SM1 - Sid challenge

[zerksees]

You guys can try anything you want - I'll play along for fun. But it seems to me that maybe we should figure out why we crashed on the Sumerian game and replay it to success. I would think we would like to learn the correct way out of this problem, and don't tell me there isn't one.

I am finally well on my way to my first Sid win and it has been quite a learning experience. I was rather particular about starting location and map setup to help me. I have already learned a few things about how important every move is on every turn from that game, and from this game too. I have a HOF agenda that probably won't allow me to play this game the right way by myself but that is an alternative I suggest.

I am no Sid expert but I certainly know a little about how to get better at Civ. Looking for a better map or a better UU certainly isn't going to prove anything. Having a plan upfront, as Smart suggested, certainly will help, and maybe that is enough.

Or maybe we should all go play the game starting before Beorn-el-Feared played it, and see who comes up with the best result. (No offense B, we just know that the 10 high immortal stack formed towards the end of those turns, and I think that is where the wheels came off. I am sure I would have fallen into the same trap. With the advantage of 20-20 hindsight, we could try pre-emptive measures to keep those stacks from forming, so next time we will be better prepared. Or maybe if we leave the EW army at home we can beat down the swarms, and that would be more benefit than pillaging. If we don't try we won't know. )

Smart - since it is your game I will go along with whatever you want to do.

 

[Own]

I agree with zerksees that a new civ won't solve our problems, but a better map will. Handy900, the warmaster himself, has, with his team, the only AWS tiny continents win ever, said that continents requires all of the following: Great play, good map with big lakes, cow and river, iron, luxuries, and RNG luck. We do need to try lots and lots of maps to find "the good one."

Sumeria IMO is the best. Heres my take on strategy-

Builds:
1. Enkidu warrior
2. EW
3. Settler
4. EW
5. Rax
6. archers, ew's, whatever need be.

Research should be:
1. WC
2. Masonry
3. Iron working
4. Alpha
5. Math
6. Writing (maybe)
7. Lit (maybe)

Cashrushing raxes and walls is a tough issue. If we don't, we might risk shields by having to if theres a threat, but we also might waste pop and happiness if there never would be a threat. Pop is more important, but so is time, so I'd say pop rush. If we have a cow we'd grow fast.

Why does not being able to build granaries make agri useless? It still is very powerful.

 

[Smart]

@zerksees
We could prevent those stacks from forming, if we had used trick with attracting EW at Sardis. I'm agreed with your suggestion, and if nobody are against, we may continue this game from my last save at page 4. We could lose this game later, because of Immortals (very early "Medival infantry" ), but let's give it a chance. Anyway we shall gain an experience, which we can use in the future games...
About the Army - I think it should stay at home while we are building EW's. When we will have enough EW's to defend our core, Army will start pillage.

@Own
I'm agreed with your research order, but build order have to be changed. Early settler is bad idea, because we need at least 5 EW's to survive vs first stack of free AI units without walls. What you think about early rax, like in this game?

 

[Own]

I think an early rax is a bad idea, because the first wave will make everyone elite anyway. If our neighbor isn't insanely close, we'll have time for twe EW's, settler, then some more EW's.

 

[Own]

Let's get this moving. How about trying nero's start and if it fails try mine. I think a new thread is a good idea.

 

[Smart]

Zerksees may play 20 turns from my save at page 4. Then we will know, what to do

 

[Beorn-eL-Feared]

For some very odd reason I missed nero's start until now. It looks really good and I'd advocate trying it.

 

[zerksees]

Got it. Will play tomorrow night (Eastern Standard Time)

 

[zerksees]

I played this game last night – for about 3 hours.

Unfortunately I crashed after 16 turns, losing SoI. Through 15 turns I was able to keep an advantage in unit counts, taking 24 and losing only 8, but then the wheels came off. In turn 16 I lost two archers attacking immortals, costing them only one hit point. On the IBT, I lost two EW and 3 archer in SoI, as my bait move on Sardis did not work and they attacked SoI, costing them 0 units. Multiple Persian immortal stacks were forming and I moved the army to address one near FoS, and the other near SoI was to be baited. Oh well. Baiting had worked fairly well before then, as invading and leaving cities open left the enemy walking around chasing wild geese.

At 1200 BC I noticed Persia had 15 cities already. We have four. So in addition to overcoming the 60% production discount and the immortal, we also have to overcome the city count deficit of 4 to 1. In order to succeed at this we are required to succeed in with a ratio of at least 10-1. For them immortals cost 12 and for us EW cost 10

Obviously I was able to make about 3-1 until my tactical abilities failed. To do this properly I need to know how the AI decides what to attack, rather than going on vague generalities – such as leaving cities open or invading their territory. I need to know how to coax them to attack our EW army. Clearly they will attack it, but I could not get them to attack it when I wanted. I also failed to attack with it – I should have been able to whack about one unit a turn on attacks. Furthermore my baiting attempts failed at the end – somehow the program determined my city was a more viable target than the invader, which it had not decided before.

I can tell you that no yellow or redlined immortals escaped, and I put full strength archers to work on immortals any time they were standing on grass or desert. Clearly something else is missing.

I know a veteran archer vs. regular immortal (or spear) defending is only about 61% successful, assuming the defender is not fortified and on grassland. This ratio is not going to cut it for us.

On the IBT after 1375 BC, six Persians fell to the EW army and backup EWs with no losses. If you take this out of the equation, my kill ratio was only 3-1.

