Are there any tricks to naval combat? Why do my battleships suck?

[SquirrelCensus]

Playing as Elizabeth, World, huge, 18 civs. I have a battleship/destroyer stack in the Baltic. One of Catherine's destroyers sinks my fresh battleship, the next destroyer sinks my destroyer. Another battleship gets sunk when a destroyer attacks her and then a transport (WTH?) finishes her off. Meanwhile in the Mediterranean, my destroyer attacks Caesar's ironclad. I'm favored to win at 99%, but I still come out of the battle with about 17/30 hp.

All my boats have at minimum Combat I or Flanking I. Some also have Drill I or II. So why are my battleships getting sunk by destroyers? Why are my destroyers getting seriously hurt by ironclads? What am I doing wrong? What should I be doing.

 

[Seraiel]

You aren't doing anything wrong, the RNG in CIV is just badly programmed. When fighting 100 battles it's understandeable that one will lose 1 at 99% odds in average. The results simply vary too much sometimes, it's possible to conquer a whole map with 20 units and sometimes one loses 10 at 1 city. With a good RNG the results would simply be closer to the results that are actually shown.

 

[Lexicus]

Well destroyers with Combat IV can fight battleships at more-or-less equal terms, which may be what's happening here.

 

[elmurcis]

I have seen combat III Destroyers from AI side pretty often.. And if its not on coast tile for extra 10% defense bonus its just your ~44 strenght (with Drill little more) vs. 30+30%->39 strenght.. I mean, its almost equeal and its totally not 95+% battle... If its Baltic sea, any turn I would end on coast atleast for this tiny 10% extra safety. Sometimes it can change odds enough so AI won't risk to attack.

 

[Dhoomstriker]

As for a trick to naval combat, support them with planes. Use your planes to weaken enemy boats and then attack for better odds, or don't even attack with your boats and reduce the AI's chance of attacking you due to the AI's reduced combat odds from having wounded boats.

Flanking I won't help your boats on defence; it's an offense-only promotion.

 

[TimANC]

I play archipelago maps a lot. The AI always has the advantage in naval combat. Particularly on defense. Avoid attacking a unit with less that 70% odds of success. Let them come at you.

Build naval units in cities that can give 2 or 3 promotions out of the box. I have not had much success with drill or flanking. I use the combat level promotions. Once a unit is at combat 4 or 5 then I'll give it more exotic promotions.

Like Dhoom says don't make it a fair fight if you are attacking. Use planes and missiles to soften them up first. Sacrifice transports if you are desperate.

Best way to destroy an opponents navy is to trap it in its cities and take the cities.

 

[Lexicus]

Stack defenders (battleships) at least one or two need Combat, a few should get drill so you can finish off damaged ships without taking much damage.

As someone said, use planes to weaken enemy ships before they get to your ships. If you have the techs guided missiles can finish ships off after the planes have hit them, but guided missiles are pretty inefficient to build since they're only one use. When ships are in potential danger from enemy ships always move them in large stacks, never let them get caught individually, and always have at least 2 or 3 battleships with combat-line as stack defenders.

 

[rbj2001]

You aren't doing anything wrong, the RNG in CIV is just badly programmed. When fighting 100 battles it's understandeable that one will lose 1 at 99% odds in average. The results simply vary too much sometimes, it's possible to conquer a whole map with 20 units and sometimes one loses 10 at 1 city. With a good RNG the results would simply be closer to the results that are actually shown.

A while ago I did a statistical simulation on my whole database of thousands of battles, and found that there was no support for the hypothesis that the RNG wasn't truly random. It's just how randomness works. You win some, you lose some. Once you've lost one, it's no less likely that you'll lose the next one, or the one after that. Odds stay the same, so losses aren't evenly spread out.

As for tactics, be sure to use the coastal defence bonus. Let the enemy attack you while you're by the coast, and you'll have better odds.

 

[Windsor]

As for tactics, be sure to use the coastal defence bonus. Let the enemy attack you while you're by the coast, and you'll have better odds.

I disagree with this tactic once battleships enter the equation. Having the benefit of collateral damage from attacking is much more important than the 10% coastal defense.

And even before battleships, planes and airships can make attacking sufficiently safe. As for promotions, the combat-line is the way to go.

 

[elitetroops]

With a good RNG the results would simply be closer to the results that are actually shown.

No, the opposite actually. A RNG that evens out results in the short term would be a very bad biased RNG. Occasional big swings are to be expected when dealing with randomness. Like rbj2001 said, there's nothing that would show the RNG is bad. I've also run tests with thousands of attacks and found that in the long run the results are very close to the odds that are shown.

