Author Topic: AI Progress scaling question  (Read 3231 times)

Offline Saan

  • Newbie Mark III
  • *
  • Posts: 31
AI Progress scaling question
« on: February 18, 2010, 11:56:06 am »
Just a general question about the effect that AI progress has as the game goes on - is it more of a flat rate of growth when you gain ai progress, or is it more of an exponential progression?

In other words:

Example A- Flat :  Your AI progress is at 10 and you gain 1 progress.  Instead of waves of ten ships you get waves of eleven ships.  (i know this isnt the right number of ships, and it is highly contigent on numerous different factors, but you get the idea).

AI Progress is at 100 and you gain 1. Instead of waves of 100 ships you get waves of 101 ships.

Example B- Exponential:   Your AI progress is at 10 and you gain 1 progress.  Instead of waves of 10 ships you get waves of 11 ships.  

AI Progress is at 100 and you gain 1. Instead of waves of 100 ships you get waves of 110 ships.

Or, a simpler way to ask the question, i guess:  Is 20 AI Progress twice as hard as 10? Is 200 AI Progress twice as hard as 100?

I hope I've been clear enough, heh.
« Last Edit: February 18, 2010, 11:59:32 am by Saan »

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Re: AI Progress scaling question
« Reply #1 on: February 18, 2010, 12:42:30 pm »
It's pretty much a linear growth, but with some complexities.  Basically, the number of ships goes up in a linear fashion based on the AI Progress.  The exact number of ships added per AIP depend on the difficulty level you are playing on, the types of AI, etc.  Suffice it to say, the number of ships in a wave (or a reinforcement event) are based on a variety of multipliers, and AIP is one of them -- perhaps the most significant one.

Then, however, the AIP also affects the tech level of the AI, as I'm sure you know.  When the tech level goes up, there is a corresponding drop in the number of ships it can then build.  That is based off the same number of how many fewer percent ships you can build of higher-mark ships as a human player, also.  So when the AI suddenly jumps up to Mark III from Mark II, or so, the size of the waves drop a bit even though their composition is tougher.  Then as the AIP increases from there, you get progressively harder in the new tech level, too.

By the time you pass something like 1000 AIP, you wind up with all Mark IV ships, and their numbers just keep climbing after that.  If you're able to survive up to 4000+ AIP, most likely it's not going to get much harder in that particular game -- generally by then all of the planets are filled with 4000+ Mark IV/V ships, the AI is sending hugely brutal waves and CPAs, and so forth.  A few players are able to get to that point when trying to (for instance) take all the planets in a large map, but it's a challenging scenario for a lot of reasons, and generally those players are on difficulty 5 or 6 for one thing, as well as having a certain style of turtling tactic that is letting them withstand those crazy waves at key defensive locations or similar.

Anyway, those are the extreme cases, which is not what you really asked about, but I think it helps answer your question.  In general, the AIP is pretty linear.
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline Saan

  • Newbie Mark III
  • *
  • Posts: 31
Re: AI Progress scaling question
« Reply #2 on: February 18, 2010, 04:36:28 pm »
Makes sense - so, with that logic, it works a bit like a flat (or regressive) tax in economics, in that a 1 point AI gain early on is actually greater than a 1 point gain later, especially when you consider that, economically, no one planet you gain is going to double your income, as your best planet is your home planet (especially early on).

So really, you dont need to be in TOO big of a rush to expand quickly early on - sitting back, scouting, and developing a plan for your first five planets or so seems to be pretty beneficial. At least, with how i play(MKIII scouts and MKII command stations off the bat). :P

Thanks for the info!


Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Re: AI Progress scaling question
« Reply #3 on: February 18, 2010, 04:45:10 pm »
Makes sense - so, with that logic, it works a bit like a flat (or regressive) tax in economics, in that a 1 point AI gain early on is actually greater than a 1 point gain later, especially when you consider that, economically, no one planet you gain is going to double your income, as your best planet is your home planet (especially early on).

This is a very good point -- I had not thought of it that way before!  However, I would point out how the reinforcements work.  Those, also, get stronger as you take more planets, but they also (usually) get more diffuse.  The diffusion has a greater effect when the map is really open (Grid, Hubs), but has almost no effect on maps like Tree, Vines, and especially Snake.

So really, you dont need to be in TOO big of a rush to expand quickly early on - sitting back, scouting, and developing a plan for your first five planets or so seems to be pretty beneficial. At least, with how i play(MKIII scouts and MKII command stations off the bat). :P

Yes, absolutely so.  You could play for a good 10-20 minutes with just scouting if that suits you, and the game won't run away with the reinforcements or anything.  It will make capturing those very first mark I planets next to you a tad bit harder (300 units instead of 50 at each, or something like that), but it's nothing notable and that's about the only negative effect.  Then, beyond that, it's always better to do something sensible in whatever time it takes, rather than just flailing and rushing around and making things worse by lack of planning.

