Testing Economic Balance Changes

[sven]

On the topic of ship upkeep cost and number of ships it seems that the problem lies in the fact that upkeep cost is a flat number while income grows in multiple ways.

If there is a worry about that pricing modules directly will lead to builds heavily favouring offence what about power 2 for Tech Tier weighting? So that higher tech contributes larger part of the ship maintenance than lower tech.

I've been hemming and hawing all morning about whether or not I think a nonlinear part pricing model is worth doing. I've pulled more numbers from an AI vs. AI test game. "Balance" has always been pretty rough in SiS, particularly as it exists between the different factions, and while that roughness is certainly in evidence in these spreadsheets, I'm not seeing clear enough anomalies to be tempted to do a serious upkeep rebalancing pass. Sure, in very late game, Orthin Gunships get some crazy good offensive efficiency numbers, but that's actually more or less working as intended. Some of the Yoral destroyers are also very effective "glass cannon" type designs, but, again, that's arguably "working as intended".

Based on the playtesting I've done, my general sense is that the upkeep balance is actually in an ok sortof place in the "in_development" build. I think I can see some relatively modest changes that ought to be made to be made to the income mechanics -- as zolobolo's posts suggest, markets probably need a higher understaffing penalty. And as other players have mentioned, the +2 trade cap bonus granted by each outpost is probably far too high. I've tried to address both these issues in the diff set for r38135, but, I don't expect the changes I'm making there to have huge impacts on the current balance.

[sven]

But no, carreirs are often damage soaking HEAVILY shielded monsters that go down way after all of their squadrons have been oblitirated.

Yeah, I think there's probably a balance problem here. The Ashdar Fleet Carrier in particular is a very "tanky" hull, with abnormally high defensive potential vs. it's offensive punch. And while there's nothing wrong with having a couple tanky hulls in the game, it's a trait that kinda goes against the concept of a "Fleet Carrier" (the very impressive combat record of CV-6 notwithstanding). Adding a 4th hanger in trade for a system slot probably wouldn't solve the issue entirely, but I think it could be a sensible tweak.

[Zoolimar]

I've been hemming and hawing all morning about whether or not I think a nonlinear part pricing model is worth doing. I've pulled more numbers from an AI vs. AI test game. "Balance" has always been pretty rough in SiS, particularly as it exists between the different factions, and while that roughness is certainly in evidence in these spreadsheets, I'm not seeing clear enough anomalies to be tempted to do a serious upkeep rebalancing pass.

Oh, I don't worry about balance between factions. Just the total number of ships on the field needed to keep parity with the AI. Sometimes you get into a position where it is hard to spread over the AI territory and as a result need to fight battles like 30 vs 40 ships. But it is pretty rare and using Ellipse galaxies seems to help.

Actually why is the formula for ship power = attack*(1+defence)? I don't think you can have 0 defence on the ship as all ships even with armour stripped still get hull points.

[sven]

Actually why is the formula for ship power = attack*(1+defence)? I don't think you can have 0 defence on the ship as all ships even with armour stripped still get hull points.

My guess is that harpyeagle wasn't certain there might not be some degenerative case where defense<1, and wanted to make sure the formula didn't ignore relative differences in attack power if that happened. Given the actual numbers we're getting in game though, the equation might as well be defense x attack. Arguably, I think you can make the case that because, statistically, some of your ships will always get a chance to fire before the enemy ever brings any weapons to bear, something like attack*(c+defense), where c is some fairly large constant, might make almost as much sense as attack x defense. (Or you could get really clever, and make 'c' a function of effective weapon range.) But; as we've both said, you really need to tune all the details of these formula carefully against the results of a large number of different carefully chosen mock battles to have any real sense of how accurate any heuristic is. I haven't done that work myself recently -- but, harpyeagle did, and I'm happy to use the formula he came up with.

[sven]

I haven't done that work myself recently -- but, harpyeagle did, and I'm happy to use the formula he came up with.