Regarding the Destroyer getting down to 17 from attacking the Ironclad, this is nothing special. If you play with a mod like BUFFY, which displays advanced combat odds, you can see why:

With >99% winning odds, it's still only ~21% chance you survive at full health. 17/30 is 58HP, so 14% chance you get exactly that. But if you count everything below that as well, you will see that there's >30% chance that you will come out of the battle with 58HP or less.

 

[Pangaea]

I've also run tests with thousands of attacks and found that in the long run the results are very close to the odds that are shown.

That's fine and dandy and all, but won't help if you attack with 5 units at 80% in 2000BC and lose all battles.

I've also read that the random function used in the game isn't the best, and has a tendency to give several unlikely wins or losses in a row. Again, if you get such "bad rolls" early in the game, it could effectively be game over. I totally agree that over 10,000 battles or whatever things will even itself out, but that is besides the point raised by Seraiel, and now me.

 

[SquirrelCensus]

Thanks for the comments, everyone. I've been upgrading my battleships and destroyers to Combat III and hugging the coast and I'm seeing better results.

Now that Louie has destroyed my North Sea platform, I'm down to hitting his ships with cruise missiles and hoping that I can shoot down a fighter every so often with my SAM infantry...

 

[Major Tom]

That's fine and dandy and all, but won't help if you attack with 5 units at 80% in 2000BC and lose all battles.

I've also read that the random function used in the game isn't the best, and has a tendency to give several unlikely wins or losses in a row. Again, if you get such "bad rolls" early in the game, it could effectively be game over. I totally agree that over 10,000 battles or whatever things will even itself out, but that is besides the point raised by Seraiel, and now me.

Playing with high risk (with a high reward) should indeed give a serious setback once in a while. With high risk I mean playing without a safety net or a plan B.

A more or less deterministic outcome would make the game rather boring.

 

[elitetroops]

That's fine and dandy and all, but won't help if you attack with 5 units at 80% in 2000BC and lose all battles.

I've also read that the random function used in the game isn't the best, and has a tendency to give several unlikely wins or losses in a row.

Well, let's not rely on "I've read somewhere" when we can test it.

Here's your typical classical era battlefield. 8 stacks of 125 C1 Vultures surrounding a city with 0% cultural defenses, defended by 1000 unpromoted swordsmen:

Spoiler :

Units where chosen to get close to your mentioned 80% odds, each Vulture will have 79.65% odds to win the battle:

Spoiler :

I do 8 stack attacks and find that after these battles 787 Vultures are still standing, so I won 78.7% of the battles. Close enough to the mentioned 79.65%:

Spoiler :

The game has new random seed on reload checked, so I repeat the process 2 more times, 2nd time I win 78.4% of the battles, 3rd time I win 80.0% of the battles. (Actually there was also in between a round where I won 80.9% of the battles, but forgot to turn on BUFFY logging, so I didn't get the data saved for the next part. That run is therefore excluded.)

That's 3000 battles at 79.65% odds. Now, with 20.35% of losing, the chance of losing 3 consecutive battles should be 0.843% or 1/119, the chance of losing 4 consecutive battles is 0.1715% or 1/583 and the chance of losing 5 consecutive battles is 0.0349% or 1/2865.

If I pick any random series of 4 consecutive battles, there should be 1/583 chance they are all losses, and so on. I have data on 3000 battles, which contains 2998 different series of 3 consecutive battles, 2997 series of 4 consecutive battles and 2996 series of 5 consecutive battles. So what we expect to find is:

3 losses in a row should occur 2998/119 = 25.2 times
4 losses in a row should occur 2997/583 = 5.1 times
5 losses in a row should occur 2996/2865 = 1 time

I open up the BUFFY logs, use find and replace to make them more easily readable and have a look at the results. The actual results:

5 consecutive losses happened 1 time
4 consecutive losses happened 5 times
3 consecutive losses happened 15 times

At this point we should note that any series of 5 losses also counts as 2 series of 4 losses and 3 series of 3 losses, and any separate series of 4 losses counts as 2 series of 3 losses. So the real results are:

5 consecutive losses: 1 time (expected 1)
4 consecutive losses: 7 times (expected 5)
3 consecutive losses: 28 times (expected 25)

Slightly above expectation, but most certainly not "a tendency to give several unlikely wins or losses in a row". The results varied a lot between the 3 different series, as expected. If I did another 3000 I would also get different numbers, possibly below expectation that time. The sample size is still too small to rule out the possibility that there could be some degree of bias in any direction or the other in the long run, but if there is one, it is so small that in practice it has no effect.

