Stellaris Dev Diary #96: Doomstacks and Ship Design

  • We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.
Huh. Overall the changes look good. The only one that has me nervous is the Fleet Disparity Bonus. It does seem pretty artificial, and I'm worried it might shift optimal gameplay into counter intuitively engaging larger fleets with smaller ones and winning solely because of that bonus. It would feel more natural/make more sense to me if it were part of one of the new military doctrines, like "Guerilla Warfare" or something. If smaller fleets engage larger fleets SPECIFICALLY with the purpose of inflicting maximum damage before disengaging, a Fleet Disparity Bonus would seem much more appropriate in that instance.
 
Huh. Overall the changes look good. The only one that has me nervous is the Fleet Disparity Bonus. It does seem pretty artificial, and I'm worried it might shift optimal gameplay into counter intuitively engaging larger fleets with smaller ones and winning solely because of that bonus. It would feel more natural/make more sense to me if it were part of one of the new military doctrines, like "Guerilla Warfare" or something. If smaller fleets engage larger fleets SPECIFICALLY with the purpose of inflicting maximum damage before disengaging, a Fleet Disparity Bonus would seem much more appropriate in that instance.

If I understand the whole diaries correctly the fleet disparity bonus is just a subtle "war attrition" mechanic implementation that is complementary with the warscore rework.

It simply means that even when you win with superior numbers you still suffer some damage. Which turns into war exhaustion which in the end forces you to stop and make peace eventually.

Without this workaround the victor would just move his doomstack around, occupy everything without ever suffering damage and only then ask for peace. Instead by suffering some damage at every won battle he should more or less quickly arrive to a point where he is forced to accept peace.
 
For this reason we have decided to introduce something called the Force Disparity Combat Bonus. The Force Disparity Combat Bonus is applied when a smaller force is engaged with a larger one in battle ('force' being every ship engaged on one side of a battle, regardless of how many fleets and empires are involved on each side), and gives a bonus to the firing speed of all ships belonging to the smaller force.

What an incredibly immersion breaking mechanic, just wow!
 
People act like the combat bonus will change the victor in a fight.

The expected result from a 15 vs 14 battle is 11-13 survivors on the team of fifteen, because as members of the losing team die they take more concentrated fire and can't kill the larger team's forces as easily, and one of them has free range to do whatever they want unmatched. leading to them taking more casualties, leading to more focused fire and less damage output, lading to a fast spiral that at some point leads to complete collapse. This is exacerbated in stellaris because of regenerating defenses and targeting. Meaning the fights are often total stomps with negligible damage to the winner

That is the problem. You expect a fleet that is 90% as strong as another to be able to cripple the stronger side at least, and take down most of their forces, but in the current situation it works out that a fleet that's 10% weaker gets completely annihilated inflicting less than 10% losses. If you throw a 900 000 mineral fleet at a 1 000 000 mineral fleet you will lose all of your ships while your enemy retains almost as much as you lost, putting you at a 900 000 mineral disadvantage. This is not recoverable during the war, and may not be salvageable during the ensuing 10 years, and your opponent on the other hand still has the capacity to defeat an 800 000 mineral fleet without spending anything, so can put the money you need to build up your fleet into their economy, or just double the size of his fleet. YOU WILL NEVER BE ABLE TO BEAT THEM AGAIN.unless the let you.

If two fleets are closely matched the bonus will help the side that starts losing not collapse with the first ship destroyed. If the weaker force manages to get lucky and pull a few kills out it does not actually help them win as they lose their bonus and will always stay behind the winning side, able to drag them down with them, but always losing in the end. And if a fleet is outmatched more than 2:1 they will still face utter destruction with minimal damage dealt in return.

It means that the winner can expect to suffer losses on par with the amount of stuff they destroyed, at best. Meaning you can grind down a fleet that doesn't take the time to fall back to their own spaceports to repair, while the defending fleet can easily jump back and forth into the fight after repairing, making wars harder to win by outright stomping into the enemies' land and destroying everything in your path until they lose because if they can't gather enough to beat you in one battle they can't hurt you.
 
I´m back and ready to answer.

Right, but that's not the "doomstack problem."

I would say that this is the effect (or maybe symptom?) of the doomstack problem.

