Bombardment Investigation

[etj4Eagle]

Originally posted by Lt. 'Killer' M.
sumthinelse: I have never seen an aqueduct destroyed - but I have seen a city of mine loose the last Temple it had. Also, you cannot sell aqueducts - so safe to say it is hardcoded you cannot destroy it (not saying you can't hit it)....


(but I guess you knew that).

That was another thing that my test city did not have (aquaducts and hospitals), since I assumed that you couldn't destroy them either. And I wouldn't be too suprised if the game attempted to say that was the building you destroyed at times causing extra misses.

 

[sumthinelse]

Originally posted by etj4Eagle


That was another thing that my test city did not have (aquaducts and hospitals), since I assumed that you couldn't destroy them either. And I wouldn't be too suprised if the game attempted to say that was the building you destroyed at times causing extra misses.

Yes, I will run some of those tests again without aquducts. What I have in mind is to make the bomber's bombard strength 1000 so that there are very few misses, and use modern armor as defender in the city -- unfortified it should be 16, the same as citizens and buildings, so the % of each hit type should be almost the same as the % in target selection: 1000/(1000+17.6) is about 98%.

 

[sumthinelse]

PTW, city with at least one bombardable building and between 2 and 5 citizens. Defender, modern armor so defense should be the same (16) for mil. unit, buildings, citizens. Bombard with modded bombers with bombard str=1000. 200 bombs dropped.

Mil. unit hits: 64, 32%

Citizens hit: 76, 38%

Buildings destoyed: 58, 29%

Misses: 2, 1%.

Still nowhere near Soren's 50%/25%/25% for mil. unit/citizen/buildings.

 

[etj4Eagle]

Originally posted by sumthinelse
PTW, city with at least one bombardable building and between 2 and 5 citizens. Defender, modern armor so defense should be the same (16) for mil. unit, buildings, citizens. Bombard with modded bombers with bombard str=1000. 200 bombs dropped.

Mil. unit hits: 64, 32%

Citizens hit: 76, 38%

Buildings destoyed: 58, 29%

Misses: 2, 1%.

Still nowhere near Soren's 50%/25%/25% for mil. unit/citizen/buildings.

Sumthinelse it looks like they have changed the ratios in PTW. Your results seem to either suggest an even probability for each structure or maybe a 30:40:30 ratio. Not exactly sure why they would have changed the ratios. The only thing I can think is that because of the new modes they decided to make it a little harder to red line the defenders.

 

[sumthinelse]

Originally posted by etj4Eagle


Sumthinelse it looks like they have changed the ratios in PTW. Your results seem to either suggest an even probability for each structure or maybe a 30:40:30 ratio. Not exactly sure why they would have changed the ratios. The only thing I can think is that because of the new modes they decided to make it a little harder to red line the defenders.

Patch 1.14f. No mention of this in the reame.txt.

 

[DANGERBOY]

Not everything goes into the readme.

 

[sumthinelse]

Originally posted by DANGERBOY
Not everything goes into the readme.

But they did put the increases in civ and building defense (2 times) in the civ3 readme.txt. PTW, different story. Would have been nice....

 

[sumthinelse]

PTW again. PBEM game. I wonder if that is different from single player?

This time modded cannon: bomb. str=1000, rof=1.

Target: warrior in size 1 city, no improvements.

400 tries, 128 hits, exactly 32% hits against warriors, like the bomber test before.

 

[sumthinelse]

@ etj4Eagle: I tried bombarding a city in civ3 and got 99 hits on a warrior out of 200 tries with bombard str=1000 against a warrior.

This agrees with your, cracker's and soren's formula of 50% target selection of military unit in a city for civ3.

I haven't tried PTW single player yet.

But you should get PTW just for setting up testing! In debug mode you can add citizens and buildings to cities, add any unit anywhere, see inside enemy cities, and of course see the whole map. I feel like I should buy PTW for you to make your life easier! Setting up a test scenario in CIV3 now feel like beating my head against a wall!

 

[sumthinelse]

PTW single player: 400 trial test, military unit hit 131/400 times = 33%, so apparently no difference between single player and multiplayer in PTW.

 

[etj4Eagle]

Originally posted by sumthinelse
@ etj4Eagle: I tried bombarding a city in civ3 and got 99 hits on a warrior out of 200 tries with bombard str=1000 against a warrior.

This agrees with your, cracker's and soren's formula of 50% target selection of military unit in a city for civ3.

I haven't tried PTW single player yet.

But you should get PTW just for setting up testing! In debug mode you can add citizens and buildings to cities, add any unit anywhere, see inside enemy cities, and of course see the whole map. I feel like I should buy PTW for you to make your life easier! Setting up a test scenario in CIV3 now feel like beating my head against a wall!

Am planning on getting PTW sometime in January. Currently in the process of preparing to move to OH at which point I will be finally employed again (ah the nicety of collecting pay checks). But until all of this chaos is taken care of I don't want to tempt myself with PTW

Is it possible that this probability change was made in the base version of PTW? If so then that could be why it never appeared in any readme. Now if only precision bombing would target units after all buildings were gone and the city was at size one we would at least have a way to avoid massive miss rates.

