Testing Economic Balance Changes

[nathanebht]

Completed an Orthin game on hard difficulty. Upload game_8173.

With that recent patch change, this win was very easy to accomplish. The AI opponents expanded quite slowly and were behind on research. Orthin do have some nice advantages though.

[SgtArmyGuy]

Did Humans on Hard, uploaded game_8175.

...I didn't like the economy changes at all. Before the patch, the AI was good at building "wide" empires (enough ships, not that great at teching, but giving a tough challenge nonetheless). In order to build "tall" empires, the AI would need to be smarter, which it isn't. The economy changes shift the game balance more toward "tall" empires with the need to optimize coin distribution between expansion and fleet supplies. The player can easily do this, but the AI at its current state cannot.

I kinda see what you were trying to achieve: in early game both the player fleet and the enemy fleets seemed smaller and more reasonable in size (before the patch - especially with humans - the AI fleets would be HUGE by the time you met other races). You had to build markets in order to supply your fleet growth. It felt slightly more "tactical", I must admit. Too bad the AI doesn't really do "tactical", it simply does "rolls over to its back and dies". This is a result of multiple contributing factors:

-AI doesn't know how to manage its economy properly, the player does.

-AI doesn't do well in ship-to-ship combat: with all odds being equal the player always wins. Because the AI also has a tendency to retreat real easily, it now ends up giving up all its ground because it has no ships to back up its low self-esteem. I fought literally no challenging combats during the whole game (and this was HUMANS on HARD - I can only imagine how bad the situation would have been with other races).

-In mid-game and forwards, the economy changes lose all relevance. You end up swimming in money anyway, so the higher upkeep costs have no meaning to you. They still DO have impact on the AI, because their empires are now smaller than yours. As a result, you snowball them much earlier and with greater force than before.

-Higher upkeep costs in early game take away both your chances to buy upgrades for your planets and to buy refits for you ships. So the "have a smaller fleet but curate it better" idea doesn't really work, as the coin is needed elsewhere. You still can't afford to do refits. This takes away from the tactical aspect I believe you were trying to reinforce.

-Small AI fleets now get eaten by Marauders again, so this old problem has re-emerged as a result of the economy changes as well.


Overall, everything I loved about SiS (the challenge, the tough AI, and the desperate struggle to survive with Humans) seemed to be totally gone. The game felt unchallenging and simply not fun due to the AI being totally crippled now. Don't listen to the Steam forums. SiS could be (and without this patch, STILL IS) the Dark Souls of 4X games: tough but extremely fair. If you now kill the AI and start to throw in cheats for it to compensate, you kill its heart. Don't do it.

SiS has one resource for warfare: metal. Money buys TIME. It gets you stuff FASTER. You can either put it into expansion or refits. This is enough. Significantly lowering the refit costs is a sound idea, since you don't want to spend time micromanaging your ships, clicking their loadouts one-by-one and refitting them at shipyards, anyway. This both eases up micromanagement and gives room for player choice. By increasing the upkeep costs at the same time, however, you take away from this, since you can't afford to use your TIME resource (coin) as it has now become a WARFARE resource as well.

Just add the fleet supply pool from MoO2 if you really wish to limit the fleet sizes artificially? This test build kills more than it gives. For the first time in the history of SiS, I'm geniuenly scared of this experimental patch and its effects on the game.

For me, it killed the game.

[zolobolo]

Agree with SgtArmyGuy on the main message but find the situation to be very complex

Read through the tow popular negative reviews to see whats up and found it is not clear cut: most of the issues mentioned spam from lack of experience with the game (the first one, the second one is more about exploits). Note that the AI is endorsed by other reviewers which is rarely the case in 4X (almost all such games have disturbingly bad AI)

During my first 50 hours of playtime I mostly enjoyed how quickly you are allowed to ramp up ship production compared to other games, but back then the AI could not put up a big fight. So far, I did not see this having been changed as I can still churn out destroyers in early game - the AI cannot anymore. Their ship quality has increased (thanks to refitting) but are fewer in numbers

This is also the case in Hard difficulty - still going through a test game there, but the AI has less ships in Hard then on normal before the change to upkeep

That being said, it is more capable on hard (for obvious reasons) but not as much as on normal before

Before I post my full analysis on hard, let me put this forward:
If the main problem is difficulty for some players on normal, is it the best way to approach this my changing a core mechanic?