One pattern you may notice jumping out of the numbers is that harpyeagle gives very large "attack" scores to ships with assault shuttles or other "combat boarding" equipment. I think this is probably roughly correct -- boarding is a very strong mechanic -- though in a strict tactical-power sense, I'm not certain it's quite as strong as his formula suggest.

[sven]

Sometimes you get into a position where it is hard to spread over the AI territory and as a result need to fight battles like 30 vs 40 ships.

I'm guessing you're playing very large maps? On 88 star maps, my experience has been that battles are generally maxing out at around 20 ships per side. If this isn't in line with other people's experience, that would be good to know about.

[Uncle_Joe]

I'm playing using 45 stars, 4 races as Ashdar Colonial on Normal. I like the economic tweaks to the system and now I feel like I actually need to build Markets and manage cash flow for the first time.

The reduced ship count does seem to increase the effectiveness of defenses (Starbases and Planetary Defenses). I can generally only afford to keep one fleet capable of battering through a defended planet. That is more in keeping with MOO2 (500lb gorilla stack) but it does remove some of the 'epic' feel of having multiple fleets sweeping across the galaxy.

I'm still re-learning the mechanics as it's been a while since I played seriously but I'm definitely liking the way the game is going.

[zolobolo]

Here is a quick status at turn 200. The player is producing only destroyers with coilgun and light cruisers with missiles
Normally this can be countered with fleets boosting both PD and firepower but as we can see the AI canot progress most of the time due to some sort of limitation

Since the production of cheap units does not require advanced tech, labs do not need to be built.
Mines are not needed in large numners, there is an abundance of it
Less factories are needed, starbases, PD and shipyards can be saved as the player can cover all area with swarms of ships that cannot be countered by 2-3 medium ships the AI can muster up at a given location - and if it does, the player can swarm its other systems in the meantime

Unused improvement slots can be used to build markets

There is no shortage of cash, due to the combination of more markets and low-upkeep ships

The AI did try to put up a decent fight though

[zolobolo]

Adding a 4th hanger in trade for a system slot probably wouldn't solve the issue entirely, but I think it could be a sensible tweak.

Sounds interesting and fun: they would pack a huge punch but be actually quite fragile compared to their offensive power

I was also wondering why Phidi have more powerfull carriers then Haduir

Another solution woul be to have some other system module that is more valuable (tempting) but I usually end up with shield like probably everyone else

Generally I think that no ship should really be incentivised to equip more then 2 shields due to the recharge rate progression and sicne they already start from a higher base number, but if such a rebalance is done, it makes sense to increase their offensive power as well to keep the balance seen in the relative advantages graphs

[sven]

Here is a quick status at turn 200. The player is producing only destroyers with coilgun and light cruisers with missiles
Normally this can be countered with fleets boosting both PD and firepower but as we can see the AI canot progress most of the time due to some sort of limitation

This seems like an interesting theory-crafting run. Looking at it more closely might also reveal undiscovered bugs in the AI. Could you upload this game?

[sven]

Since the production of cheap units does not require advanced tech, labs do not need to be built.

This is an interesting strategy. I've been worried that destroyer swarms might be too strong in the mid game, but, if you're using them this successfully with tier 1 techs, that actually suggests that the "problem", if any, is that light ship upkeeps may just be too low across the board.

The AI did try to put up a decent fight though

Depending on the difficultly level, this may suggest that there's not actually any real balance problem here. I think a player with your level of experience is likely to be able to roll through most 'Normal' maps without much difficultly, and if you can do so using a low-tech strategy, as well as playing more conventionally, that doesn't necessarily mean the game balance is broken. Particularly if you felt like the AI was offering reasonable resistance -- this could be a "working as intended" case.

In my mind, the test of a truly "broken" strategy is that using it makes the game feel like a cakewalk on 'Brutal'. By this metric, a Tinkers missile ship swarm may be boarderline broken -- and I wouldn't be surprised if Yoral torpedo destroyer swarms are in a similar category. I'd be surprised if a low-tech Ashdar CL rush is on the same level though.

[zolobolo]

The reduced ship count does seem to increase the effectiveness of defenses.

it does remove some of the 'epic' feel of having multiple fleets sweeping across the galaxy.