We have to remember the changes to starbases as well. Starbases will now be on par with fleets. If an attacker takes enough losses in a fleet versus fleet engagement then he still might not be able to take the system because his ships will be too damaged to take on the starbase controlling said system. This gives the defender time to repair and rearm. Even if they lose that system then there's always another starbase waiting.

This is true, but remember the accepted definition of the term doomstack the forums have been using for the last few years: A force big enough that it can pretty much rampage through enemy territory unimpeded, with minimal (if any) losses.

Stronger starbases mean that, in the new system, a fleet will have to be bigger, in order to be considered a doomstack.

Which will encourage you to put even more ships in one place, at any time.


Depends on how small of a fleet... But if enough to actually capture the starbase, then it's a very viable tactic if you heavily outnumber the enemy. Attacking an undefended starbase is a hell of a lot better than attacking one guarded by a buffed fleet.

True, but remember that, in the new system, it will be required to have an Outpost in EVERY SINGLE SYSTEM: Considering maintenance costs (which will prevent you form upgrading every system), if you are facing only non-upgraded outposts, chances are that you are hitting only unimportant targets.

Also, if your fleet, while smaller than the enemy´s, is actually capable of defeating sucessive upgraded Starports (with low or zero losses) until you reach the enemy´s Core systems, then it is probably big enough to be used against the actual doomstack, if you position them around one of your own upgraded starbases, which would make that option... not really attractive.

And how have they solved Doomstacks? They havent, you cant really fully "solve" the doom stack problem. You can only mitigate it. Which is what exactly they have done. You will still generally lose against an overwhelming force. But now you can make this win very expensive for them and possibly reduce your own losses.

Have they mitigated it? 1- You still can put as many ships as you want in a system/battle; 2- Gutting Warp and Wormholes will remove strategic options for flanking and maneuvering, locking you into "artificial" choke points; 3- Starbases will be found in *every* *single* *enemy* *system*, and they can be a lot stronger than they were.

Fact 2- mean that you will often NOT be able to avoid enemy presence in a region you need to traverse, and Fact 3- mean that you WILL want overwhelming force when facing those static defenses, if only to reduce your own losses.

So you are STILL encouraged to put as many ships as possible in only one place., in other words, making doomstacks.

I'm not sure if someone else has pointed this out yet, but the real limit of what is possible to command in any kind of space combat has absolutely nothing to do with the capabilities of the admiral in question but the capabilities of the computers available. This is due to the distances involved. Ancient admirals did not need to deal with effectively infinite battle space and the difficulties of possible enemy positions spiraling into being completely unpredictable as time goes on. At space distances, even minor maneuvering of any of the ships involved at a large distance exponentially increases the difficulty of acquiring a firing solution, and that's without taking into affect any possible countermeasures that might exist (Such as physical CM like generating an expanding radiation cloud due to high energy particles firing through stellar gas, or electronic counter measures like bombarding incoming missiles/torpedos with fake positions of involved ships to confuse targeting sensors). The amount of factors that would need to be calculated simply to fire one weapon from one allied ship against a single enemy ship is astronomical, and even if we had instant communication of some kind eventually you would reach a point where every additional ship you add that has to be coordinated in the fleet wouldn't increase the capability of the fleet by the expected amount due to increasing the complexity of the calculations. And that's assuming that you have instant communications. If you had to deal with orders being given at the speed of light then in a large fleet you would likely be looking at minute or more delays to every order reaching the ends of the fleet further complicating any kind of calculations of enemy positions as you have to predict where they will be at that particular moment. You can just handwave it and say that the computers are perfect and can do it no problem but that's equally as arbitrary as saying that they can't and that there's a limit to what can feasibly be coordinated in each individual fleet.

This line of reasoning was born of a poster claiming that it is impossible for a single admiral to command too many ships in a battle. The examples i gave proved him wrong.

Also, the capacity of command for a battle leader (an Admiral , in this case) isn´t solely due to Computers or means of communication, but also (some might say: chiefly) the Military Organization: For each decision-making officer, there are a greater number of sub-officers who will in turn relay, implement and, if necessary, adapt those orders.

An Admiral isn´t controlling the actions of each individual crewmenber if his fleet; Instead he is sending order to the Battle group commanders, who, in turn, relay orders to the ship captains, who then relay orders to their Second officers, and so on. Thats how a single officer commands a force numbered in the thousands.