I have nothing against tweaking core mechanics but only as long as the difficulty does not decrease by it. This is a genre where it is absolutely possible to offer endless levels of difficulty by fine-tuning parameters. A simple blunt solution could be to:
1. Easy turns down AI aggressiveness against both AI and player (e.g: base diplomacy relationship bonus). Note that it does not alter the resource gain of the AI - by reducing its overall aggressiveness, this would lead to smaller empires (less conquest) and consequently less overall resources compared to player but leaves all core game mechanics unchanged
2. Normal turns down AI aggressiveness against player only (AI empires would thus be allowed to grow but still unlikely to go against the player - does not help of course if the player decides to step up against an empire or alliance that is much stronger )
3. Hard (no tweaks, no bonuses whatsoever) - give a hint here for players that this is the real deal and the AI will come at you mercilessly but without bonuses. Most players would probably stick with normal and if they chose hard, the name says it all
4. Hardest (bonus to AI) - for those who would like to play the systems
Can also introduce new difficulty levels in between, or sliders giving direct control over each aspects, there are options
Currently the icons and color or normal suggests that it is unaltered so that is a problem of course with the above concept

Again, I support adjusting the upkeep system but also propose improving the AI scripts to use the new system at least as effective as before the change and address difficulty via other means, else the nerfing of the AI reduces the challenge on all difficulty levels and needs a lot of workaround to keep above the water especially when other systems are touched upon additionally

We would all like the game to succeed and get extended and that the systems that have not yet gotten enough attention to get treatment but it should be considered that the game has a strong challenge without apparent bonuses to the AI which is something that can set it apart from other titles and that this stance has come with a lot of effort already put into the game by the developers and others

[zolobolo]

One more thing to note: It seems like a good idea to revise the coin system as before this, buyout and refitting was much more easy for the player

This has lead to a lot of micro as in each turn the player has been buying out improvements (I have rarely waited for a mine or market to be built normally) and ordering refit. Reducing the amount of these actions should improve the flow of the game but since refit was already expensive it makes sense to address the abundance of coin in general instead

Regarding AI: It has classically been neglecting building markets at all and only recently starting building them at all.
My assumption at this point is that the foreign trade routes that should be generated via diplomacy do not actually apply - if this is the case this would also help the player as well of course but might help the AI get back in track via bugfix. Long-term they need to build a bit more markets and this will of course come at the price of other resources

[zolobolo]

The changes might be messing with the build priorities when starbase is involved as the only really successful empire on hard (Tinker) has started building two starbases in almost all of its sectors and prioritizing mining in 80% of its planets probably to stack up the additional metal to build new bases

Just a hunch but might be due to the now huge difference in upkeep between ships and bases: maybe the AI thinks this is a more cost effective investment overall (which is true from coin perspective but not from metal and overall war effort perspective)?

BTW: Seems like I have recently passed 1000 hrs of SiS playtime on steam

[zolobolo]

Suggest to have tinker forge only restore munition, system and hull but not armor

Reason: Restoring both system and armor seems like a double bonus for the same thing. System is the vital part so restoring this still gives them the intended edge but does not produce unbalanced cases where armor is high enough to absorb all damage and even if some gets through, hull restores on top of that (so the bonus actually stacks and applies double be reducing damage twice)

This is not a problem for player focusing fire and using AP, but gives a huge advantage to tinker ships over other AI

[sven]

Did Humans on Hard, uploaded game_8175.

...I didn't like the economy changes at all.

Thanks for the testing, and the detailed feedback. The difficulty posed by the AI has been swinging around a lot recently, as I've been experimenting with different options for making it play better under the high-upkeep balance. On Sunday, I had a developmental build where I thought the AI was clearly too difficult, so I nerfed it heavily on Monday, removing all the AI's starting resource bonuses on 'Hard', and shifting 'Brutal' down to where 'Hard' had previously been. That was a dramatic set of changes -- probably too dramatic, in retrospect.

I think a lot of our impressions of how well the AI is playing are tied to how quickly it expands in the early game, and that, in turn, is very much a function of how many starting bonuses I give it. I'm about 2/3rd's of the way through a 'Brutal' playthrough with Humans under the latest balance, and the AI has felt ok to me, challenge wise, though maybe a bit easy for the top level difficulty.

All of which is a long-winded way of saying that, insomuch as the problem is mostly that the AI isn't feeling challenging anymore, that may actually be fairly easy to fix. Just putting the starting advantages back to about where they where before r38045, and making the new 'Hard' closer to the current 'Brutal' might about do it. I certainly appreciate that you like where the balance is in the current steam build -- but, if you peak behind the curtains, the truth is that the AI has never really been "extremely fair"; there's always been some amount of "cheating" involved in trying to make it feel about right.

