Soliciting Thoughts on Alliance Barriers To Entry

GM ThunderClap wrote: The ablity to seize a city has the same population requirement to settling - doesn't mean you can "raise it to the ground" though [weird phrase]

I think you mean "raze" - http://www.merriam-webster.com/dictionary/raze
/pedant

GM ThunderClap wrote: Naturally we will put a more detailed guide to seige mechanics in the help which will reveal all

This is going to be very important very soon - so the guide will be much appreciated.

Psst.  The phrase "Raise it to the ground" probably sounds weird because of the replacement of "raze" with "raise."

And it gets worse...after that unfortunate homonym, there's still the redundancy of "to the ground."  Raze means to level, demolish completely, bring to the ground.

Anyway, my concern isn't about the outcome of legitimate conquests, but rather what happens when a player has a bunch of sock-puppet accounts that want to hand over cities, possibly already built up, with as little resistance as possible.

Ok all, I've slept on it a bit and this is now the modified position I'm considering:

HonoredMule wrote: On points 1 and 2:  - I'd like to still see the in-game ability to announce loyalties and enmity...i.e. show non-aggression/defensive/offensive pacts and war declarations, as you do now.  They would just be for informational purposes, except that pacts should still prevent non-secret military attacks, again as they do now.  Making war declarations auto-propagate to a party's offensive pacts (barring a conflict with other pacts, which should just nullify both) would be a nice touch, and might prevent them from being taken lightly.

I agree (with NAPs and Confederations), and there's no reason why we can't keep a state of war as well.  I think you meant "war declarations auto-propagate to a party's defensive pacts"? Again, which I support.  Or do you mean both ways?  ie One member of a Confed declares War so all members do, and similarly if one member has war declared on him then the aggressor effectively declares war on the whole confederation.

HonoredMule wrote: But there's value in maintaining some protection from people who label themselves your friend.... forcing the dissolution of pacts to require an early announcement (like the current 24hour war system) alongside inability to openly attack each other would invest just a little bit of commitment and prevent them from being entered into lightly.

Again, I agree with this.

HonoredMule wrote: On point 3:  - Yup, I agree.  Escrow really looked like a novel and beneficial idea to me at first, but now I don't think it can very easily be made to work at all.

I think we can keep Escrow simply as an optional offering for sweetening eg NAP and Confed offers.  

HonoredMule wrote: On point 4:  - I'm in favor of a token fee, if only to encourage isolated players to engage others and at least get up to their ankles in the meta-game.  Players who never look outside their city walls are really missing out.  (And I'm a very anti-social person who's saying this.)

Me too.

HonoredMule wrote: On point 6:  - Alliance taxation is already optional, so I have no problem with it being included so long as the protection of all funds gathered remains limited by some factor that scales appropriately...such as the sum of all alliance members' vault levels.  To be honest, I really liked the idea of taxation because it encourages a player to invest into his relationship with the alliance, forging a bond and possibly a genuine symbiotic relationship.

You're, again, right!  If the collection mechanic is there (and now appears to be working ) there is no reason to get rid of it as it is already optional for Alliances to participate in - but we need to find the right level of protection.  I want it to be simple (and CPU cycle efficient), and getting every member of alliance's vault data does the former but not the latter.  So I'm considering a flat protection amount of around 1,000 Gold protected per alliance member.

HonoredMule wrote: Will there be a population requirement to conquer a city?  Will there at least be a cooldown before conquered cities contribute to that population? GM ThunderClap wrote: Also the newly seized city won't be fully operational at first - as clearly the population won't be fully on your side to begin with and you'll have to crush those resistence cells; also they may be emaciated from the seige etc.

ThunderClap is spot on.

The population requirement for conquering a city is the same as for settling a new city. If you don't meet the pop requirement you will only have the option to raze. 

Conquered cities will contribute to the overall population level of the Conqueror, but will be severely damaged during the Conquering phase (ie large population reductions via buildings levelling down).

Essentially, conquering a city has the advantage that you do not need to produce time- and gold-consuming settlers, but you do still have to meet the population requirements, and you do have to win the battles.  You also get a head start - with existing buildings remaining in place but substantially levelled down.  Of course, this might not be much of a head start if the buildings built here are not what you want (ie if you have a specialist purpose for the city, and so need to demolish some stuff.  Whilst it is the case at the moment that there is a plot for pretty much one of every building you can build, this won't be the case for long).

The Population requirement for each new city scales geometrically, and we think that about the maximum number of cities an end-game player will be able to reach is around 12-14.

KillerPoodle wrote: You can quell the exploits to some extent by increasing the personal cooldown timers such that leaving an alliance takes a full 24 hours to become effective and that this action is visible to other players. So, if the war is already hot the player leaving is still attackable for the next 24 hours, and they cannot attack if their alliance cannot (during the 24 war warm-up timer for example) until they are completely out of the alliance.

