Opinions about current simulators

[VVidi]

Dear Settlers

Since the demise of what was probably the favourite battle simulator amongst many TSO addicts, and the marvellous creation and improvement of the others, I would like to ask for some experienced opinions on how best to calculate a reasonable result - not necessarily the absolute optimum but as close as possible :).

May I please take as an example of the penultimate boss camp in the bottom right-hand corner of the Secluded Experiments adventure where there are 3 crazy cooks (not Swedish chefs); 50 dastardly gunmen; and 125 snakes with sabres. There are a couple of excellent guides for this adventures, and in all of them that I have read this camp is tough enough to warrant 4 attack waves. If this was the only and/or the last camp that's no problem, but there are lots more!

My thinking is that if I am going to have generals recovering, preventing me from continuing that adventure, I may as well send every single fast general I have into a battle, plus any additional slow generals that I will not need for the next 4 hours.

So, the question is how to use the battle simulators to work out what I need? If this camp were composed of different troops maybe I should send 1 slow general with 187 bows then a fast general with 123 longbows and 50 recruits then a vet with 240 recruits then my MMA with 5 elite and 165 cannon so that after they have all recovered I can send my MMA back with a feather to finish it off.

Hoping this isn't a silly question :)
VV

[Gerontius]

It's down to you to work out how to approach the attack, then use the simulator to see what troops are needed for each stage

With this particular camp, you have the cook defending the rest of the troops, so you either need to use cavalry to kill everything else first, or you need to kill the cooks first then mop up the rest.

A quick simulation will show that the cavalry option isn't really feasible, so you have to go for the second option.

A few simulations with bows and longbows will reveal the details of the first 2 attacks you specify above. Although, personally, I use a different B/LB mix

The rest is then pretty straightforward.

[VVidi]

Thank you for your answer, Gerontius.

I have changed my original post because I inadvertently suggested that a certain combination of attacking troops may be feasible, when they were just random example attackers.

From what you say there is an assumption that people should know things like if there is a crazy cook kill everything else first. Firstly in this camp there are three crazies so there are two others to kill first also, secondly it is exactly that kind of knowledge I lack, and so probably do 90% or more of players - what are the "obvious" facts and where do we learn them?

[Gerontius]

You learn the facts by (a) studying the guides and (b) understanding how combat works.

From that anyone can work out WHY the guides do what they do and form their own strategies.

[Gavona]

You learn the facts by (a) studying the guides and (b) understanding how combat works.

From that anyone can work out WHY the guides do what they do and form their own strategies.

Gerontius is being a bit over-modest in not linking to his excellent guide - http://forum.thesettlersonline.com/threads/25571-Guide-Detailed-combat-guide

Thanks for that!

[Errol_Chin]

Gerontius is definitely being modest - personally I think his guide should be stickied with a note to read it before using the adventure guides or simulators. If you think of the adventure guides as teaching you a recipie, Gerontius guide teaches you how to cook :)

[pauantbro]

+1
well said :-)

[Errol_Chin]

From what you say there is an assumption that people should know things like if there is a crazy cook kill everything else first. Firstly in this camp there are three crazies so there are two others to kill first also, secondly it is exactly that kind of knowledge I lack, and so probably do 90% or more of players - what are the "obvious" facts and where do we learn them?

Just to expand on that - the obvious facts are the unit data for the troops in a camp and the order in which they are placed and you can learn all about them in Gerontius' guide.

For example, having read his guide I know from the unit data for the crazy cook (5000 hitpoints, 200-300 damage 66% accuracy, first strike, splash damage) and the number one position in the camp that:
a) the only player unit that wont attack the cooks before anything else are cav, who will attack the gunmen, then the petty officers because of the "attacks weakest" attribute.
b) because the cooks have first strike, I will take a hit from all three no matter what. That the strength (damage) of that hit will be one of the following 3x200, 2x200+1x300, 1x200+2x300 or 3x300 hit points.
c) because the cooks have "splash damage", a single cook can attack many other units at once.
d) because the cooks don't have the "attacks weakest", their hit will be applied to my melee units first in the order recruit, militia, soldier, elite soldier before going on to the ranged units.
e) With 5000 hit points each cook can take a lot of damage before dying and will recover any lost hit points between waves, so all 5000 damage has to come in a single wave

now you can do some very simple number crunching before getting into the simulations.

We can quickly see that the cav option is limited. There is no combination of gen and cav that can possibly kill a cook. While 50 cav will kill the 50 gunmen, it will take between 6 and 12 to kill a single petty officer. (that's up to 1550 cav you wont be seeing again and you still have the three cooks to worry about).

From there you can start number crunching the troops that can absorb the cooks initial hit and protect the other units so they get a hit in - max damage (900) / unit hit points {recruit 40, militia 60 etc} so 23 recruits, 15 militia. Then add on the troops that can take out the cook - hit points (5000) / min unit damage {bows 20, longbows 30} so 250 bows or 167 longbows. So from that I would use 23 recruits and 167 longbows as a starting point in the simulator to see what the minimum cost is to guarantee killing a cook - you can probably see now why it is 4 waves - 1 for each cook and a final mop up.

Before I read the battle system guide last week, I could not have explained any of that.

[Gerontius]

Thanks for the praise for my guide.

I didn't highlight it because it isn't actually finished. I'd meant to expand it to answer this very question directly but never found the time or the words.

But I'm glad what is there has proven useful :)

[Bluesavanah]

I won't comment on that particular camp but what I have started doing is looking at unit cost per camp, Fairytale camps in particular before the changes I just use whatever was needed to take the camp out in as few hits as possible. Say for example I'd worked out I could kill a camp with 100 soldiers and 170 cannons losing 100 soldiers I'd of been more than happy to do that, but now I'm not so happy with it. The reason being that 1000 steel swords sells for around 1000 gc which I could buy 20,000 bronze swords for example. Being as I have plenty of generals I now consider can I kill that camp for less money.

There are many variables to consider when I sim an adventure, can I reach the end of the adventure in a timely manner,how much am I spending to do the adventure, how long does rebuilding take ?

[Sharpielein]

What I always do when using my guides:
Given any camp that can't be outright cracked with 1 single wave ...

I first simulate a wave of 250 ::R:: or ::B::, then see what's left.
I then typically go for Wave 2 with an MMA or Major, depending on whether there are low HP units or not. That wave I load with a mix of ::R:: and ::K:: and try to adjust upward or downward.
If the amount of lost ::R:: looks fairly low, I will see if I can already save troops on W1, if there's no way in hell freezing over to not lose ::K::, I introduce a wave either before or after the first. Usually, I introduce a MMA Suicide wave before the first and see if that helps.

In calculating the amounts of troops, I use a method called "binary search", so for example if I would simulate with 100 ::R:: 150 ::K:: and see I lose cannons, I go to 170 ::R:: 90 ::K::, and see if I still do. If I don't, I try 135 ::R:: 135 ::K::, and up and down until I got the perfect number.

It usually takes about 2-3min to simulate a camp. I don't claim my solutions are always optimal, but I figured that given my losses for no-block VLT are just about 5000 troops , I am probably doing something right.