Now can we please lay this "the RNG is horrible"-talk to rest once and for all. Losing streaks happen, as they should. And if you want to play the most efficient possible game and build only the amount of units you expect to need, this will bite you in the behind every now and then. If you keep all your eggs in one basket and play lots of "win the battle or abandon game"-situations, then expect to abandon games every now and then.

Here is the (long) list of battles in the order they happened. 1 = win, any other number = strength of surviving defender. You will see some long streaks of nothing but victories and some areas where losses are a lot more frequent. But this is exactly what you should expect from a RNG. A list where 4 wins is always followed by 1 loss would have absolutely nothing to do with randomness.

Spoiler :

1
(0.48/6)
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
(1.86/6)
1
1
1
1
(1.86/6)
1
(3.24/6)
1
1
1
1
(4.62/6)
1
(3.24/6)
1
1
1
1
1
1
1
(3.24/6)
1
1
1
(0.48/6)
1
1
(3.24/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
(0.48/6)
1
1
1
1
1
1
1
(4.62/6)
1
1
1
1
1
(3.24/6)
(0.48/6)
1
1
(1.86/6)
1
1
(0.48/6)
1
1
1
(0.48/6)
1
(4.62/6)
1
1
1
(0.48/6)
(0.48/6)
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
1
1
1
1
1
(4.62/6)
1
1
1
(0.48/6)
1
(0.48/6)
1
1
(0.48/6)
(0.48/6)
1
1
1
1
1
(4.62/6)
1
1
1
1
1
1
1
(0.48/6)
1
1
1
(0.48/6)
1
(4.62/6)
1
1
1
1
(0.48/6)
1
1
1
(1.86/6)
1
1
1
1
1
(1.86/6)
1
1
(0.48/6)
1
1
(0.48/6)
(6.00/6)
1
1
(1.86/6)
(0.48/6)
1
1
1
(3.24/6)
1
1
1
(0.48/6)
1
1
1
1
(0.48/6)
1
1
(0.48/6)
1
(0.48/6)
1
1
1
1
1
1
1
(3.24/6)
1
1
1
1
1
1
1
1
(4.62/6)
1
1
1
1
1
(0.48/6)
1
(0.48/6)
1
1
1
1
(0.48/6)
1
1
1
1
1
1
(1.86/6)
1
1
(3.24/6)
1
1
(3.24/6)
(1.86/6)
1
1
1
1
1
1
(0.48/6)
1
(3.24/6)
1
1
1
1
1
1
1
(1.86/6)
1
(3.24/6)
1
1
1
1
1
1
1
(1.86/6)
(1.86/6)
(1.86/6)
(0.48/6)
1
1
1
1
1
1
(1.86/6)
1
1
1
1
(3.24/6)
(3.24/6)
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
1
(3.24/6)
1
1
1
(0.48/6)
1
(3.24/6)
1
1
1
(0.48/6)
1
1
(1.86/6)
1
(1.86/6)
1
1
1
1
1
(0.48/6)
(3.24/6)
1
(0.48/6)
1
1
1
(3.24/6)
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
(4.62/6)
(1.86/6)
1
1
1
(3.24/6)
1
1
1
(0.48/6)
1
1
1
1
1
1
(3.24/6)
(0.48/6)
(3.24/6)
1
1
1
1
1
(3.24/6)
1
1
1
1
1
1
1
(1.86/6)
1
1
(3.24/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(4.62/6)
1
(1.86/6)
1
1
1
1
(1.86/6)
(0.48/6)
1
1
1
1
1
1
(0.48/6)
1
1
(1.86/6)
1
(0.48/6)
(0.48/6)
(0.48/6)
(3.24/6)
(0.48/6)
1
1
1
1
(0.48/6)
(1.86/6)
1
1
1
1
(1.86/6)
(3.24/6)
1
(1.86/6)
1
1
1
(6.00/6)
(0.48/6)
1
1
1
1
1
(1.86/6)
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
1
(3.24/6)
1
1
1
1
1
1
1
(0.48/6)
(1.86/6)
1
1
1
(0.48/6)
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
1
1
1
(3.24/6)
(3.24/6)
1
(1.86/6)
1
(3.24/6)
1
1
(4.62/6)
(0.48/6)
1
1
1
(3.24/6)
1
1
1
(0.48/6)
1
1
(3.24/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
(3.24/6)
1
1
(0.48/6)
1
1
1
1
1
1
1
1
(4.62/6)
1
1
1
1
1
1
1
(0.48/6)
1
1
1
(0.48/6)
