The probabilities of artillery fire (vanilla)

[choxorn]

I have always been curious as to these probabilities, of how often catapults, cannons, arty, and radar arty do the various bombard effects. So, I put
together a scenario- 50 catpults vs. an Ancient Age city (small, few buildings, defended by spearmen), 50 cannons vs. a Middle Age city (middle-sized, fair amount of buildings, defended by musketmen), 50 arty vs. an Industrial age city (large, many buildings defended by infantry), and 50 radar arty vs. a Modern Times city (huge, nearly every building, defended by mechs).

Here were my end results. The probabilities are probably not exact (that would take more testing, which I am too lazy to do), but fairly close. Also note that I only show the outcomes that happened, not any that didn't happen. Well, anyway, the results:

Catapult shots:
-nothing: 33 (66%)
-destroy building: 6 (11%)
-1hp from defender: 11 (22%)

Cannon shots:
-nothing: 26 (52%)
-destroy building: 2 (4%)
-1hp from defender: 19 (38%)
-kill citizen: 3 (6%)

Arty shots:
-nothing: 23 (46%)
-destroy building: 6 (12%)
-1hp from defender: 11 (22%)
-2hp from defender: 4 (8%)
-kill citizen: 6 (12%)

Radar Arty shots:
-nothing: 15 (30%)
-destroy building: 7 (14%)
-1hp from defender: 16 (32%)
-2hp from defender: 4 (8%)
-kill citizen: 8 (16%)

As you can see from this, catapults will do nothing 2 in 3 times, and Cannons will do nothing 1 in 2 times, so they have a tendency for being useless, and doing nothing is the most probable outcome for arty, still. And for radar arty, nothing is close to the most probable. But radar arty are obviously the best, then non-radar lol arty, then cannons, then catapults. This was what was expected, of course, but you can never be too sure! Like I said earlier, to find the exact probabilities would take some more testing, which I don't have the time- or the attention span- to do. If anyone wants to, feel free to verify my tests.

 

[gmaharriet]

This makes an interesting test with some merit, but there are several things which would skew the results...things like:

are there enough defenders for (potentially) 50 hit points to be knocked off?
are the units fortified?
are the towns/cities on a hill? across a river?

Since I play only C3C, a test would probably use fewer arty-type units (maybe 25?), but would include trebs against pikes. I'd probably have random seed turned OFF and run each test many times to allow for RNG differences.

A large difference in C3C is that arty always goes for units first with one big exception...if the civ being attacked has The Great Wall, buildings get hit first and even Great Wonders can be destroyed.

At any rate I'm too lazy to do the work setting up a test, so I'll just have to use my own judgment based on in-game experience.

 

[TimBentley]

Like I said earlier, to find the exact probabilities would take some more testing, which I don't have the time- or the attention span- to do. If anyone wants to, feel free to verify my tests.

They've already been tested, but I don't have the time to search right now.

 

[choxorn]

@gmaharriet:An answer to your questions:

1. No, but not every unit redlined. However, I might have only put 6 buildings in the catapult bombarded city...
2. No, they are not fortified because I don't know how to fortify them in the editor.
3. They were put in a randomly generated map, as I didn't know there was a difference, but at the beginning, the catapult city was size 5, the cannon city was size 10, the arty city was size 17, and the radar arty city was size 28.

As for the random seed, like I said, too lazy.
@TimBentley: link would be nice- I thought I'd find it in the strategy articles forum, but found nothing about arty probabilities.

 

[TimBentley]

http://forums.civfanatics.com/showthread.php?t=39577

I think that's sufficient for vanilla/PTW (although it looks like it did no sea bombardment testing; if there were differences in land or air bombardment it didn't find them); the probabilities of hitting unit/citizen/improvement obviously changed in C3C. I seem to remember posting in a thread discussing that.

I can't find that thread, but apparently I did some testing. Land bombardment will destroy walls in towns (or do buggy stuff if the walls come from Great Wall), then units, then population or improvements. Sea bombardment will destroy coastal fortress, ships, air units, land units, then population or improvements. Air bombardment will destroy units (air units, then ships, then land units), population, or improvements.

 

[choxorn]

Okay, so it has been done, but this one is a strategy article. Plus, I don't see it testing catapults/cannons/arty/radar arty. Still, I might read that one. And I also might make a test in C3C and put in Trebuchet. But I'm too lazy to test sea and air bombardment. (Note: Even if I do test all bombardment, which I won't unless I get really bored, I'm not testing H'wachas, Dromons, F-15's, or any type of missle because they have lethal bombard, which would screw me up.)

 

[#1 Person]

My opinion is that 50 bombard units is such a small sample that it is statistically meaningless. If I did the experiment I would have used a larger sample, like 100 or 200. Of course I will not, as I don't care THAT much.

 

[choxorn]

I don't really care either, but wanted to get rough estimates. As you can see, these are probably several percents off. But I'm not going to the trouble of making more units and doing the test again. Anyone else who wants to, can, though.