As for how the firing speed increase doesn't make sense from a realistic or immersion perspective, I'm thinking about it like this. Any form of space targeting is all about probabilities. There is no 100% exact targeting unless your weapon hits instantly regardless of distance, your sensors give instant information of enemy positions, that information is processed and coordinated with allied ships instantly, and is sent to your weapons instantly. In a situation where the enemy has, lets say, double the ships in the same arbitrary amount of space that you do, it could feasibly be faster for each of your weapons to target due to them having the possibility of calculating a solution that has high probabilities of hitting one out of several enemy ships. And if you say that the fire rate of a weapon is based significantly on the time it takes to acquire a firing solution, then it makes sense of a kind. Does it make perfect sense? No, but it's good enough for me.

I acknowledge you reasoning, but I could argue that, since this is SPACE, the enemy ships will (obviously) be hundreds of thousands (maybe millions) of kilometers apart form each others, so no trageting bonus should apply.

In fact, I could argue that, when faced with double the ships you have, The outnumbered fleet should be forced to disengage earlier.

Last but not least, You do realize that, when you say "its good enough for me", you are authorizing me to answer "No, it is not good enough for me", do you? And them the discussion would not go anywhere.

As for how this update deals with doomstacks, I think it remains to be seen.

I consider this a dangerous line of thougth.

We have information. We are capable of using said information to predict how the system will work. We can make educated guesses and semi-precise predictions about how the system will work.

I say that we should keep making those predictions and, hopefully, get changes/ajustments while the update is still in development, instead of "waiting and seeing".

It might be too late then.

Hope you're exams go well.

I´m OK. Just a slight deficiency in Vitamins b and D.

I disagree that this just adds more clicks to doomstacking. I think the scenario in most people's heads are two navies clashing over two opposing starforts. Since the likelihood of bordering another person with just a single chokepoint is laughable, let's assume that our empires have multiple contact points, entrances and star forts. Sure you can shove all your ships into one area, but you'd just make it easy for another player to beat you unless you had a significant advantage in size.

If one side does NOT has a significant advantage in numbers, then we are NOT discussing Doomstacks, are we?

For example, if we both had 25k FP and we each had 3 star forts relatively along our border (A-C) and (D-F), you could take all of your navy to attack my fort D. Say I saw you massing for the war-dec and moved 15k FP up to fort D. Let's say all our forts are 10k FP. You come in expecting to stomp my navy and my defenses, cause 1.8 is fresh on your mind where defense stations are little more than beefy roadblocks. However, I don't keep my fleet anywhere near the front of the starfort. In fact, I place them behind the fort on the outskirts of the system. You engage the fort, while I meander around the outside waiting for your cruisers and corvettes to over commit to the fort. I then come in at an angle to your larger fleet and being pounding your back line. Your vettes and bruisers now have to peel off of the fort which continues blazing away at your ships while they attempt to shield your back line.

If YOU can see my fleet coming, then *I* probably can see yours, too. Remember: the whole point of doomstacking is to get an overwhelming advantage: In this case, I would simply NOT engage, if I knew you would get reinforcements in time.


This whole time, my cruisers and corvettes get a bit of free time chunking your battleships and destroyers down before the combat system forces them to intercept your cruisers. I'm not sure if the smaller fleet bonus is added since my fort equalizes the total fleet power between us, but if it does, this battle goes even worse for you. I don't need to win. I just need to hurt you badly and stall. Since my ships won't be insta-wiped now and I can pull back with limping ships, I can confidently engage in this battle knowing that I have my other 10k FP attacking your fort C which is 15 systems away across from fort F. In that even fight, let's say I pull ahead and begin winning. It hurts me, but you can see that I will capture your fort C.

Didn´t you say that the Forts are 10k strong? if you send a 10k fleet to attack it, chances are that you will lose most of it even if you win. In this particular case I wouldn´t exactly mind, as the Starport is indestructible and would be retaken later easily, because after that 10kvs10k battle, you would almost certaily have to retreat that fleet.