1
(1.86/6)
(3.24/6)
1
1
(1.86/6)
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
1
(0.48/6)
1
(1.86/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
(0.48/6)
(0.48/6)
1
(1.86/6)
1
1
1
1
1
1
1
1
1
(3.24/6)
1
1
(0.48/6)
(1.86/6)
1
1
1
1
1
1
1
1
1
1
(3.24/6)
1
1
(3.24/6)
1
1
1
1
(4.62/6)
(4.62/6)
1
1
1
(4.62/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
1
1
1
(3.24/6)
1
1
1
1
1
(1.86/6)
1
1
1
(4.62/6)
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
1
(0.48/6)
1
(0.48/6)
1
1
1
1
1
1
(3.24/6)
1
1
1
(6.00/6)
1
1
(6.00/6)
1
1
1
(0.48/6)
(0.48/6)
1
1
1
1
(1.86/6)
(1.86/6)
1
(3.24/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
(1.86/6)
1
1
(0.48/6)
1
(3.24/6)
(0.48/6)
1
1
1
1
1
1
1
1
1
1
(4.62/6)
1
(3.24/6)
(0.48/6)
1
1
1
1
1
1
1
1
(0.48/6)
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
1
(1.86/6)
(1.86/6)
(0.48/6)
(0.48/6)
1
1
1
1
(1.86/6)
1
(1.86/6)
1
1
(0.48/6)
(3.24/6)
1
1
(0.48/6)
1
1
1
(1.86/6)
1
(1.86/6)
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
1
1
1
1
1
1
1
1
(4.62/6)
1
(3.24/6)
1
1
1
(1.86/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
1
1
(1.86/6)
1
1
1
(4.62/6)
(1.86/6)
1
1
1
(1.86/6)
1
1
1
1
1
1
1
(0.48/6)
1
1
(0.48/6)
(3.24/6)
1
(3.24/6)
1
(0.48/6)
1
1
(0.48/6)
1
1
1
(0.48/6)
1
1
1
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
(1.86/6)
1
(1.86/6)
1
1
1
1
1
(1.86/6)
1
1
1
(0.48/6)
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
(4.62/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(3.24/6)
(3.24/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(3.24/6)
1
1
(3.24/6)
(0.48/6)
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
(3.24/6)
1
1
1
(1.86/6)
1
1
1
1
1
1
(0.48/6)
1
1
1
(3.24/6)
1
1
(1.86/6)
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
(6.00/6)
(3.24/6)
1
1
1
1
1
1
1
(3.24/6)
1
1
(3.24/6)
1
(1.86/6)
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
1
(1.86/6)
1
(3.24/6)
(6.00/6)
(4.62/6)
1
1
(1.86/6)
1
1
1
1
1
1
(4.62/6)
1
1
1
1
(3.24/6)
1
1
1
1
(1.86/6)
1
1
1
(0.48/6)
1
1
1
1
1
1
1
(1.86/6)
1
1
1
(0.48/6)
(0.48/6)
1
1
1
(0.48/6)
(6.00/6)
1
(1.86/6)
1
1
1
1
1
1
(3.24/6)
1
(4.62/6)
1
(0.48/6)
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
1
1
1
1
(3.24/6)
1
1
1
(3.24/6)
1
1
(6.00/6)
(4.62/6)
1
1
1
(0.48/6)
(0.48/6)
(1.86/6)
(0.48/6)
1
1
1
1
(3.24/6)
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
(3.24/6)
1
1
1
1
1
1
1
1
1
1
1
1
(3.24/6)
(0.48/6)
1
1
1
1
1
(1.86/6)
1
1
1
(3.24/6)
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
(1.86/6)
1
1
1
1
1
1
1
(4.62/6)
1
1
(3.24/6)
1
(0.48/6)
1
1
1
(0.48/6)
1
(3.24/6)
1
(1.86/6)
1
1
1
1
(4.62/6)
1
1
1
1
1
1
1
1
(1.86/6)
1
1
(0.48/6)
1
1
(4.62/6)
1
1
1
1
(0.48/6)
(0.48/6)
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
(1.86/6)
(0.48/6)
1
(1.86/6)
1
1
1
1
(0.48/6)
1
(0.48/6)
1
1
1
1
1
(0.48/6)
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
1
1
(4.62/6)
1
1
1
1
1
1
1
1
(1.86/6)
(1.86/6)
1
1
1
1
(0.48/6)
(1.86/6)
1
1
1
(3.24/6)
1
1