Trouble is that we already have exploits going the other way.

ie. Player A (not in an alliance) uses his neutrality to attack Player B (who is in an alliance)
Player A then immediately joins (or forms) an alliance, forcing Player B to declare War, pay escrow to do so, and giving Player A 24hrs of protection before war commences.

Changing the cooldown timer from 6hrs to 24hrs etc also doesn't really alleviate the problem very much.  We're all thinking in terms of 6hrs at the moment because that is the time fairly much set by the military troop distance between all the current player cities. 

There will come a point where 2 players are 12/24/36/48hrs away from each other, and there's (especially with the Stealth / Covert Ops skills) ample exploitable opportunity for one player to heavily kick another player and be out of whatever anti-exploit timer we have put in place before the retaliatory troops arrive to extract vengeance.

KillerPoodle wrote: Being able to have one city on the map that is not able to be taken is good.  being wiped off the map completely is unlikely to persuade someone to stay in game.  The city that is "invulnerable" to conquering doesn't need to be the original one though it could just be whichever one is left till last.

Is what we originally considered but is too exploitable given that Siege Encampments & Blockades need to be set up on an unoccupied square next to the city being sieged or bloackaded.  ie you could make a 3x3, 4x4, 5x5 block of cities, with all the outer cities being "semi-throwaway" alt cities as invulnerable capitals, thereby making all the cities inside the ring unsiegable and unblockadable.

I'm not against leaving the "your first city is unconquerable" rule in place if people still want it, and I guess it does provide some sort of protection in that a player who want to force you out of the game would need to permanently maintain a siege camp next to your city (thereby using up at least one army slot, and having his troops committed externally).  So I was wrong when I said there was no difference between being razed to the ground and perma-sieged.  There is a difference, although it might seem slightly moot to the targetted player.

So, the modified position is:

• Drop the concept of War Pending, the 24hr War Pending timer and War Dec Fees
  
• Keep NAPs and Confeds, and Escrow can be offered for these
• Make NAPs & Confeds enforce non-aggression, and make quitting a NAP or Confed have a cooldown timer before the enforced non-aggression lapses.  The enforced non-Aggression applies to military units only.
  
• Except as above, allow anyone in game (Allied or Neutral) to attack anyone else without warning.
  
• Increase Alliance Setup Fee to 5000
  
• Drop Alliance upkeep fees
  
• Continue with Alliance Forum development and Alliance data sharing to provide some ingame benefit for alliances
• Keep Alliance Taxation, Tax Collection, Alliance Finances for those who wish to use them
• Alliance Coffer to protect 1K gold per alliance member
  

Edited by GM Stormcrow - 14 Mar 2010 at 10:36

rescendent wrote: GM Stormcrow wrote: Except as above, allow anyone in game (Allied or Neutral) to attack anyone else without warning. Maybe have it so initial state between Alliances is neutral as no diplomatic relations exist and you have to actively declare peace? Like actively declaring war?

Think that's covered by NAPs.

Firstly, I like the idea of a distinction between what I guess are "Offensive Pacts" (ie Confederations) and "Mutual Defensive Pacts", but it's a further down the line idea.

Seondly, I think we're largely done on the original point of this post, regarding alliances, as we're now moving on to some unrelated issues (ie Siege mechanics). Please correct me if I'm wrong on this, but generally I'm taking my last proposition as what should happen with the changes.

HonoredMule wrote: Regarding vault protection:  You'd think being agreeable with so much that I say would make me happy.  Well, I'm very happy, but there's still "one more thing."  What about using a cache table that only sums vault levels per alliance once per few hours or day?  Using a hard per-player value is ok, but not great...it infers extra protection to undeveloped masses and demerit to small, mature alliances.  It's not a big deal, but I just can't see it putting that much strain on the event queue.  The values could even be maintained as  just a hash table in memory, or could be calculated only when theft of alliance property potentially occurs.  I won't presume to know what kind of architectural constraints you're having to encounter (I'm a LAMP guy, personally), but if it could be managed, it would be nice.  If there's another value more easily accessible, such as sum of player population (already maintained on the alliance score listing), that would also work. Constants trouble me, because we always often end up with problems around factors of 0, 1, and arbitrarily large n.  For example, for a static 1000 value one-man alliances still gain substantial protection for the one-time cost of alliance setup, possibly funded by a friend.  This becomes a bigger problem if you ever want to allow alliance protection to cover equipment as well as gold.  Alliances consisting of many sock-puppet accounts would become someone's personal Cayman Islands, and even real alliances may be tempted to stuff the roster a little.  Trying to root out cheaters is a necessity no matter what, but rooting out motive to cheat is more thoroughly effective.  Anyway, if you still disagree, I'll shut up...I've made my viewpoint sufficiently clear, and I'll know you've given me as much consideration as you can.

Well, I misspoke a bit. 