 

[sumthinelse]

Originally posted by etj4Eagle

Is it possible that this probability change was made in the base version of PTW? If so then that could be why it never appeared in any readme. Now if only precision bombing would target units after all buildings were gone and the city was at size one we would at least have a way to avoid massive miss rates.

I think it's very likely to be the same in base PTW as PTW patch 1.14f (but different from civ3) but don't want to uninstall to find out.

There is a way to avoid massive miss rates of course -- edit the bombard strength in the editor!

One thing I will do to make the probabilities the same as citizens get killed by bombardment: Change the defense bonus for both city and metropolis to 0 so you can start at 30 citizens and the defense bonus does not change as the population goes down....

 

[Catt]

Originally posted by sumthinelse


I think it's very likely to be the same in base PTW but don't want to uninstall to find out.

Wait - don't your tests indicate pretty clearly that the unit targeting percentage is probably 50% in Civ 3 but probably 33% in PTW?

 

[sumthinelse]

Originally posted by Catt


Wait - don't your tests indicate pretty clearly that the unit targeting percentage is probably 50% in Civ 3 but probably 33% in PTW?

Yes. Eagle asked if I thought that the %s were changed (from the numbers in civ3) in base PTW (this might explain why there was nothing about it in any of the PTW reademes) so I said yes, probably, but I don't want to uninstall PTW patch 1.14f to find out (to test base PTW). I think it's 50% in civ3, and a little over 30%, probably 32%-33% in PTW.

 

[sumthinelse]

Latest results for PTW bombardment:

Selection of targets in town/city/metro:

-----------UNIT--------------CIVILIAN------------IMPROVEMENT

PTW-----30%--------------30%------------------40%

CIV3-----50%--------------25%------------------25%

I changed the civilian and improvement defense bonus to 1, and bombarded warriors (defense 1). I set the bombard strength to 1000 so that when a target was selected, the probability of a hit was almost certain (1000/(1000+1.1))=.9989. I bombarded a city with only a palace, because a miss would almost certainly be a selection of improvement. Out of 400 tries, 121=unit, 118=civilian, 161=miss (improvement).

 

[sumthinelse]

I found a note in the civ3xedit help: "Any Improvements or Wonders that have a defensive value against bombardment (> 0) is considered walled."

I think City Walls is the only improvement that has such a value, but what would "walled" mean besides having a defense against bombardment???? AFAIK the only way to attack a building is to bombard it.

 

[cracker]

You have to be very careful here in these tests because there are several built in bugs in the bombardment system that make the results come out significantly worse than expected.

For example, a town with pop 1, and no defenders and no buildings can never sustain a bombardment hit because there are no targets.

a town with pop 1, a palace, and no defenders can can never sustain a bombardment hit because there are no targets.

You can bombard a town with a free granary from the pyramids, and you could hit it with 10,000 cruise missles and never score a hit.

you can bombard naval units legitimately but when they are in cities or towns you can never score a hit on them with bombardment.

the miss rate artificially goes up in towns or cities with "undestroyable" buildings and/or the last undestroyable citizen and/or military units that sometimes cannot be targeted.

Unless you carefully document these "screw up conditions" it is tough to draw any logical conclusions for bombardment in towns or cities.

I provided Thunderfall with a number of these examples in screenshots for the SOTD, but I think they got pushed to the side because they were not something insanely screwy with PTW.

 

[sumthinelse]

Originally posted by cracker
For example, a town with pop 1, and no defenders and no buildings can never sustain a bombardment hit because there are no targets.

a town with pop 1, a palace, and no defenders can can never sustain a bombardment hit because there are no targets.

Absolutely. I made no such tests. But it's not really a bug -- it's just Working As Designed - WAD.

You can bombard a town with a free granary from the pyramids, and you could hit it with 10,000 cruise missles and never score a hit.

Unless there is more than one citizen or a military unit with hp>1 in the town so they would have something to hit.

you can bombard naval units legitimately but when they are in cities or towns you can never score a hit on them with bombardment.

Didn't know that but no such test. Sounds like a bug.

the miss rate artificially goes up in towns or cities with "undestroyable" buildings and/or the last undestroyable citizen and/or military units that sometimes cannot be targeted.

Yes I agree. What I did was to bombard a city with a single improvement and modded the bombard strength =1000 and civilian defense = 1 and the units were unfortified warriors (defense=1.1) so that if a unit or civilian was targeted it would hit about 99.9% of the time, thus all misses indicated the building was targeted. Also I set the city and metro defense bonus=0 so I could kill many civilians without changing the odds.

I was thoughtful and careful but I welcome your questions.

I provided Thunderfall with a number of these examples in screenshots for the SOTD, but I think they got pushed to the side because they were not something insanely screwy with PTW.

Thanks Cracker, I would like to see them. Are they uploaded here?

 

[Jaybe]

Does city size (town, city, metro) affect bombardment defense for units or cits/improvements??

 

[sumthinelse]

Originally posted by Jaybe
Does city size (town, city, metro) affect bombardment defense for units or cits/improvements??

It should affect both defense against a "normal" attack (attack by knight, tank etc.) and also against bombardment. But I have not actually tested bombardment against cities (7-12), 13+.