(0.48/6)
1
(1.86/6)
(3.24/6)
1
1
1
1
1
(3.24/6)
1
(0.48/6)
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
(0.48/6)
1
(1.86/6)
1
(3.24/6)
(0.48/6)
1
(1.86/6)
1
1
1
(0.48/6)
1
1
1
1
1
(1.86/6)
1
1
(0.48/6)
(3.24/6)
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
1
1
1
(1.86/6)
1
1
1
1
(1.86/6)
(1.86/6)
1
1
1
(0.48/6)
1
1
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
(3.24/6)
1
1
1
1
(1.86/6)
1
1
(1.86/6)
1
(3.24/6)
1
(0.48/6)
(1.86/6)
(3.24/6)
1
1
1
1
1
1
1
(4.62/6)
1
(1.86/6)
1
1
1
1
1
1
1
1
1
1
1
(3.24/6)
1
(0.48/6)
1
1
1
1
1
1
(1.86/6)
1
1
1
(0.48/6)
1
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
(3.24/6)
1
(4.62/6)
(3.24/6)
1
1
1
(0.48/6)
1
1
(0.48/6)
(3.24/6)
1
1
1
1
1
(3.24/6)
1
1
1
1
1
(3.24/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
(0.48/6)
1
1
1
1
1
(1.86/6)
(0.48/6)
(0.48/6)
1
1
1
1
1
1
1
(0.48/6)
1
1
(1.86/6)
1
1
1
1
1
1
1
1
(3.24/6)
(3.24/6)
1
1
(3.24/6)
1
1
1
1
(0.48/6)
1
1
(4.62/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
(0.48/6)
1
1
1
1
1
(0.48/6)
(1.86/6)
1
1
1
(3.24/6)
1
(3.24/6)
1
1
(1.86/6)
1
1
1
(3.24/6)
1
1
(0.48/6)
1
1
1
(1.86/6)
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
(0.48/6)
1
1
(3.24/6)
1
(1.86/6)
1
1
1
1
1
1
(4.62/6)
1
1
1
1
1
1
1
1
(3.24/6)
1
1
1
1
1
(1.86/6)
1
1
(3.24/6)
(0.48/6)
1
(1.86/6)
1
(0.48/6)
1
1
1
1
1
(4.62/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(3.24/6)
1
1
1
1
(4.62/6)
1
1
1
(1.86/6)
1
1
1
1
1
1
1
1
(1.86/6)
1
1
(0.48/6)
1
(3.24/6)
1
1
1
(3.24/6)
1
(3.24/6)
1
1
(3.24/6)
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
(0.48/6)
1
1
1
1
1
(1.86/6)
(1.86/6)
1
(6.00/6)
(0.48/6)
(3.24/6)
1
1
1
1
1
1
1
1
(3.24/6)
1
1
(0.48/6)
(0.48/6)
1
1
1
(4.62/6)
1
1
1
(1.86/6)
1
1
1
1
(3.24/6)
(0.48/6)
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
1
(0.48/6)
1
1
1
(0.48/6)
1
1
1
(0.48/6)
(0.48/6)
1
(0.48/6)
1
1
(3.24/6)
1
1
1
1
1
1
1
1
(4.62/6)
(0.48/6)
1
1
(4.62/6)
1
1
1
(1.86/6)
1
1
1
(0.48/6)
1
1
1
1
1
(1.86/6)
(1.86/6)
1
1
1
(3.24/6)
(0.48/6)
(1.86/6)
(1.86/6)
1
1
(1.86/6)
(0.48/6)
1
1
1
1
1
1
(6.00/6)
1
1
1
(3.24/6)
(6.00/6)
1
(1.86/6)
(4.62/6)
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
(1.86/6)
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
(0.48/6)
1
1
(3.24/6)
1
1
1
1
1
1
1
(3.24/6)
(1.86/6)
1
(1.86/6)
1
1
1
1
1
(6.00/6)
1
1
1
1
1
1
1
(1.86/6)
1
(3.24/6)
(0.48/6)
(0.48/6)
1
1
1
(0.48/6)
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
1
1
1
1
1
1
1
(3.24/6)
(3.24/6)
1
1
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
(4.62/6)
1
1
1
1
1
1
1
1
1
1
1
(4.62/6)
1
1
1
1
1
1
1
(0.48/6)
(6.00/6)
(0.48/6)
1
(1.86/6)
1
(0.48/6)
1
1
1
1
1
1
1
(0.48/6)
1
(0.48/6)
1
1
(0.48/6)
(0.48/6)
1
(0.48/6)
1
1
1
1
1
1
1
(0.48/6)
1
(3.24/6)
1
1
1
1
1
(6.00/6)
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(3.24/6)
1
(0.48/6)
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
(3.24/6)
(0.48/6)
1
1
1
1
1
(1.86/6)
(1.86/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
