The Great Wall Debate

[Tacit_Exit]

After receiving mixed messages from posters here at CFC on the nature of the Great Wall's effects on cities (size 7-12), I decided to use the editor to set up some tests.
Scenario: Human Player-Carthage
Player2(AI) -France
I placed 1000 Carthaginian warriors adjacent to the French start position, and 1000 French warriors on the French start position.
All terrains grassland, no rivers, France given a Town or City in start position depending on test.From this basic template I created 8 separate 'Test' scenario saves, as follows:

Test1- French Town,Wall no, Great Wall no
Test2- French Town,Wall yes, Great Wall no
Test3- French Town,Wall no, Great Wall yes
Test4- French Town,Wall yes, Great Wall yes
Test5- French City,Wall no, Great Wall no
Test6- French City,Wall yes, Great Wall no
Test7- French City,Wall no, Great Wall yes
Test8- French City,Wall yes, Great Wall yes

The reason I created the apparently redundant combinations was partly for completeness, partly as an additional check for variation.
I figured the 1000 warriors would be enough repetition that each test should only need to be run once.I used stack move at the beginning of turn 1 to initiate combat vs the french, then left it to run, counting the surviving warriors once done to determine victories/defeats (I used 'unit costs' on advisor screen to get the number)
The results were suprising, to say the least!

Test1- 456 victories,544 defeats
Test2- 279 victories,721 defeats
Test3- 429 victories,571 defeats
Test4- 177 victories,823 defeats
Test5- 262 victories,738 defeats
Test6- 174 victories,826 defeats
Test7- 285 victories,715 defeats
Test8- 113 victories,887 defeats

Initial appraisal seems to indicate that there is some benefit to having a wall in a city, regardless of ownership of the Great Wall, which has nothing to do with what I was looking for!
I am loathe to draw firm conclusions at this stage, however, as I may be missing something.
I think I may have to run repetitions of these tests after all, despite the n=1000 warriors, just to get an idea if these results are reproduceable.
But before I do I'd like some constructive feedback on my test method.
(Pdescobar suggested to remove the veteran/elite promotion bonuses, but after some thought, I figured in this test I'm only looking for significant differences in results, so with 'all else being equal' the obfuscation factor of this aspect should only contribute to slight variation, particularly with warrior n=1000).

 

[amirsan]

In my opinion The Great Wall is a waste of shields. Dunno

 

[DaveMcW]

Obviously 1000 trials is not enough.

The two cases you sound interested in are city/wall/g-wall and city/none/none. Turn off unit promotions and run them both 10,000 times. That should be enough to see if there is a real difference.

 

[Tacit_Exit]

Originally posted by DaveMcW
Obviously 1000 trials is not enough.

The two cases you sound interested in are city/wall/g-wall and city/none/none. Turn off unit promotions and run them both 10,000 times. That should be enough to see if there is a real difference.

I hope you mean 10*1000 warrior assaults not 10000 runs of each test scenario! Each trial takes ~ 8 mins to run!

The original intention was to see if the 100% defense bonus conferred on walled towns by ownership of the Great Wall was reduced to the generic 50% city (7-12) bonus when it grows past 6.

I believed that walls just became redundant in cities (7-12), rather than 'deactivated', and that a Great Wall enhanced wall should continue to add 100% defense bonus to cities that built a wall before growing to 7+.

However, many posters here believe that the GW enhances cities to 100% defense bonus, irrespective of walls, mainly (I believe) because the editor label for GW states "Doubles City Defenses".
(interpreted as doubling the 50% bonus inherant to cities (7-12).

And yet more posters believe it has no effect on cities.

So I figured the only way to get to the truth was to test it!

And preliminary results seem to prove that there is some effect of walls in cities!(but no difference if only gw present; rejecting the idea that GW enhances the base 50% inherant in cities(7-12))

Having said that, I think I'll definately have to run these multiple times; I'll do as you suggest, 10 trials of each test, with promotions off this time( and preserve random seed off too of course!)

 

[Strider]

Interesting.... I normally build the Great Wall just because it doesn't take long and it gives you better defense (I think). Interesting effects through....

 

[GenMarshall]

I duild the Great Wall just to reap the benifits of an increace in Barbarian Defense . THough I wished it still had the "City Wall in every city" effect.

 

[nihil8r]

i just build it because it looks badass on the city screen and because it gives 2 more culture per turn to my nation =]

 

[Tacit_Exit]

Originally posted by CivGeneral
I duild the Great Wall just to reap the benifits of an increace in Barbarian Defense . THough I wished it still had the "City Wall in every city" effect.

Just remember the Great Wall has 2 distinct effects; the Barbarian Attack bonus, AND doubling of defense value of walls (vs any attacker)

 

[greynite1]

the Great Wall in most of my games has never been an issue because generally I take up my time with other wonders. Oracle, GL and hanging Gardens. I'm always staying out of conflicts unless I can absolutely help it and build up to a point where i'm ready to fight. Then usually the counter attack never comes. The other problem is because I never have enough Cities available to build ONE MORE Wonder. I'm always struggling with those first couple. Those are huge especially the GL and its a toss up between the hanging Gardens or the Oracle.

 

[TheNiceOne]

Great test. Unlike DaveMcW I think you have a large enough sample to make a conclusion.

If we compare with the numbers from the combat calculator we get the following expected results:

Def bonus / Attacker wins
10% 456
60% 291
110% 194
150% 144

Your test results get close enough to these expected number sto conclude IMO.

