Solving the Lack of Barbarian Problems

[Jamey]

I've been trying out various ideas to help keep Sarmatia from going deeply into the red. Right now, I've made these changes:

1. Changed the happiness bonus on their government buildings to a law bonus, thus reducing corruption.
2. Changed Tanais and Campus Alanni to be large towns at the beginning of the game. This effectively built the building, since they had the population already.
3. Changed Campus Sakae to start with a Temple of Law. That gives an extra 10% happiness bonus, which seems to make Sarmatia hold onto Campus Sakae rather than seeing it rebel.

With these changes implemented, Very High Taxes, and the capital in Tanais, I'm seeing a net profit of 1400 per turn. Despite that, the AI goes negative in around 10 turns of play. I suspect that this means that the AI is not cranking taxes and is building units.

My next change is going to be to move some troops to be in position to attack Campus Sakae, which should provide more income sooner. Right now, the AI pulls troops up there, but it takes around 5 years to complete the conquest (normally the first for Sarmatia). By that point, the AI is already in debt and it apparently just keeps snowballing.

[Solaris]

Lowering upkeep if they insist on recruiting could work.

[Quinn Inuit]

Definitely. Also, should we just give them Campus Sakae?

[Jamey]

Using Tanais as the capital yields the highest income for Sarmatia. That remains constant even with the conquest of any combination of the Crimean settlements, Kotais, and Albania. It changes when you start counquering down into the Balkans.

I tried giving the Sarmatians Campus Sakae. What I hadn't completely thought through is the effect of distance on the fringes of the Sarmatian economy. Roughly 3/8 of all income is lost to corruption without law bonuses. Even with a solid governor (+25% in law and influence bonuses) and a temple, it requires 5 units to hit a Very High tax rate (another 60% bonus to order). Obviously, that can be reduced with either a better governor or more temples, but we're starting to get silly with the Campus Sakae start point. We could also reduce Campus Sakae's population, but that would eliminate the value of controling it.

Campus Massagetae is even worse, clocking in at a 60% loss of income due to corruption. That pretty well eliminates value of economic development there. Without a truly staggering law bonus, I'm not convinced that it's possible to make Campus Massagetae provide actual profits after its garrison once it grows.

Anyway, that's the point I've reached. I'm contemplating options. I suppose we could apply a larger law bonus or a law bonus and an unrest penalty to help with the corruption, but that seems like it's a bandaid rather than a solution. The answer may be reducing the maintenance of Sarmatian units further, but that will make them pretty unstoppable if they take any better land. :hmmm:

[Quinn Inuit]

Ah hah! What if you give them a law bonus only where there's the AOR_Scyth resource, and maybe a lesser one without it? That way they can at least hold onto their early settlements. Conquest will still be difficult, though maybe we can help with a client state solution. Nomads frequently left local cities alone as long as they were nominally in charge--take a look at Bactria after 150 B.C.

[Jamey]

Ah hah! What if you give them a law bonus only where there's the AOR_Scyth resource, and maybe a lesser one without it? That way they can at least hold onto their early settlements. Conquest will still be difficult, though maybe we can help with a client state solution. Nomads frequently left local cities alone as long as they were nominally in charge--take a look at Bactria after 150 B.C.

Are you inside my head? I was thinking about how we could represent the loose authority of nomadic tribes over areas they conquered when I woke up.

Too bad we can't change hard coded values. I'd support giving Sarmatia a flat corruption penalty across all settlements rather than one that grows with distance if we could.

[decimator22]

Yes, that would be the best for the sarmatians. If you see my AI faction bet, they can't expand, well they get like 2 settlements and then they rebel from them, They also are always in red money. I do reccomend giving them campus sakae, as they always try to get it and then it rebels from them. DO give them a give law and happines bonus in the steepes.

[Quinn Inuit]

Are you inside my head? I was thinking about how we could represent the loose authority of nomadic tribes over areas they conquered when I woke up.

Too bad we can't change hard coded values. I'd support giving Sarmatia a flat corruption penalty across all settlements rather than one that grows with distance if we could.

Heh. Well, that's part of my idea for the social structures. You'd have the option of leaving kingdoms in place or supplanting them, and we could easily steer Sarmatia toward the former option.

I agree about the hard-coded values. Nomad factions should get a boost to that from their heavy use of horses, IMO.

Yes, that would be the best for the sarmatians. If you see my AI faction bet, they can't expand, well they get like 2 settlements and then they rebel from them, They also are always in red money. I do reccomend giving them campus sakae, as they always try to get it and then it rebels from them. DO give them a give law and happines bonus in the steepes.

Should I put the law bonus in the next hotfix? If so, what should it be?

[decimator22]

Let me get back to you once I play a lil with sarmatia.

[Jamey]

Another complication for Sarmatia is what we are trying to make it represent.

Is this the core Sarmatian tribal group? They controlled roughly Campus Sarmatae and Campus Alanni at 280 BC.