Now you have to make a choice, do you pull away from our battle over D, or do you press on with a severely softened battleship group and bet all the marbles? We've both taken similar losses over D, but the power I lost to my fleet was mainly front line, while I took out a few of your battleships costing you more in resources to replace. You decide this isn't going to plan and back out. Now I have your fort and you've gotten nothing for your doom stack. Keep in mind that this ignores the bonus from the defensive ascension which would have given my ships more fire power and increases my build speed for replacement ships. This also ignores any auras coming off of the fort that would have weakened your ships. With those factored in, your mindless doomstack rush looks like a bad idea. We could push the scenario further with me bringing my full power to crush your defenses over A while you try to reclaim C, then putting my 15k FP over E while waiting for you to doomstack your way over to retake A while my other 10k FP takes C away again, and then attacking B while you're committed to taking A, but by now you should see that these changes seriously wreck all plausibility of doomstacking. This is without even mentioning about how you now have to have at least 2 or 3 admirals of varying strength in control of your doomstack.

Considering my answers just above, I think this specific scenario changed enough to warrant different questions.

Taken in conjunction with all that's been revealed so far, warfare in this game is getting a vast systematic improvement. I offer anyone the chance to play me in a match after the patch hits. I have found ways to use tactics/strategy in the doomstack era other than "let the ai doomstack pound a planet while I snipe it's armies". All I see in these DD's are opportunities to have more strategic wars.

Sorry, but no. The mere fact that 2/3 of the movement options are being gutted shows that things are getting worse. The fact that EVERYWHERE will be defended (in different degrees) is a thinly-veiled way to include attrition where none should exist. This is NOT a facilitator for strategic maneuvering or thinking.

I will admit though that fighting other players who outstrip you in FP is virtually impossible to beat currently atm and in 2.0, will by no means be a cake walk, but in 2.0, I can picture scenarios where a decent strategy will prevent total annihilation, if fleet power isn't vastly different, forcing status quo or even white peace (PLEASE GIVE US WHITE PEACE WIZ). But if it's like 140k FP vs. 60k FP...then you toast in a 1v1 bruh, no matter what you do, 2.0 or not. Rage quit, ask to be a vassal, delete the file... I don't know.

I don´t picture any scenarios in the new system, where a 140k FP vs 60K FP won´t stomp the other even harder than now. Specially when the smaller one will almost certainly take more losses to static defenses than the bigger one. Sure, the bigger one will lose more ships than now, but the smaller one will be even less capable of actually damaging the enemy core systems.

Currently, the smaller fleet can evade doomstacks and defenses and hit SpacePorts/mining stations to cripple/damage the enemy´s production capacity. This will almost certailny not be possible in the update.

i see this showing up a lot, and this was mostly to deal with needing only 1 admiral for your nation, a side effect of doomstack, not to outright get rid of them. now you don't outright lose admiral bonuses by splitting your fleet, and so might feel less reluctant to split on a moments notice.

I agree on the 1 admiral per empire. In my games I tend to have 3-4 fleets, because my emprie gets big in the mid-late game, and, since I Purge, my wars tend to involve a LOT of territory.

However, you realize that, in the new system, the onyl reason why you WON´T lose admiral bonuses when spliting fleets is because you won´t have them in the first place? Because the admiral will only command one fleet and the rest will be following him?

Nowadays we have unmanned drones, self-driving cars... why is there a need for a fleet crew? In the future you could probably control all forms of transportation from a main computer very much like a videogame. So, an admiral could control large fleets from a main computer, no problems. It would of course kill much of the roleplaying fun of the game...

I do think those things work better as an abstraction, in the game. (unless, of course, managing those systems is the point of the game, like FTL)

So you can sned them one hyperlane further, unleass you are fighting a 1-3 system empire connected to hyperlanses that are also connected to you, in which case there is no need for a recon force, is there ?

well, I think the practical effect is the same: if the enemy sends a smaller force and it is destroyed by a upgraded Starbase, your territory isn´t invaded. If the enemy scouts the area and finds said Starbase, he won´t send his smaller force to die a pointless death, so your territory remains uninvaded.

Don´t you agree?

P.S: What a wall of text, heh?

I THINK I have now said Everything I had to say about this issue. I hope I have been clear.
 
There's also something to be said for keeping the total pool of leaders relatively low. Scientists are already nameless bundles of stats to me, with no personality, throwing in more leadrs doesn't seem to be something that would make the ones we already have seem more special.
For more personality you would need a more CK2 like system with more traits and events (including chains), both triggered and cyclic, which take the traits into account.
 
Imagine kevlar vests. They stop fast travelling bullets but wont hold off a simple kitchen knife.
The point of impact for a beam weapon would be much smaller than that of a missile. Especially if this was the way to penetrate shields - they would be engineered to strike less surface area for that purpose.