What has me more concerned are your observations that there are various ways in which the game is feeling less-fun as a result of the high coin upkeeps. When the concept for this patch was coming together, Arioch and I did a lot of hemming and hawing about the different options available to us for cutting down on fleet sizes. We did consider a Moo2-style "soft" fleet cap, as well as a metal-based upkeep cost. High upkeeps in pure coins seemed like the simplest of the various possibilities, and it's a system that does seem to work fairly well in Civilization, so we gave that a try.

But the early game is arguably the heart and soul of a good 4X, and high upkeeps do cut down on the amount of free coins you have to micro your colonies in the early game -- especially if you're playing cautiously, and don't neglect your military over the first 100 turns or so. Maybe that means a pure coin upkeep isn't the right solution for us. I'll need to think this over.

In the meantime though, I'd be curious what you think of the current 'Brutal' balance. It's probably playing closer to the old 'Hard', and insomuch as the AI behaviors there still strike you as problematic, I'd like to know about it.

My own impression is that AI has always had big problematic blind spots, and I think those are hurting the game no matter how I balance it. Mistakes in fleet routing are the category of issues that most has my attention right now -- particularly in mid/late game, when you have large empires and/or big alliance blocks, the AI's tendency to send ships scrambling from one side of the map to the other, missing out on all the critical battles that are taking place while they're in transit, is a real problem. That's true under the old balance as well, though it's perhaps more pronounced now, because games are lasting longer. I'm going to try and do something about that issue as soon as I can find time for it. I do think there's a chance that once the routing issues are fixed, the high upkeep game will start playing fairly well. But, if people other than me don't like the way these changes are headed, that's certainly something I want to keep hearing about.

[SgtArmyGuy]

The changes might be messing with the build priorities when starbase is involved as the only really successful empire on hard (Tinker) has started building two starbases in almost all of its sectors and prioritizing mining in 80% of its planets probably to stack up the additional metal to build new bases

Just a hunch but might be due to the now huge difference in upkeep between ships and bases: maybe the AI thinks this is a more cost effective investment overall (which is true from coin perspective but not from metal and overall war effort perspective)?

BTW: Seems like I have recently passed 1000 hrs of SiS playtime on steam

Oh, I forgot to mention in my feedback that this "starbase anomaly" was also present in my game. Two starbases in almost every system that I encountered. Weird behavior.

[nathanebht]

Didn't realize that hard difficulty had been made easier and that brutal was the new hard.

Think it would be better if there was a game creation setting for AI smarts and another for the AI getting free resources. Allow people finer control of their difficulty.

[sven]

If the main problem is difficulty for some players on normal, is it the best way to approach this my changing a core mechanic?

It's probably worth clarifying that I'm experimenting with changing the economic balance because I think changing the economic balance is a good idea. In particular, I'm interested in how the game's economic balance ends up affecting fleets sizes. To take two classic 4X games as examples -- Moo2 is balanced for relatively small fleet sizes, where individual ships can have a fairly big impact on the overall strategic situation. Sword of the Stars, in contrast, is balanced for huge fleet sizes, where individual ships tend to get lost in huge blobs of dozens or even hundreds of ships crashing into each other.

The balance of SiS has always been a sortof hybrid of things I liked about SoTS and things I liked about Moo2, but, as far as fleet sizes go, my preference has always been for a Moo2-ish balance, which emphasized the importance of individual ships, over the SoTS-ish balance, which is more about managing competing swarms.

SiS has a Moo2-like tactical game, but, pre-patch, it's economic balance is tending towards almost SoTS-like fleet sizes. I think that's a poor combination. After the early game, players end up being encouraged to just autocombat their way through most battles, because the fleet sizes have progressed well beyond anything that's fun to manage by hand.

So, as I said, I've been thinking about various ways to try and get fleet sizes under control. And I still think high coin upkeeps might end up being a viable route.

Difficulty is really a separate topic. And it's a tricky one, because what I end up doing is to tune the game so that there's one difficulty that I enjoy playing on, and then sortof guess at sensible parameters for all the others. Most recently, most of my playtesting, and thus most of my tuning, has been focused on 'Brutal', which was probably a mistake. Given the community of players we have, I should probably be tuning mostly for 'Hard', and then defining 'Brutal' and 'Normal' relative to that.

[zolobolo]

Maybe that means a pure coin upkeep isn't the right solution for us. I'll need to think this over.