(4.62/6)
1
1
1
1
(0.48/6)
1
1
(0.48/6)
1
1
1
1
(1.86/6)
1
1
1
1
1
1
(4.62/6)
1
1
1
1
1
(6.00/6)
1
1
(0.48/6)
1
1
1
(0.48/6)
1
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(4.62/6)
1
1
1
1
1
1
1
1
(0.48/6)
(1.86/6)
1
1
1
(0.48/6)
(1.86/6)
1
1
1
(0.48/6)
1
1
1
1
1
(0.48/6)
1
1
1
1
1
1
(1.86/6)
1
1
1
(4.62/6)
1
1
1
1
1
1
1
(0.48/6)
1
1
1
1
1
1
1
1
(4.62/6)
1
1
1
1
1
(1.86/6)
1
(1.86/6)
1
1
1
1
1
1
(3.24/6)
1
(3.24/6)
(0.48/6)
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
1
1
(0.48/6)
(0.48/6)
(1.86/6)
1
(1.86/6)
1
1
1
1
(1.86/6)
1
(6.00/6)
1
1
1
1
1
1
1
1
1
(0.48/6)
(0.48/6)
1
1
1
1
1
(0.48/6)
1
(4.62/6)
1
(0.48/6)
1
1
1
(0.48/6)
1
1
1
1
1
(0.48/6)
1
1
(6.00/6)
1
(3.24/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(3.24/6)
(1.86/6)
(0.48/6)
1
(0.48/6)
1
1
1
1
1
(0.48/6)
1
1
1
1
1
(6.00/6)
1
(0.48/6)
1
1
(0.48/6)
1
1
1
(1.86/6)
1
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
(0.48/6)
1
1
(0.48/6)
(3.24/6)
1
1
(3.24/6)
1
1
(0.48/6)
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
(3.24/6)
(6.00/6)
1
1
1
1
1
1
1
1
1
1
(3.24/6)
1
1
1
1
1
1
1
1
1
1
1
1
(3.24/6)
1
1
1
1
1
1
(0.48/6)
1
(0.48/6)
1
1
1
1
1
(0.48/6)
1
1
1
1
(0.48/6)
1
(0.48/6)
(4.62/6)
(3.24/6)
(0.48/6)
1
1
(0.48/6)
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
1
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
(3.24/6)
1
1
1
1
1
1
1
(1.86/6)
1
1
(1.86/6)
(0.48/6)
1
1
(1.86/6)
1
1
1
1
1
(1.86/6)
1
(1.86/6)
1
1
1
1
(3.24/6)
1
1
1
1
(1.86/6)
1
(1.86/6)
1
1
1
1
1
(3.24/6)
1
1
1
1
(1.86/6)
(4.62/6)
1
(0.48/6)
1
1
(0.48/6)
(1.86/6)
1
1
(3.24/6)
1
(1.86/6)
1
1
1
1
1
1
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
(3.24/6)
1
1
1
1
1
1
1
(0.48/6)
1
1
1
1
1
1
(0.48/6)
1
1
1
(0.48/6)
1
1
1
(3.24/6)
1
1
(0.48/6)
(1.86/6)
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
1
1
(1.86/6)
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
(0.48/6)
(1.86/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(3.24/6)
1
1
1
1
1
1
1
1
1
(3.24/6)
1
1
1
1
1
(0.48/6)
1
1
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
(0.48/6)
1
1
1
1
(0.48/6)
1
1
1
1
(0.48/6)
(1.86/6)
1
1
1
1
1
1
1
(4.62/6)
1
1
1
(1.86/6)
1
1
(3.24/6)
1
1
1
1
1
(0.48/6)
1
1
1
1
1
1
1
(1.86/6)
(0.48/6)
1
1
(3.24/6)
1
1
1
(3.24/6)
1
1
1
1
1
1
1
1
(1.86/6)
1
(3.24/6)
1
1
1
1
(3.24/6)
1
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
1
1
(0.48/6)
1
1
1
1
(0.48/6)
1
(1.86/6)
(1.86/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
(4.62/6)
1
1
(3.24/6)
1
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
(0.48/6)
1
(0.48/6)
1
1
1
1
(4.62/6)
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
(3.24/6)
1
1
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
(3.24/6)
1
1
(0.48/6)
1
1
1
(0.48/6)
(0.48/6)
1
1
1
(1.86/6)
1
1
(1.86/6)
(1.86/6)
1
1
(1.86/6)
1
1
(1.86/6)
(1.86/6)
(0.48/6)
1
1
1
1
1
1
1
1
1
1
1
1
(1.86/6)
1
1
1
1
1
1
1
1
1
1
1
(0.48/6)
(0.48/6)
1
1
1
1
(3.24/6)
(1.86/6)
1
1
(0.48/6)
1
1
1
(1.86/6)
1
1
1
1
1
1
1
1
1
1
1
1
1
1