We can come up with any silly reason we want to justify this change to missiles. I'm saying it isn't very believable and me pretending otherwise is just going to be me trying to turn a blind eye to a big inconsistency so I can enjoy the rest of the game - which I've done before I'd just rather not have to.
 
Nowadays we have unmanned drones, self-driving cars... why is there a need for a fleet crew? In the future you could probably control all forms of transportation from a main computer very much like a videogame. So, an admiral could control large fleets from a main computer, no problems. It would of course kill much of the roleplaying fun of the game...
well take The Legend of the Galactic Heroes for example in LOGH electronic warfare has gotten to the point were it's actually worse to use automation instead of humans
 
I feel like the force disparity modifier is a bad idea. You SHOULD be punished for getting into an engagement with a larger enemy, and the disproportionate results are just a fact of how combat like that works. I think it'd be better to just provide more opportunities for the smaller fleet to gain an advantage rather than just making them shoot faster somehow. Let the smaller fleet lay an ambush and catch the bigger fleet as they drop from FTL and engage them while they're stuck cooling down and unable to fight back, add in a kind of ship that lays down subspace netting to slow the enemy fleet and let your backline take advantage of superior weapon range, let the player set a broad tactical plan for specific fleets(something like the FF12 Gambit system, where you go "okay if you run into a fleet that is more than 125% of your fleet power then try to keep your distance and disengage as many ships as possible, if you come under attack from torpedoes then focus the ships that launch them"). It just seems like a lazy approach to the problem to go "if you have less ships they shoot faster" rather than introducing the sorts of combat complexities that render Lanchester's Law inapplicable to real battles.
 
not at fantasy choke points mandated solely by the RNG

I don't like this notion.

everyone has to deal with choke points militarily due to "rng". no one planned where they'd be born, and so didn't choose who or where they needed to protect. what chokepoints they had to deal with was completely out of their hands.
 
Looking forward to these changes to combat, borders, and movement. They were obviously difficult decisions to make, and not going to please everybody, but I think things really needed a rejig to make war in Stellaris feel more rewarding.

I loved the idea of playing with wormholes, and very occasionally using a different form of FTL could lead to tactical maneuvers against vastly superior opponents (warping in and taking out wormhole generators and trapping enemy fleets), but more and more play with warp-only just to add a feeling of geography and strategic movement to the game, otherwise it is more or less b-line to the enemy.
 
People act like the combat bonus will change the victor in a fight.

The expected result from a 15 vs 14 battle is 11-13 survivors on the team of fifteen, because as members of the losing team die they take more concentrated fire and can't kill the larger team's forces as easily, and one of them has free range to do whatever they want unmatched. leading to them taking more casualties, leading to more focused fire and less damage output, lading to a fast spiral that at some point leads to complete collapse. This is exacerbated in stellaris because of regenerating defenses and targeting. Meaning the fights are often total stomps with negligible damage to the winner

That is the problem. You expect a fleet that is 90% as strong as another to be able to cripple the stronger side at least, and take down most of their forces, but in the current situation it works out that a fleet that's 10% weaker gets completely annihilated inflicting less than 10% losses. If you throw a 900 000 mineral fleet at a 1 000 000 mineral fleet you will lose all of your ships while your enemy retains almost as much as you lost, putting you at a 900 000 mineral disadvantage. This is not recoverable during the war, and may not be salvageable during the ensuing 10 years, and your opponent on the other hand still has the capacity to defeat an 800 000 mineral fleet without spending anything, so can put the money you need to build up your fleet into their economy, or just double the size of his fleet. YOU WILL NEVER BE ABLE TO BEAT THEM AGAIN.unless the let you.

If two fleets are closely matched the bonus will help the side that starts losing not collapse with the first ship destroyed. If the weaker force manages to get lucky and pull a few kills out it does not actually help them win as they lose their bonus and will always stay behind the winning side, able to drag them down with them, but always losing in the end. And if a fleet is outmatched more than 2:1 they will still face utter destruction with minimal damage dealt in return.

It means that the winner can expect to suffer losses on par with the amount of stuff they destroyed, at best. Meaning you can grind down a fleet that doesn't take the time to fall back to their own spaceports to repair, while the defending fleet can easily jump back and forth into the fight after repairing, making wars harder to win by outright stomping into the enemies' land and destroying everything in your path until they lose because if they can't gather enough to beat you in one battle they can't hurt you.