I think the approach of higher upkeep is correct in order to reduce buyout micro and abundance of coin from mid-game on

It should also reduce the cases where AI is spamming battleships without escorts which should be a big plus and make AI more competitive

So far what from what I can tell:
1. The AI being stalled does not derive from the fleet upkeep cost
2. Foreign trade routes do not apply cutting the income very short for AI which tend to build much much less markets then the player hence their relative weakness
3. AI is concentrating on building starbases all over the place and seems to be saving up metal to do that (not just tinkers) but only becomes visible after 200 turns or so for smaller empires. This might come from the now high upkeep and hence relative priority of starbases but is a build logic error in its core

[zolobolo]

SiS has a Moo2-like tactical game, but, pre-patch, it's economic balance is tending towards almost SoTS-like fleet sizes. I think that's a poor combination. After the early game, players end up being encouraged to just autocombat their way through most battles, because the fleet sizes have progressed well beyond anything that's fun to manage by hand.

We are testing on different difficulties, yes
On Normal, I didn't have issues with massive fleets - even after 200 turns. What was annoying is fighting 6-8 battleships with no other ships in the fleet but this should be resolved with the higher upkeep.

I am also no against big fleets per say. It is normal not wanting to play each battle after early game: on a huge galaxy you can have 3-4 battles each turn with minimal stakes and 1-2 big battles. Naturally you will autoplay most of them

[zolobolo]

The stall in research is also present in Hard: at turn 200, almost all AI empires still only have PD coilgun and nuclear missiles

They seem to avoid advanced lasers, heavy weapons and rockets/torps when selecting research - not sure yet if this has something to do with the new con balance in some way

[zolobolo]

The balance of SiS has always been a sortof hybrid of things I liked about SoTS and things I liked about Moo2, but, as far as fleet sizes go, my preference has always been for a Moo2-ish balance, which emphasized the importance of individual ships, over the SoTS-ish balance, which is more about managing competing swarms.

Sorry I haven't played either of them but would recommend you stay in between: it is fine to manage manually at the beginning and then autoplay and only control the fleet when deciding battle is done this way the player experiences both styles and the game is more interesting by it. If you limit it to either of the options, you limit the potential experience the player can have

If someone just wants small scale battles, they can play tiny galaxy, but if you select to have 200 stars, you expect to be able to field and fight hundreds of ships (even if not all of them at once )

A fine example was UFO the original: all the smart heads say it was popular because you could name you soldiers and manage them individually, but you know how me and pals use to play it?: Send dozens of soldiers everywhere with multiple fire teams to cover each other. Losses of 30-60% were common, but I didn't even notice their experience matters until the 3rd or so play-through and didn't matter: for me, the game was large-scale global invasion simulator and always build like 6 bases with 3 troop transports It was fun throwing soldiers at the enemy and dint worry if yellow hear dude survived

[Dragar]

My own impression is that AI has always had big problematic blind spots, and I think those are hurting the game no matter how I balance it. Mistakes in fleet routing are the category of issues that most has my attention right now -- particularly in mid/late game, when you have large empires and/or big alliance blocks, the AI's tendency to send ships scrambling from one side of the map to the other, missing out on all the critical battles that are taking place while they're in transit, is a real problem. That's true under the old balance as well, though it's perhaps more pronounced now, because games are lasting longer. I'm going to try and do something about that issue as soon as I can find time for it. I do think there's a chance that once the routing issues are fixed, the high upkeep game will start playing fairly well. But, if people other than me don't like the way these changes are headed, that's certainly something I want to keep hearing about.

I have a lot of thoughts, but I'm going to keep them mostly to myself for now. I'll come back to why in a moment.

I think you're focusing on the right thing here. The fleet routing/transit issues hurt the AI a great deal, and smaller fleet sizes may even exacerbate this issue.

I'm not going to give a detailed review of the economy changes or thoughts of fleet sizes, because I think until some of these issues have been resolved somewhat it's really impossible to give meaningful feedback. It doesn't really matter what the AI is doing or how big fleets are if the AI can never bring its forces to bear on an opponent. As a completely singleplayer game with simple, board-game like mechanics (a good thing!), SiS leans perhaps more on the AI for a good player experience than other games in the genre do.

My own preferences for SiS are smaller battles to larger ones, and mixed fleets to nothing but one ship class. Unlike SotS, battles get kinda unmanageable with too many ships. But that makes individual ships more interesting, and potentially much more so given some development there.