In general, the game was designed to be something of a middle-ground, taking away the advantages of both turtling and rushing.  If you try to turtle, the AI will also be reinforcing, and eventually that is just not a good strategy.  If you rush early you can snap up a few planets very quick, but then you hit a roadblock before long where you have to step back and take a more measured tempo if you want to win.  There is a pretty big gradient there with how people are able to play even so, but I specifically wanted the "normal" way for this game to be played to be more of a thoughtful 4X-style experience rather than your typical RTS pvp rush-fest.  And I'm a turtle/boomer by nature, so I also wanted the game to kick me out of the nest, so to speak.  It's a lot more fun to actually play, rather than turtling for 40 minutes for then 20 minutes of glory.  But in most RTS structures, I don't like feeling so exposed when I don't turtle, hence the problem; and hence the middle ground of AI War.

Thanks for the info!

My pleasure!
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline triggerman602

  • Sr. Member Mark II
  • ****
  • Posts: 366
Re: AI Progress scaling question
« Reply #4 on: February 18, 2010, 05:39:29 pm »
wait are there always markI planets next to your home system because i think iv seen markIVs neighboring me which are always bad because by the time youve gotten strong enough to even tread in a markIV system its bloated to over 4000 units.

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Re: AI Progress scaling question
« Reply #5 on: February 18, 2010, 06:20:22 pm »
wait are there always markI planets next to your home system because i think iv seen markIVs neighboring me which are always bad because by the time youve gotten strong enough to even tread in a markIV system its bloated to over 4000 units.

Anything that hasn't been upgraded to III or IV that is next to your home planet will be I, rather than the standard II.  Referring to diff 7 and thereabouts, specifically.  Any IIIs or IVs that are next to your home will be weaker in that they won't have an ion cannon on them.  But, that is still a challenge since they are right next to you from the start, though.  Part of the variety in the scenarios that can be generated!
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline triggerman602

  • Sr. Member Mark II
  • ****
  • Posts: 366
Re: AI Progress scaling question
« Reply #6 on: February 18, 2010, 06:41:30 pm »
oh ok thanks.

Offline Doddler

  • Sr. Member
  • ****
  • Posts: 287
Re: AI Progress scaling question
« Reply #7 on: February 19, 2010, 12:17:37 am »
Anything that hasn't been upgraded to III or IV that is next to your home planet will be I, rather than the standard II.  Referring to diff 7 and thereabouts, specifically.  Any IIIs or IVs that are next to your home will be weaker in that they won't have an ion cannon on them.  But, that is still a challenge since they are right next to you from the start, though.  Part of the variety in the scenarios that can be generated!

It's pretty lame when you start with only Mk IV systems adjacent to you at the start though. :(

Offline triggerman602

  • Sr. Member Mark II
  • ****
  • Posts: 366
Re: AI Progress scaling question
« Reply #8 on: February 19, 2010, 01:07:13 am »
that would be really lame. id probably just quit and get another map.

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Re: AI Progress scaling question
« Reply #9 on: February 19, 2010, 09:35:33 am »
Anything that hasn't been upgraded to III or IV that is next to your home planet will be I, rather than the standard II.  Referring to diff 7 and thereabouts, specifically.  Any IIIs or IVs that are next to your home will be weaker in that they won't have an ion cannon on them.  But, that is still a challenge since they are right next to you from the start, though.  Part of the variety in the scenarios that can be generated!

It's pretty lame when you start with only Mk IV systems adjacent to you at the start though. :(

There are hidden benefits, though. Since there are only so many upgraded to mk4 on every map, that means the rest of the map is easier. And since they start out next to you without ion cannons, it makes them easier than they would be off elsewhere. Still a harder scenario, but it's not just bad.
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!

Offline Ruges

  • Jr. Member
  • **
  • Posts: 56
Re: AI Progress scaling question
« Reply #10 on: March 18, 2010, 04:36:04 pm »
Quote
If you're able to survive up to 4000+ AIP, most likely it's not going to get much harder in that particular game

So 4,000 AIP is the softcap? We have a goal in our current game to break 10,000 AIP. But your saying we will not notice a difference between 4,000 and 10,000 AIP?

Offline x4000

  • Chris McElligott Park, Arcen Founder and Lead Dev
  • Arcen Staff
  • Zenith Council Member Mark III
  • *****
  • Posts: 31,651
Re: AI Progress scaling question
« Reply #11 on: March 18, 2010, 04:41:30 pm »
There isn't a real softcap on AI Progress itself.

However, regarding reinforcements, when AI planets fill up with ships, they simply can't add any more.  They will upconvert them over time, until you wind up with every planet simply having around 4,000 Mark V ships, and then there's nothing else for the AI to do reinforcement-wise.  Playing normally, I imagine you'd hit that point long before 10,000 AIP.

Regarding waves, there is a hard cap on the size of how big an individual wave can be.  So if you're already getting Mark V waves of the largest size, then it doesn't matter what your AIP is, because the wave can't get any bigger.  I think that's something like 2400 ships in one wave or something along those lines.  (Note that that's per line item in the incoming wave list, which of course often contains 2 items in solo, or 16 items in 8-player).

Regarding CPAs, I don't think there's a cap on that.  So those would just get increasingly terrifying, I guess.  To my recollection.

To be honest, that sort of thing isn't something I've thought a lot about because the general expectation is that we're talking about a game in a state many hours after players should have died. :)
Have ideas or bug reports for one of our games?  Mantis for Suggestions and Bug Reports. Thanks for helping to make our games better!