Are we including the Scytho-Sarmatians (blending of Scythian and Sarmatian culture with a Sarmatian takeover? They controlled roughly Campus Scythii, Tanais (but not the city itself - it was Bosporan), and part of Crimea (with their capital apparently at Neapolis, which is inland near Khersones).

Are we including the Sakae and the Massagetes (they were related and merged later under pressure from the Huns)? They controlled roughly Campus Sakae and Campus Massagetes at the time, and possibly much of the rest of the NE corner of the map.

EDIT: I should note that a written record from the Bosporan Kingdom indicates that in 309 BC the Scytho-Sarmatians fielded an army of 20,000 infantry and 10,000 cavalry in defense of their then Bosporan allies. That's a very infantry heavy army for horse nomads. Combined with the archaeological evidence, it shows that the Scytho-Sarmatians were relatively settled.

[Saint_jimmyz]

This may sound obvious, but try augmenting the value of each province to reflect the distance from the capital.

eg if you have a settlement that should be producing 300 gold per turn, and the distance is reducing that by 200 gold, set the income to 500 gold.

[Jamey]

Here are the changes I've made, which sets the faction up nicely for a human player:

1. Added aor_scyth to Campus Scythii
2. Modified the law bonus on government buildings to 20% in aor_scyth, 10% in aor_dacia and not aor_scyth, 5% elsewhere. This gives a somewhat higher law bonus to the Crimean towns and Thracian lands that Sarmatia should expand into.
3. Added a port and a stables to Tanais. Since Tanais was historically a trade port, this makes sense. The stable is there to provide a location for building generals at the start of the game. Since general production is pretty much the only value of the stable to Sarmatia (they get their cavalry from their base barracks), the 4k price tag is too high for them to pay for a long time.
4. Added Campus Sakae to their starting territory.
5. Added a new commander, Haldita (Partatua's 16 year old son), as governor of Campus Sakae.
6. Gave Zipoetes Warlord 2 and Partatua Warlord 1 at the start of the game. I'm planning on giving Warlord at some level to most faction leaders and faction heirs as I work my way through the barb factions.
7. Made Tanais the starting capital.

These changes bring the base income of the faction to around 800. If you maximize all taxes, the income goes to around 2500 (probably not the best plan given your low starting population). For a human player, that gives you the ability to either build your army and start conquering or sit tight and build your settlements, but not both. I'm about to start testing how the AI responds to these changes.

EDIT: Here are Haldiata's start stats:

character    Haldita, named character, age 16, , x 249, y 152 
traits GoodAdministrator 1, Intelligent 1, HarshJustice 1, Handsome 1, Bloodthirsty 1, Leader_Bonus 1
ancillaries tribal_elder

[Jamey]

I've run several games through 10-20 years and the same thing has happened in each of them. Sarmatia's military deployment looks something like this:


Note how the military is a scattering of small forces. Not once in any of these games has Sarmatia concentrated its forces and made an attack on a Crimean settlement. It has sent an army as large as 6 units to attack Locus Herulae. Also, Tanais tends to have a fairly sizable force (10-12 units) in it which never moves.

Also, in each of these test games Sarmatia has done decently with building ports, roads, and mines. However, it also builds units to scatter around its lands and runs out of money at around turn 20.

Does anyone with more experience have suggestions on how to make the Sarmatian military perform better? Is this a problem with the way the rivers and coasts limit movement to the Crimean settlements? Or is this a problem with the AI dealing with enormous territories? Or is this a problem with the AI dealing with widely scattered starting forces? :hmmm:

[Jamey]

This is very unrelated to the root topic of this thread, but I did come across a list of Scythian and Scytho-Sarmatian kings in Tamara Talbot Rice's book on Scythia. The names we do not have already are:

Galatus
Saitapharnes
Targitaus
Spargapeithes
Lycus
Gnurus
Octomasades
Aertes
Agarus

Unless someone has a strong argument against doing so, I'm going to add these names for Sarmatia.

[decimator22]

I've run several games through 10-20 years and the same thing has happened in each of them. Sarmatia's military deployment looks something like this:


Note how the military is a scattering of small forces. Not once in any of these games has Sarmatia concentrated its forces and made an attack on a Crimean settlement. It has sent an army as large as 6 units to attack Locus Herulae. Also, Tanais tends to have a fairly sizable force (10-12 units) in it which never moves.

Also, in each of these test games Sarmatia has done decently with building ports, roads, and mines. However, it also builds units to scatter around its lands and runs out of money at around turn 20.

Does anyone with more experience have suggestions on how to make the Sarmatian military perform better? Is this a problem with the way the rivers and coasts limit movement to the Crimean settlements? Or is this a problem with the AI dealing with enormous territories? Or is this a problem with the AI dealing with widely scattered starting forces? :hmmm:

In evey game, make it full AI or not Sarmatia never expands to the peninsula, It just scatters there forces and some times gets like 4 settlements and then they rebel from them lossing what they won. This repeats very often.

[Quinn Inuit]

Jamey: Your changes sound good. Could you send me the files? I'll integrate them into the 3.4 build. Or you can hang onto them and integrate them later. Your call.

