questions here (maybe): How "big" are the rolls, and how/when does the game decide to withdraw. (It could in theory roll to withdraw even before the first combat - but that wouldn't make sense, and doesn't happen. When exactly does it happen?)

It seems almost like a calculus question. When does it kick in, versus your odds of attack, in the middle of an attack? How many "mini rolls" are involved in an attack?

I know there's a Combat Log, but I've never studied it in depth relative to this. Perhaps somebody else has already?

Thanks if you can help! - MK

 

[Kiershar]

I believe the "Combat Explained" chapter in the Academy explains all that.

 

[JackOfClubs]

Retreat rolls happen when a round ends that would kill the attacker. In that case, a special role is made against the attacker’s retreat chance. If successful, the attacker retreats with the number of hps he had before that final, potentially killing round. The attacker’s chance of winning combat is never decreased by having retreat chances. They only come into play when the attacker would die.

From Arathorn's Combat Explained thread in the Strategy Articles section. You might want to take a look at that article as it has more on the subject than I have quoted.

@Kiershar: Jinx!

 

[ModernKnight]

Thanks Kier and Jack, that clears it all up - and makes it very simple. It's a "back end" thing... if you were set to be killed, you get your Withdrawal chance of retreating. Simple and cool.

Thanks!! - MK

 

[ModernKnight]

P.S. What's this about an "academy". Jack, your link brings up exactly one (great) post. Kier, you mention an Academy, but I don't see it on the listing of forums here.

 

[Maydrock]

On the CFC home page on the left side you will see a link to the "War Academy". Inside the academy are links to really good articles various players wrote on different topics. I believe most(actual posts) can be found in the subforum("strategy articles") to this forum.

 

[JujuLautre]

Thanks Kier and Jack, that clears it all up - and makes it very simple. It's a "back end" thing... if you were set to be killed, you get your Withdrawal chance of retreating. Simple and cool.

Thanks!! - MK

Just consider Withdrawal as being a save roll against death

 

[ModernKnight]

Thanks Mayd and Juju - sounds like good advice, both ways.

Er - Mayd - can I ask your royal highness if you're a lady?

 

[Chemtech]

My intial response was:

"It doesnt" or at least it doesnt ever seem to work with my horse units, although it seems to work great with catapults.

I seem to lose an inordinate amount of "flanking" promoted units - I usally send them in ahead of a city attack (after cats of course) trying to do a little additional damage to the main defender before sending in the ground troops - but I can't even count the number of times Ive sent in 3 or 4 horse units and lost ALL of them - now the odds are I should lose 3 out of 4 or 4 out of 5 - so losing them all sometimes makes sense, but it just seems to happen ALL the time to me.

On the other hand, I usually manage to keep 2 out of 4 suicide cats so I guess it kind of balances out.

 

[Perugia]

My intial response was:

"It doesnt" or at least it doesnt ever seem to work with my horse units, although it seems to work great with catapults.

I seem to lose an inordinate amount of "flanking" promoted units - I usally send them in ahead of a city attack (after cats of course) trying to do a little additional damage to the main defender before sending in the ground troops - but I can't even count the number of times Ive sent in 3 or 4 horse units and lost ALL of them - now the odds are I should lose 3 out of 4 or 4 out of 5 - so losing them all sometimes makes sense, but it just seems to happen ALL the time to me.

On the other hand, I usually manage to keep 2 out of 4 suicide cats so I guess it kind of balances out.

I agree cats retreat often when attacking cites, especially with additional flanking promotions.

This is somewhat of an anecdotal response but I too think mounted units do not retreat when attacking cities but they do retreat in the field. Can anyone disprove this?

 

[ParadigmShifter]

Chariots can definitely retreat when attacking cities. I even used flanking War Chariots (egypt UU) to soften up spearmen in my last winning game. Chariots count as mounted units.

The chance of retreat seems to be adjusted based on the chance to win though, so weak units like chariots and catapults seem to withdraw more often than stronger horse archers.