It's really no real overhead at all (we're talking the difference between a tiny fraction of a ms and 2 tiny fractions of a ms). 

Without going all techie, the general concept of the programming design is that the server doesn't know anything it doesn't need to know until it needs to know it.  So with every server tick that goes past I'm afraid to say it doesn't recalculate how many hundredths of a decimal place your wood level has gone up by.  It just knows what your wood was the last time it checked, and how much wood you produce. And so, when something 'initiating player active' happens (eg you refresh your screen), it does check to see whether your wood level has increased, or if something 'initating player passive' happens (eg some thieves arrive to steal your wood) then it similarly makes the  calculation. 

But generally the DB holds concepts of things, but doesn't calculate unknown-knowns unless they're demanded to become known-knowns, by the act of observation. 

I'm not sure who would be more proud: Donald Rumsfeld or Schrodinger.

So, the overhead of this would be tiny, but we're equally determined to eke out every single piece of overhead saving we can.  This server is designed to hold c. 110K active players with a concurrency of about 50K, and we'll throttle / release growth as necessary at various milestones we reach.

HonoredMule wrote: Regarding conquering cities:  Maybe I'm missing something.  Every response so far has carried the implicit assumption that conquered cities must be severely damaged husks.  Does that mean that they must be conquered via siege, or reduced to a certain population level?  Or maybe the original population migrates out and some other factors force the rebuilding process to be slow?  Because my concern is over a well developed-but non-defended city.  For example, someone builds up a city to about 500 population but does not defend it, and an all-out attack by a handful of first-tier military units gains the city.  Can someone clarify whether this is a possibility in any form?

It's a few questions at once, so I'll answer as best I can, and forgive me if I've posted this elsewhere (and am repeating myself).

A city cannot be conquered unless it is first sieged.

A Siege Encampment must be set up in a neighbouring square to the city under siege.

The Siege Encampment, once arrived, takes a minimum period to setup.  Once setup, it starts Sieging at a very substantial penalty.  During this period the Sieging party can choose to use Siege Weapons to attack but suffers massive penalties to doing so.

Over time, these penalties are reduced by x% per hour.  The actual percentage they reduce per hour will depend on the size of the city being Sieged.  The larger the city, the longer till the percentage reaches a tipping point.  As a general guideline we're definitely talking hours and in the case of large population cities, days.

With every passing hour, the Siege camp gains strength and does more damage with the siege engines, as they bed in / find range etc.  Equally, every passing hour means the defender has a chance to organise a defence of the besieged city (be that bringing in reinforcements or returning armies from abroad).

Once the Siege Penalty timer reaches a certain point, the city can be Stormed.

The Sieging party can choose not Storm the city at this point, in which case the Siege Penalty timer continues to tick upwards and ultimately into the positive, meaning that the Sieging party actually gains bonuses to their encampment's bombardment of the city, and indeed the ultimate attack.

The Sieging party does not need to Storum the city with the same army.  Any allied army may, in fact, Storm the city coming from elsewhere if the attackers so choose.

If there are no defenders in the city at this point, the city will be taken.

If there are defenders in the city at this point, the city will be fought for, and the winner will hold the city.

One decision we haven't yet decided upon is whether:
a) Defending Siege Weaponry in the City should be able to used to target the Siege Engines of the Sieging Army
b) Defending Assassins should be able to attempt assassination on the besieging army, thereby killing the commanders and actually potentially forcing the siege to end (as you can't do anything with an army without commanders). Given that armies cannot (currently) use diplomatic units as reinforcements to defend, I am against this option.

The mechanics favour the defender enough as it already is (Terrain choice, city wall, etc), so I'm still trying to decide on the above 2 items, and would welcome some thoughts.

History favours the defender of a fortified and guarded city - so much so that armies of many, many multiples of the defensive strength have smashed against the walls of various strongholds and still failed.  Illyriad is not quite so forgiving to the defender, but the defender still has a goodly advantage. 

I haven't run too many iterations of the numbers, but think an attacker (or multiple attackers) with less than (a combined advantage of) 2.5-3x the force might actually not win against a properly defended city - which makes timing of the attack, feints, other strategic opportunities critical. Whittle them down, draw them out, choose your timing, etc will make the difference here.

To specifically answer your questions, HM:

Yes, an undefended city will be taken fairly swiftly after siege begins, and the time depends on the size of the city.

Even if the Sieging party has not used Siege weapons to destroy the city first, many of the citizens of the recently captured city will attempt to deny the invaders a very good proportion of their wealth and buildings, by setting fire to lots of things in their rush up into the hills to live as outlaws and bandits. ie, taking off the RP hat, regardless of the offensive tactics used the city buildings will self-destruct themselves to a certain extent, to help ensure that "domino-effect" growth by a single player is curtailed.




Edited by GM Stormcrow - 14 Mar 2010 at 22:35