As for the units, I'll bet that's an artifact of the extreme province size. I have a proposed solution: create a Sahara-like province covering the whole north of the map, leaving those provinces with only their southern third or so. What do you think?

[Jamey]

I tried putting a bridge south of Tanais to facilitate movement. This seems to allow the AI to consolidate its forces and bring them southward. Unfortunately, Kotais and Albania are its primary targets; I'm trying to get the armies to head to the Bosporan settlements. This also changes the road layout of that area of the steppes dramatically. On the plus side, it seems to help with units getting stuck near Tanais.

The most entertaining of these games is here:


Check out Illyria's expansion. Illyria allied with Macedon and just devoured Thrace. It is about to sweep into Sarmatia as well, which is why I terminated this test game. You can see a 19 unit army backed by a 14 unit army near Tanais. Based on their positions and watching other games develop, the AI is currently building up forces for two attacks, one on Phanagoria (larger southern army) and the other on Olbia.

-----

Sarmatia still has the problem of immediately losing the settlements that it gains. I do not have a solution to the problem, but I believe that it is based on the fact that cavalry units are small and autoresolve really hurts them even when they win.

I recently observed an army of 2 Sarmatae Basilikon Hippeis and 12 Sarmatae Hippotoxotai take Olbia. About 6 units survived this attack, leaving the contentment of the settlement at 15%. Had the AI exterminated, it would probably have kept control of the settlement. As it is, it rebelled to the Greeks in a couple of turns, winding up with a garrison nearly as strong as it was before the first conquest.

-----

So, I think that there are two problems that still need to be resolved. First, getting Sarmatia to conquer Crimea before going further afield. Second, making sure that Sarmatia can hold onto the Crimean settlements once it takes them.

Probably the simplest solution to the first problem is to position Sarmatia's attack force near the weakest of the Crimean settlements and perhaps also weakening a garrison for Sarmatian conquest.

As for the second issue, I don't have a good solution. One solution would be to put an even bigger content bonus of some type on settlements with Greek colonial and either Thracian or Sarmatian AOR. That seems pretty arbitrary and it would need to be a pretty huge bonus to be meaningful. The simple truth is that the game is not well designed for cavalry armies (with their small size) taking settlements.

I'm open to suggestions. :hmmm:

[Jamey]

Jamey: Your changes sound good. Could you send me the files? I'll integrate them into the 3.4 build. Or you can hang onto them and integrate them later. Your call.

As for the units, I'll bet that's an artifact of the extreme province size. I have a proposed solution: create a Sahara-like province covering the whole north of the map, leaving those provinces with only their southern third or so. What do you think?

Looks like we cross posted.

I'd like to get things going a bit more smoothly before passing it along. It's a definite improvement, but Sarmatia still tends to stall out because it can't or won't take and hold Crimea.

[Jamey]

Here's my work on this to date:
http://rapidshare.com/files/151351136/3.4_dev.zip.html

This was done with the files from the the 3.3.4b hotfix. None of the changes are in the files from later hotfixes, so it should be compatible. This may not be savegame compatible. Also, you will need to run campaign_map_reset.bat to guarantee that all of the changes appear in the game.

Finally, while I've had no problems during testing of this, this isn't an official update. Use at your own risk. I'd advise that anyone who tries this do so in a parallel version (just copy the ExRM folder in your RTW folder and then make a new shortcut).

-----

Here are the screenshots from my longest test run:

Spoiler Alert, click show to read: 

270 BC


260 BC


250 BC

In this run, Sarmatia was its most successful. Normally, it has conquered 1-2 settlements by 270 BC, in this game it has 3 and is pushing for 2 more. It's apparent preferred order is Kotais, Olbia/Phanagoria (they appear about even), Theodosia, Khersones, Albania in the near east. It also tends to send troops eastward to conquer Locus Herulae and Jaxarta.

I kept this going to 250 BC to see how Sarmatia would do in direct contact with Macedon. Sarmatia lost because it normally waits out the entire duration of the siege until the garrison starves, probably because its armies are mostly or entirely cavalry. It could have had at least Campus Getae and and Sarmisegetusa - it sieged both simultaneously for at least 4 turns with massively superior forces before Macedonian reinforcements arrived.

Given another decade, Sarmatia would have produced a more balanced army, which would have also helped. Around 250 BC, Macedon has the dominant position. At the same time, Sarmatia has finally started producing heavy infantry in Crimea. I know from experience that Armored Hoplites backed by Sarmatian Cavalry are a very fearsome force.

-----

I'm thinking about making changes towards the eastern edge of the map.

I'm going to remove the landblock to see how the balance works out there with a stronger Sarmatia.

I'm also thinking about removing the spawning Sarmatians and possible replacing them with equivalent rebels. Captain Kudurmabug tends to wander back across Asia rather than fighting anyone. I saw one migrate all the way from Bactria with his force intact until he lost a siege against Campus Getae.

[Jamey]

Quinn, I seem to recall you posting a link about embedding garrison units in the settlement declaration in descr_strat. It appears that I forgot to make that a bookmark and I cannot find the thread now. If you still have that link, I'd like to get it.