I think your analysis is correct, and i agree that the bonus would help correct this issue.

But the bigger question is this: why is it that a fleet that is only 10% smaller than it's opponent gets eliminated while only dealing 20% or so damage to their opposing fleet?

I think fundamentally there is something at work with how damage is being registered and allocated over time that causes this snowballing effect. The force disparity mechanism is a band-aid. It might work just fine as a sort of catch-up mechanic, but I still wonder whether it could've been handled at a more fundamental level with how combat is resolved.

At an extreme end (and in a simplified example), say Side A has 100 ships and Side B has 90. The extreme result is that side A loses 90 ships (10 left) and side B loses their whole fleet. Right now, it feels like Side A loses about 25 ships and Side B loses their whole fleet. What's the root cause for battles playing out this way?
 
I don't like this notion.

everyone has to deal with choke points militarily due to "rng". no one planned where they'd be born, and so didn't choose who or where they needed to protect. what chokepoints they had to deal with was completely out of their hands.

In better games where the RNG doesn't generate a road-map to mandate exactly where you must focus your forces and defenses, the decision about where to place and move your asset's is commonly referred to as "strategy".
 
People thinking that announced changes dont will fix doomstack should read the star base and the warscore/war exhaustion dev diaries and after this, read the doomstack dev diary again.
 
That the smaller fleet fires faster sounds strange at first but it can be argued that it is a very abstract representation if the larger fleet not being able to bring all guns to bare because if the combat width for large parts of the battle.

Actually that abstract representation would be great. Having a reduction in fire speed as the fleet grows (regardless of the opposition - but they get the same malus based on their size) would probably do the same thing, yet make a more direct sense.

And it's not as much combat width as just "don't want to shoot friendlies, so have to be more conservative when taking shots". The effect could also differ between combat styles (close attacks means less reduction in fire speed for those swarming the opposition at "melee range" - but more deduction for those hanging back - and the swarmers are easier to hit for the opponents since they are definitiely within range)
 
RE: Force Disparity. It is a bit 'gamey' but Stellaris is still at best a strategic simulation. Some things are going to be 'gamey'. It's just a rough math trick to represent not just the mentioned 'ease of maneuvering and target rich environment' but also other intangibles like faster response due to shorter chain of command, lower institutional inertia, etc. A lot of the 'technologies' are abstract representations of those same concepts as well, this is just one more layer of them. No worse or better in and of itself than any other, and only really relevant in terms of how well it does or does not end up making the simulation entertaining.

ETA: Whoa, posted after only reading to page 3 and the post before mine also uses 'abstract representation' as a description. AbstractMind!
 
Some people seem to think this means a small fleet will be able to inflict more losses than it takes when fighting a large fleet. That's not the case from what I can tell - it still won't even inflict equal losses, but it'll be less unequal than it is now.
 
Actually that abstract representation would be great. Having a reduction in fire speed as the fleet grows (regardless of the opposition - but they get the same malus based on their size) would probably do the same thing, yet make a more direct sense.

And it's not as much combat width as just "don't want to shoot friendlies, so have to be more conservative when taking shots". The effect could also differ between combat styles (close attacks means less reduction in fire speed for those swarming the opposition at "melee range" - but more deduction for those hanging back - and the swarmers are easier to hit for the opponents since they are definitiely within range)

Wiz already answered this though, they're buffing the small fleet rather than nerfing the big fleet because positive buff percentages scale better than negative malus percentages.
 
At an extreme end (and in a simplified example), say Side A has 100 ships and Side B has 90. The extreme result is that side A loses 90 ships (10 left) and side B loses their whole fleet. Right now, it feels like Side A loses about 25 ships and Side B loses their whole fleet. What's the root cause for battles playing out this way?

It has to do with focused fire and the effect of removing combatants. Using your 90-100 ship example, assume all ships are equal with 10 health, and do 1 dmg a round. After 1 round the 100 ship side would remove 10 ships, and the 90 would remove 9. So 2nd round is 91 vs 80, and 9.1 removed vs 8. This keeps going with the smaller sides capability to remove ships decreasing at a faster rate. You can math out the rest if you want, but this is why the disproportionate losses happen, and why focusing fire to remove targets quicly is so effective.