Yes, these are on my short list as well:
1. Smaller scale (especially on huge systems the discrepancy between empire size and military flexibiltiy/capacity is considerable)
2. OP planetary defenses
3. Slower gameplay (along a number of factors)
4. Less competitive AI and/or increased AI bonuses to coupe with the system
5. Metal becoming irrelevant
6. Making trade and the very cool trade pool and trade transport production less relevant
7. No money for anything else like refit, or buying stuff: Since building less ships is probably suicidal especially when the overall number is too low to cover much area, the remaining surplus income should always be filled with ships
7. Making either small or large vessel unviable (depending on the balance direction)

Galciv2 and in part 3 was a particular game where one could win a theater with a single ship
MOO4 did not particularly have fleet size limit as far as I can recall, but the game was too small scale uneventfull to really leave a mark
Stellaris had fleet limitation but large ships were not viable there - stopped playing afterwards so don't know abouth the new balance changes

What if there were a galaxy map setting for fleet size limits: Unlimited (default), 1 ship/colony, 2 ships/colony and 3 ships/colony?

This way, everyone is free to limit the number of ships for performance, GUI or preference reasons

Counting the number of ships and colonies should be easy enough: these numbers arealready being calculated each turn, but of course it would need to be figured out how the limit is applied

[Zoolimar]

Well current dev versions upkeep costs seem to not be that restrictive. On humans with 6 planets I had 6 cruisers (3 light, 2 heavy, 1 assault) and could add 2-3 more. That's with level 1 markets. Then Gremak came and put a dumper on my attempt to ally with Colonials and get out of the corner.

[sven]

Well current dev versions upkeep costs seem to not be that restrictive. On humans with 6 planets I had 6 cruisers (3 light, 2 heavy, 1 assault) and could add 2-3 more. That's with level 1 markets. Then Gremak came and put a dumper on my attempt to ally with Colonials and get out of the corner.

Yeah, humans are a bit of a special case now, because I've added a racial perk for them that drops all ship upkeep costs by 50%. Given how relatively difficult their starting position is, my hope is that this isn't too wildly unbalanced. But, it does mean that human empires can punch well above their weight in terms of maximum ships active at once.

More than that, it usually means that if you're playing as humans, you probably won't "feel" the impacts of ship upkeeps nearly as much as you would as one of the other races. (My own experience with humans has been that I only really notice their upkeep constraints if I really go nuts capturing pirates in the early game.)

[zolobolo]

This seems like an interesting theory-crafting run. Looking at it more closely might also reveal undiscovered bugs in the AI. Could you upload this game?

Sure, uploaded: game_8317

Findings from my side on coin:
- Orthin seemed to have been focusing on conducting research projects for a while instead of producing ships even though they had the income
- Was surprised to find how many of the buildings I usually need became obsolete with this strategy: could have saved like half of the mines I ended up building
- For the little upgrade I did, there was always enough coin and metal to upgrade everything
- With around 30 ships I did not find these to be a huge amount considering the amount of planets, systems and turns passed (almost had no losses at all so I produced around 40 ships tops over 200 turns)
- Did not close trade deals with AI to not boost their trade income and mine for a valid test
- Did not attack the AI to see how far they can/will build up when left to develop - otherwise the game would have been over at around turn 100
- When I say the AI was putting up a fight what I mean is that they have pulled together decently upgraded ships for what was there, and they focused them into a single attacking fleet of various ships types. But they were lacking any sort of defensive ships, or reserve which opens them up for quite a lot of nasty surprises. On larger maps this behaciour will cause issues for them as all the pressure seems to be going on one target

Additional findings:
- At one point Orthin have offered up 3 systems fo peace! I strongly suggest to not have the AI offer up any systems or jsut as a last resolve. Metal and coin is better as it does not hurt their economy so badly: though considering the coin balance direction it might
- Orthin got "Error at the food shortage report formatter" error right at the savepoint
- Orthin do not yet want ot research Shield capacitators and use them in their designs