Seafaring Sinking Chances = 25%

[Catt]

I set up a quick test in the editor because I was curious what the reduced sinking chances were (and I noticed others were as well). Appears to 25% in both sea and ocean based on ~1000 independent trials in each.

Catt

[Arrian]

It's a 25% reduction, or the chances of sinking are 25%?

What's the normal chances? 50%?

Oh, and when you do these 1000 trials... what do you do, set up a stack of 1000 galleys and goto them into ocean, and then hit end turn?

-Arrian

[Catt]

Chances of sinking are 25%. Normal chances are 50%.

I set up a scenario with ~100 galleys in coast around 6 little islands and preserve random seed off. First turn move all galleys into sea. Save game. End turn. Get another beer while the sinking animations run. Check F3 to see how many galleys survived. Note results. Reload save. Repeat. Do the same for ocean.

It would have been a better experience with ~1000 galleys since I would not have been dead drunk at the conclusion of the test and could have posted the results last night .

Catt

[alexman]

Catt, you are the ultimate Civ3 manual!
Thank you for the tests!

Arrian, if Catt means what he said, it's a 50% reduction, which is the same as 25% chance of sinking. [Edit: yes, I guess Catt meant what he said, and his hangover is not too bad because he's quick with the reply button ]

[alexman]

By the way, I don't think you are right, because I don't think that the seafaring trait should get a reduced chance of sinking. It's unbalancing!

[/CyberChrist@CFC]

[Dominae]

I always thought the chances of sinking in Ocean were higher than in Sea...

/me starts looking for a similar test as an excuse to get dead drunk.


Dominae

[Catt]

Originally posted by alexman
By the way, I don't think you are right, because I don't think that the seafaring trait should get a reduced chance of sinking. It's unbalancing!

[/CyberChrist@CFC]

Yes, that was a bizarre discussion. [edit](my surrender smile is not working)[end edit]

Originally posted by Dominae
* Dominae starts looking for a similar test as an excuse to get dead drunk.

You could set up a map with lots and lots of volcanoes and increased eruption chances to see if the "2 tile" lava danger is volcano + 2 tiles or volcano +1 tile a la the outpost 2-tile visibility. I have a sneaking suspicion that it is volcano + 1 tile, menaing it is perfectly safe to build a city with a volcano in the radius but dangerous to build right next to a volcano.

Take a drink with each eruption while you conduct the test -- you can even select your own rate of inebriation by playing with the eruption chances in the editor, whereas my rate was controlled exclusively by the length of the sinking animation and the hard-coded sinking chances.

Catt

[Nor Me]

On what difficulty level? I've seen circumstantial evidence for different sinking chances on different levels in PTW. (Something like less than 1/3 on chieftain and more than 2/3 on deity.)

[Catt]

Originally posted by Nor Me
On what difficulty level? I've seen circumstantial evidence for different sinking chances on different levels in PTW. (Something like less than 1/3 on chieftain and more than 2/3 on deity.)

I tested on regent. Good point -- I didn't do any tests with respect to other potential influencers I've seen speculated upon such as difficulty, government, culture lead, etc. Since I have the scenario I can check difficulty pretty easily, though I am not interested enough to try and rule out the inumerable other potential factors. I'll do Chieftan and Sid to see if there is a difference . . . how much beer do I have left in the fridge . . .

Catt

[Nor Me]

You know that J does work for one square at a time?

A quick test on 100 galleys in PTW gives 48 surviving on chieftain in sea, 56 in chieftain on ocean, 52 deity and sea and 56 pm deity in ocean.

It looks like 50% is right.

[Catt]

Beat me to it (but I moved each of 96 galleys into different squares, worried that the sinking chances might be affected by being in the same tile )

My results tie as well with ~500 trials in sea for each of Chieftan and Sid.The differences in sinking rates were statistically insignificant.

Catt

[DrSpike]

Well done Catt.

I asked this here at at CFC and nobody knew. As suspected it is incredibly unbalanced. Even if you need 3 turns at sea/ocean you only have a 58% chance of sinking. Due to the movement difference and the sinking probability difference the comparable number for non-seafaring is just over 3%. Granted it wouldn't always be 12 squares from coastal square to coastal square, but the results are striking.

Given the power of early contacts with another continent seafaring becomes too strong.

[Thoth]

Catt: If you've still got some beer left , does the experience status of the ship (ie regular, veteran or elite) make any differene to the sinking chance?

Dominae: If you're still looking for an excuse to get drunk , do any techs affect the odds of sinking? Astronomy negates the chances of sinking in sea, but does it affect the odds in ocean? Does the movement cost of the terrain have any affect? (My ingame experiences seem to indicate that it does, but my sample size was far too small to have any statistical signifcance)

[DrSpike]

Solutions:

1) The sinking probability could be changed, either back to 50% or much closer to it.

2) The probability of sinking could be made to depend strongly on turns at sea/ocean. In the simplest case make the probability of sinking the product of turns at sea/ocean and the usual probability. The first turn it is still 25%, then 50%, then 75%. As such the probability of surviving 3 turns becomes 0, and the probability of sinking at or before the second turn is around 63%. That is just an illustration........more sophisticated methods following the same pattern would be possible.

Tbh I doubt many would cry too hard if solution 1 (the easy way out) was adopted. Seafaring civs still have a tidy advantage though, being 4 times more likely to survive the 3 turn suicide run. As such maybe something along the lines of (2) would be better.

[Catt]

Originally posted by DrSpike
[. . .]

As suspected it is incredibly unbalanced.

[. . .]

Given the power of early contacts with another continent seafaring becomes too strong.

It's even more unbalanced than that in the sinple player game since the AI is coded to never risk a sinking by ending its turn in a tile where sinking is possible. "Suicide galleys" are the exclusive province of the human player in SP.

(note: the above was true in PTW and I've seen no evidence that C3C changed this).

Catt

[Catt]

Originally posted by Thoth
Catt: If you've still got some beer left , does the experience status of the ship (ie regular, veteran or elite) make any differene to the sinking chance?

Ugh! That would require a lot of time in the editor to change each unit to a different status (they default to regular). I am going to rest on my laurels and say that the unit experience plays no role whatsoever, even though I havent actually tested it. Don't tell anyone, though.

Catt

[Thoth]

I may give it a go myself next week. I've got a string of 12 hr midnights coming up so I won't have time over the weekend. I'd do it now but I want to get some more civing in.

[Arrian]

It's not unbalanced for MP, I don't think. I mean, between humans who are all capable of trying suicide galley runs, +1 movement & 50% reduce sinking chance is definitely nice, but is it really out of line with other trait bonuses? I don't think it is.

The real problem is in SP, because the AI will never try a suicide galley run. So all the AI gets out of the trait is the +1 movement, which is still nice, and the 1/2 cost harbors.

-Arrian