Of course all this analysis is in hindsight.

So if anyone has insight into getting the AI to attack the army I would appreciate it.

I also learned I need help with MM. Three times I lost a city for a turn because I did not MM properly. Too much effort on positioning and not enough on city management.

I might try this again but I don’t want to hold up the game. It is clear that I still have something to learn here about tactics, and might replay this a few times to see what I can learn.

 

[schadenfreude]

unless it is in a city. Even then, they will bypass that city to head to another softer target. You can either A), take your army out for pillaging (if it gets yellowlined, it will possibly be attacked), or B), bait the AI to attack with slow one move units into a empty front line city, where you have multiple cats and archers and spearmen waiting on a road behind it, ready to move in and blast their stack. When (if ever) the stack is gone, pull forces back behind again and repeat. This of course only works if they don't have horses.

I would try this game but with picking your opponents. Always War Sid is IMO tough enough without the Persians. Maybe not as glorious, but a greater shot at pulling it off.

 

[Own]

You are absoulutely right about the armies, they won't get attacked in a city unless there's no other target.

 

[Whomp]

Are we going to attempt Nero's start? I think we should at least try.
If not, I agree with schadenfreude. Sid is tough enough without an offensive UU. I'd prefer to try Romans, Persians or Iroquois.

 

[Own]

Romans may not be a bad idea. Persians sound good too. Lets try nero's first.

 

[zerksees]

You are absoulutely right about the armies, they won't get attacked in a city unless there's no other target.

If I understand what you mean, this is not 100% correct. They did in fact attack the city with the army on the IBT before 1350BC. 5 immortals and an archer died trying to attack it. There were undefended cities further away that definitely were targets, and they did not move toward them.

What I need to know is how many units the other cities need to have in them so that the AI will decide "since it will take me 10 turns to get to city B, I will attack city A which is closer". I am hoping it is not just because the army was new (and the AI did not know better yet) that it was attacked.

Maybe they are not aware of the army for one turn. Maybe I can get them to approach the city and put the army in there just before the attack and they will go through with it.

Decoding the decision tree that the AI uses when attacking is an interesting task.

 

[Smart]

AI will not attack an army 100%, if it's outside the town and fully healed. If army is in city, there is chance, that AI will attack it (remember your last turnset in Own7 )

It's interesting to see the AI algorithm, but to do this you need source codes . We can also make maps in editor, and see the AI behaviour in different situations...

 

[Own]

That army was wounded, but it wouldn't have been attacked had it been close to full strength. There was another empty city two tiles away.

Source codes aren't really needed, just testing. Some of my knowledge comes from my experiances, some from others.

 

[Smart]

http://forums.civfanatics.com/showpost.php?p=3509381&postcount=83

Try the last save - AI will attack 100% healed army in the walled city on hill

 

[zerksees]

After playing my 20 turns in an improved manner, all I can say is "That EW army is da bomb!"

It just took three tries to figure out how to use it. It was really quite simple. All I did was load it up with 3 EW and put it in SDC with a couple EW and a couple archer backing it up. I put at least one or two EW on the other cities every turn, and stayed away from Sardis, so they would not be distracted. I also gave the workers a EW escort for good measure, up to the end when we acquired the slaves. Every turn immortals came and attacked the army, and every turn they all died! This avoids the large buildup of units and we can handle four or five at a time.

Why? I am not sure, but it may be the mountain and hill approach that they like - as you know they like to travel places where they feel safe.

Final kill ratio: 47-5. I lost an archer taking down Zohak, lost an EW at Port Arthur, lost three other archers on pillaging spearmen. We need mathmatics to get catapults. It will be here soon.

I built two cities (you guys can rename Umma if you like).

It started as an experiment on whether they would attack the armies and it just kept going.

Two more cities and we can get another army.

Other highlights:
- Our army made it to red three times, once down to only one hit point, but it lived. Only one of the EW in the army got promoted.
- I used the army once near the end to take out a spear when a 2-stack of spears came by SDC.
- Persia did drop an immortal near Port Arthur, but I was able to get archer there so only one EW was lost.
- Persia dropped a settler off with an archer escort right next to our elite archer. A gift from Xerxes - how thoughtful.
- No pop rushing
- Lost one turn in FoS to disorder

Sorry if you don't like the city locations. This was a quasi-experiment that went real well. Umma was put on a hill like the others, and Really Bad-tibira was put on the ruins of Zohak to try to monopolize the space up there. It seems like a bad choice but I did not want to waste another turn moving it again. Someone with better MM skills than me should be able to make this little empire rock!

I hope you will want to pick this up again. It may take a couple tries but we can definitely learn how to beat Sid this way.

Other:
- That is a settler pair S of Sardis. Move the EW off the hill 2 SW of Umma and they might start to go north again and get us more free workers.
- Having the two archers up north will be handy in case any more settlers land until we own all of it.
- I suggest we should get up to eight cities ASAP, and build a second EW army to hide our offensive under. Then take Sardis, then start moving south...

 

[Beorn-eL-Feared]

That should put Whomp up, and I think it's a clever way of thinking outside the box - that is, to neither hurry our arses to pillage nor to build an offensive army. With catapults, we should be able to slowly gain terrain - the Persians won't get bigger and badder will they?

 

[Whomp]

Very, very nice work Zerksees. I will pick this up and play it tomorrow.