Don't forget Horse Archers in Warlords get -10% city attack too.

 

[ABigCivFan]

I tried giving flanking to my mounted units, they still died too often. So I never give them flanking anymore, it is much better to give them combats+specific counter promos (shock, pinch, formation), which gives them much better winning percentage rather than chicken away percentage.

For the same reason i always give my city-attacking siege units CR promos.

 

[ParadigmShifter]

I give mounted units combat promotions if it is early in their lifespan and run away promos if their time is drawing to an end.

 

[JackOfClubs]

Getting a chariot up to Flanking II makes them immune to first strike which can be helpful against early archer defenses.

Flanking II leads to Mobility which can be useful for active defense if you have a large territory to cover, though railroads kind of negate this. But Mobility is very useful for troop carriers if you can get them up to 10 XP. (Theocracy, Vassalage, Drydock and either Pentagon or Westpoint will do it.)

 

[ModernKnight]

Thanks folks -

The chance of retreat seems to be adjusted based on the chance to win though, so weak units like chariots and catapults seem to withdraw more often than stronger horse archers.

Paradigm, if i understand you right, this is contrary to the guide i was referred to earlier. You withdraw if you would've died - period. Is what the guide says.

Does anyone know how to pull out the code (html?) for flanking/withdrawal? It makes sense, in a way, if cavalry don't do well in cities. But you would've hoped the Civilopedia or manual would mention this important point.

No one's mentioned attack choppers versus withdrawal. They should certainly be able to withdraw, even for cities - but not counting anti-air interception.

Jack, I never tried Mobility for naval units - I thought it only meant that hills and forest/jungle don't effectively stop you. And thus would be useless for ships (although I love Navigation). Please tell me what Mobility does for ships?

Thanks all! - MK

 

[ParadigmShifter]

No hard evidence, it just seems if I have a 90% to win I have a low withdrawal chance. Maybe that is factored into the odds since it is so unlikely you would lose hence withdrawal is equally unlikely. Whence... (I've always wanted to say that but have never had a 3rd point to back it up).

 

[ModernKnight]

Paradigm, looking again at Arathorn's Combat Explained thread that Jack kindly pasted, I see more complexity past what I said. Take a look while I do too - I don't have the time to work the math right now, but I see it's agreeing that weaker attackers (cats) are better for withdrawing. See the Retreat section.

Hmm. An infinitely weak unit with every possible barrage upgrade? ... (checks the Civilopedia)

Jack, what have I missed about Mobility?

Huh - Arathorn looks a little young. God. When are old people supposed to turn senile, 80 or 40??

 

[ParadigmShifter]

Mobility reduces movement by 1 for any tile with 2 movement cost or more. Check out Shaka's Impis. They can move 2 tiles over hills and forests except when they start by moving into a forested hill (which would be 3 movement). Ships need Navigation to get an extra movement since all sea tiles are 1 movement only.

 

[Spitefire]

If you have 30% chance of retreat then it works like this (with a 0% chance to win you will retreat 30% of the time) (with a 50% chance to win you will retreat 15% of the time) (with a 75% chance to win you will retreat 7.5% of the time) it effects your lose odds so that 30% means that 30% of the time you would have died you instead run, are we understanding yet?

 

[cabert]

If you have 30% chance of retreat then it works like this (with a 0% chance to win you will retreat 30% of the time) (with a 50% chance to win you will retreat 15% of the time) (with a 75% chance to win you will retreat 7.5% of the time) it effects your lose odds so that 30% means that 30% of the time you would have died you instead run, are we understanding yet?

exactly.
The way retreating odds are displayed is misleading although right.
What you see is :
odds to win : X%
odds to retreat : Y%.

What it means is that you have
100%-X%-Y% odds to die,
X% to win,
Y% to retreat.
If you didn't win, you don't have X% to win anymore. You have 0%, and there you're back to your unit's ability to retreat.

Amongst the retreating units, my all time favourite is the submarine .