 

[Re-loader]

Thanks for the comments, everyone. I've been upgrading my battleships and destroyers to Combat III and hugging the coast and I'm seeing better results.

Now that Louie has destroyed my North Sea platform, I'm down to hitting his ships with cruise missiles and hoping that I can shoot down a fighter every so often with my SAM infantry...

Draw his ships out of his territory, either by blockade or just floating around near his tiles. The AI will try to draw you in to his territory.

Have a couple of submarines in your naval stack, and some missiles in them.

Have some carriers and air support in your naval stack.

Always keep the stack together, any stragglers will get picked off by the AI.

I prefer Drill over Combat. Drill is equally good for defending. I have a small amount with just Combat in a stack, and majority with Drill, land or naval.

OP's strategy is the problem, not his choice of odds.

 

[ConfusedCounsel]

Snip.

Excellent post. I've always been highly suspicious about 'faulthy RNG' talk in games, but it is nice to see it proven once again with such rigor.
While true randomness is indeed difficult to program, this is more of an academic or philosophical problem. Randomness sufficient for a computer game is a peace of cake. In fact, it is probably more difficult to get a RNG that is slightly off in some way then a normal well functioning one.

 

[6K Man]

Lots of good advice in this thread already (using planes/missiles to soften up AI ships, ending turns on coast, avoiding Flanking, etc). I’ll just add a few points:

Don’t expect naval combat in Civ4 to work the same way real-world naval combat does/did. In the real world, a sail frigate armed with cannon would be unable to damage anything more modern than an ironclad, and a battleship would rarely suffer significant damage in single combat with a destroyer. Obviously Civ4 doesn’t work like that. Rationalize that how you like (Destroyer in Civ4=squadron of destroyers, for example), but you’ll find the game easier once you accept how it works.

Ships don’t get the opportunity to collect terrain or situational defensive bonuses the way land units do, and there are no defense-specific promotions for ships. So, a Destroyer defending on Coast against another Destroyer will lose about 1/3 of the time, and suffer significant damage most of the time it wins, making it an easy kill for a second attacker (By contrast, an Axeman fortified in a hill city behind 20% culture will almost always win defending against another Axeman, and might even be able to beat a second or third attacker). So, keep your ships in stacks. Unless the AI comes at you with a bigger stack, you’ll have a chance to heal up your wounded ships between turns. I’ve had solo Destroyers die after being attacked by 4 Frigates, and Battleships die after being attacked by 2 Destroyers… so now, my ships move around in pairs, at least.

If you’re in the late game, stack Attack Subs with Battleships to avoid having the Battleships defend against AI subs.

Hang on to Airships for use in sub-spotting; other aircraft can’t do that.

Don’t upgrade all your Destroyers to Stealth Destroyers, because the latter won’t defend against ships that can’t see them, and they can’t see submarines.

TL;DR: Quantity will usually beat quality in Civ4 naval combat.

 

[Seraiel]

Neither elitetroops nor anybody else understands the point.

With the low number of units a player attacks in early times, the deviation of the RNG just leads to completely idiotic results. It's possible to play 2 days on a round and work hard for every advantage and then the RNG just says "I have a bad day, now you lose all of your troops, have fun and play another game, all your effort is worth nothing" and that is simply not fair. A player needs to be rewarded for good moves and the RNG is just where this fails most often. For me this is completely understandeable and it's a point raised by good players again and again. Yes, randomness enriches a game and brings fun but randomness that is too great is definitely not good and I don't understand what a statistical analysis should help. Deviation in randomness is what leads a player to risky play, it's not the players that play risky because they are greedy, it's the too great deviation in randomness that demands that kind of playstyle just because it's possible to win a lot and in a single case statistics are of no matter because there is only one or the other, it doesn't matter if in 100 games the results will be fair, what matters is that one single game.

Seriously, how is it possible that players get blamed for playing risky if they play the game like it was programmed and seriously, how elitetroops can a great deviation in RNG be called good if it's the source of risky play... . If a single game is all that matters than the deviation in the RNG must not be great.

 

[elitetroops]

No, Seraiel. You don't understand randomness.

Randomness means that every roll is independent of the outcome of the previous roll. When you lose an 80% battle, there is still 20% chance you lose the next 80% battle. And when you lose that one as well, there is still 20% chance you lose the third one. This will over time lead to great deviation. How often you should expect to see these deviations over a larger sample can be predicted by simple math, and my post above showed that the RNG in CIV follows these predictions very closely. Hence, it is a good RNG.

You are asking for a RNG that would not work like this, but be more likely to give a win after a few losses. You'd probably want it to be 100% chance of win after losing 2-3 80% battles. That is not random!! If you want to keep complaining, at least stop using the words "bad RNG", and instead say "CIV IV is badly programmed because it uses a RNG". That is essentially the point you are trying to get through.