Test1 and Test3 with 456 and 429 falls into the 10% category and shows that a Great Wall have no effect on towns without walls.

Test2, Test5 and Test7 falls into the 60% category and shows that a town with wall has the same defense as a normal city.

Test4 falls into the 110% category and shows that the wall bonus in a town is increased from 50% to 100% with the Great Wall.

Test6 is really interesting. The result is close enough to the expected 110% category 174 vs 194 wins, which shows that a wall increases city defense from 50% to 100%. This makes walls much more valuable than I have thought, since the wall is effective for all cities (but possibly not metropolises).

Test8 confirms what we find in test7, that the wall and now also the Great Wall increase the defense of a city. Test8 shows 113 wins, which is a bit less than 144 wins of a 150% defense, but the real defense is 160% (which cannot be set in the combatcalc), so 113 wins seems close to expected for a 160% defense.

Now, maybe you could test this for a metropolis as well, if the wall increase the defense there from 100 to 150%?

This sure means that I'm going to build some walls in my border cities.

 

[Phantom Lord]

Nice test. My conclusion is that I will keep building walls in endangered towns but never build the great wall, unless I'd have to waste a prebuild otherwise. I'm also wondering why the AI often goes after the Great Wall but almost never has city walls in its towns.

 

[Tacit_Exit]

Well I ran multiples of the test as suggested, this time with veteran and elite HP set at 3 (seems to have only altered results for test 8 significantly).

I'm really rusty on my stats, so the deviations are probably calculated wrongly; please let me know and I'll fix it.

Thank you TheNiceOne for your great insights.
I used your figures to try to present what %defense categories each combo fits into.I will as you suggested, test Metropolises with the same test scenarios as a base and post them later; but for now, my head hurts too much from trying to dredge up what my poor math teacher tried to teach me!

I think its pretty clear that there's something weird about walls!

They do not, as advertised, become redundant in cities at all!

I set up this test to see if The Great Wall enhanced walls gave a benefit to cities (since I figured the 100% to 50% reduction in defense multiplier upon growth to 7+ in this case seemed illogical) and have uncovered what is probably a PTW 'bug'.

These tests do refute the opinion of some posters that the Great Wall's editor label 'doubles city defenses' means it doubles the base 50% defense bonus inherant to cities.

It also does not support the idea I originally had that wall's 50% bonus existed simultaneously (but not additively) with the 50% inherant to cities (hence my conclusion that a city with a Great Wall enhanced wall should have 100% defense bonus; not 50%).

This test was done with PTW 1.21f.

 

[Tacit_Exit]

And heres a zip of the test scenarios I set up if anyone doubts my results or wishes to check for 'artifacts' in the experiment.

All 8 Bix' is too big to attach so you'll have to add wall/great wall combos to this "Town,No Wall, No Great Wall" version to simulate all 8 test bix'.

 

[DaveMcW]

Wow. So the Great Wall works as expected, but city walls never become obsolete! This will definitely give the defender an advantage in multiplayer games.

 

[Tacit_Exit]

Originally posted by DaveMcW
Wow. So the Great Wall works as expected, but city walls never become obsolete! This will definitely give the defender an advantage in multiplayer games.

Yeah!

I'm gonna test the Metropolis' (metropoli?) next, but given that the town-city 'switch off' of walls doesn't seem to be working theres no reason to believe it does for the 12-13 transition either.

Metropolis + wall + civil defense =

 

[Hygro]

All I can say is "whoa"

Also test to see if the GW still works after obsolence. Probably doesn't but might be worth checking (hill metro behind river, civil defense, wall, GW, civil defense, rader tower, 4 elite mech infantry in an army, )

 

[Tacit_Exit]

Originally posted by Hygro
(hill metro behind river, civil defense, wall, GW, civil defense, rader tower, 4 elite mech infantry in an army, )

Just doesn't pay to walk out your front door these days!

 

[Hygro]

yeah really!

The evidence pointing to wall effectiveness after cities is amazing... this means I;m going to always build walls before it's too late!

Tacit I like the new avatar!

 

[pdescobar]

[Quoting Tacit_Exit]These tests do refute the opinion of some posters that the Great Wall's editor label 'doubles city defenses' means it doubles the base 50% defense bonus inherant to cities.

Good job Tacit. Interesting results. So the inherent defensive ability of a settlement is *not* a "city defense" for purposes of the great wall. Frankly, I consider that in itself a bug. At the very least, the editor flag for this should be renamed to "doubles defensive improvements" or something else less-ambiguous. It might be instructive to run a test using civil defenses to see if other improvement-based defenses are affected by the Great Wall.

It also does not support the idea I originally had that wall's 50% bonus existed simultaneously (but not additively) with the 50% inherant to cities (hence my conclusion that a city with a Great Wall enhanced wall should have 100% defense bonus; not 50%).

Yep. We were all wrong Oh, one other thing: the "expected" value for a 160% defense is 134, putting your experimental value for test #8 of 136 almost dead-on.

 

[Grille]

Great work on discovering this wall bug. Wonder if it gets fixed, either allow cities (and metros?) to have a build option for walls or put the defense bonus right.
Seems like city size .gt. 7 just triggers the wall graphic to disappear. How can I estimate enemy's defense bonus w/o investigation (and not doing an *in-game-test*)?
edit:typo