If a single game is all that matters, then it is up to you how risky you want to play. Losing streaks will happen. They should and must happen when dealing with randomness. It's up to you how to deal with it when they inevitably do happen.

 

[Seraiel]

No, Seraiel. You don't understand randomness.

Randomness means that every roll is independent of the outcome of the previous roll. When you lose an 80% battle, there is still 20% chance you lose the next 80% battle. And when you lose that one as well, there is still 20% chance you lose the third one. This will over time lead to great deviation. How often you should expect to see these deviations over a larger sample can be predicted by simple math, and my post above showed that the RNG in CIV follows these predictions very closely. Hence, it is a good RNG.

You are asking for a RNG that would not work like this, but be more likely to give a win after a few losses. You'd probably want it to be 100% chance of win after losing 2-3 80% battles. That is not random!! If you want to keep complaining, at least stop using the words "bad RNG", and instead say "CIV IV is badly programmed because it uses a RNG". That is essentially the point you are trying to get through.

If a single game is all that matters, then it is up to you how risky you want to play. Losing streaks will happen. They should and must happen when dealing with randomness. It's up to you how to deal with it when they inevitably do happen.

elitetroops my friend...

I understand how you see that now but plz try to also understand what I write:

I'm not against an RNG because I think that randomness is something good as it brings fun, but the flip-side of fun is frustration and you can believe me in this moment, that I don't currently suffer from the last but have over years because of the RNG, that's something personal though, I seriously made a statistical analysis of my games and found out that I had so much bad luck with the RNG like I myself cannot understand, because I found out that in the game in which I had the most luck I actually only got the chances that were predicted .

An RNG that evens out results though is something good imo. and you can also trust me that I understand randomness, maybe not fully but what you explained I understood years ago already and that what you mentioned, complete true randomness that doesn't underlie any rules that prevent it from being too powerful is the problem imo. It causes great luck and great happiness but also great frustration, one can lose a whole game in one attack and that without having made a single fault or having made a bad decision and that goes too far for me. Randomness in games needs to be very carefully controlled so that the results don't get too extreme because (this should be easily understandeable) great luck is great but if there is no limit, then players will play more and more risky seeking greater and then even greater luck and the flip-side is that they also get greater and greater suffering and too much of something (and I know that very well from having very strong feelings i. e. ) is not good anymore, so very slightly controlled "a little less luck" and therefore "a little less suffering" would be better.

And yes, I didn't want to call the RNG bad but write that the RNG is badly programmed imo. and that CIV could have become a way better game if the DEVs i. e. had invested into implementing the above.

[EDIT]

What also got mentioned btw. was another thing that I now call "badly programmed" I'm confident in writing this and that is the function via which HP-losses are distributed. If a Machine Gun is in a city there are good chances that a Cavalry will beat it, this already is stupid, do you know what Machine Guns do with Cavalries? I hope you don't. And if in a 99% fight the chances to get out with 50% of the HP are 15%, that again is too high for my taste. A 99% fight means that one unit is lightyears further in developement and that unit should maybe take a few scratches and ofc. it can lose if i. e. having a malfunction or whatever but losing 50% HP and that with 15% chance? Not in the world is that realistic, a Destroyer could run over an Ironclad and probably take only minor damage, a Destroyer would need one shot to sink the Ironclad while the Ironclad probably could not even scratch the armor of the Destroyer and yes, ofc. , the Ironclad could wait somewhere, the Destroyer passes and then the men from the Ironclad jump on the Destroyer and capture it, but 50% HP at 15% chances is ridiculous.

Addition: The Combat rounds are also too few in number. You once demonstrated that there are cases where a single +10% STR promo can lead to needing 1 combat round less sometimes raising chances from 30% without that promotion to 60% with it. And I'll go again one step further and now also say that the whole unit-upgrading function is a complete fail aswell, because unit-upgrades are basically not adviseable in CIV and the reality (in states other than the WW2 UDSSR) the humans are the valuable part and not the weapons. Ok, training the humans costs some money, but how much difficulties do you think an ancient Archer would have had with a Rifle? Archery is way more difficult than handling a rifle, the Archer would have laughed at how easy his job became. Actually the whole process of upgrading is stupid imo. , because a Warrior cannot live 5000y so the Warrior 5000y after the initially build Warrior are the children of the children of the children (...) so they'd live in the time where those new weapons would get developed and could easily adapt.

And there was one other function that I didn't like aswell, I'll remember it in a further discussion maybe and I beg to understand the above correctly now.