Test results -- Bombers -versus- spearmen in Cities

Using the “Early Air Power, Army, and Artillery Test Map” .bic file, I played through three quick sample scenarios (same map, just different random civs and different start positions.)

The start positions are all fairly close together so there can be lots of early contact and potential interaction. This is a tiny map with all four (4) test civs on the same continent. All engagements are in a fully up to date version of V1.21 with lethal bombardment engaged for all the aircraft.

All the civs can build the WWII style bombers and fighters right away without any special resources.

As a test, I try to build a fighter and a bomber early in the build cycle just as soon as I get the first ground defenders out of the way and start setting up for the settler expansion.

This gives the opportunity for lots of early engagements between warriors, spearmen, archers, and bombers.

The bombers have the standard Firaxis defined default bombardment power of 8 and rate of fire of 3, but I upped the range to the Firaxis maximum of 8 so I could perhaps reach some targets even in these early games.

Amazingly in all three cases, the AI civs targeted me for abuse and blackmail because I think they may be programmed to ignore the military power of any unit that does not have A and D values. I cannot yet confirm this, but in all three games on Emperor level, I was at parity or ahead in city build rate at somewhere near 6 or 8 cities at about 1700 BC when hostilities broke out. I was also at about parity with total numbers of units overall. In all three cases the AIs declared war on me, because I declined a total treasury + tech + map extortion demand. I could then usually reach only one or maybe two cities in AI territory at best. Usually the enemy capital was in range of the bombers but just beyond fighter range and barely visible at the recon fringe.

I stopped playing every 4 turns and used the CIV3 multitool to collect data from the cities that were under attack. Then I placed the data in spreadsheets to track the combat results.

I only used one or two bombers because I wanted lots of engagements to see what the hit success rate would be and I really was not interested in necessarily killing anybody outright.

The next posts to this thread will give you the engagement results:

All of these engagements are in capital cities with no other improvements except the default Palace. All three cities were located on plains or grassland. The typical civilian populations in the cities only range from 1 to 3 citizens at the most. In most cases the defenders are regular spearmen with a defensive value of 2 and 3 hit points of health. The expected 1 on 1 engagement ratio would have bombers with an attack of 8 vs spearman defending and result in 8/10 contests going to the bomber with 2/10 or 1 out of 5 going to the defender. With 3 hit points of health that would give the spearmen a 50/50 chance of surviving a bomber attack.

When defending a city the bomber attacks are more complicated, because the bombing mission could target civilians, improvements, or the defenders. In the case where a city has 1 improvement, 2 civilian workers, and 2 defenders there are five possible targets that could be hit by bombs.

In observing how these engagements play out, I can tell that the Firaxis code first picks out which target will be engaged using some sort of random number generation process. Then the target squares off against the attacking bomber using its defensive value in one of the standard Firaxis CIV3 “slugfests”. I am not sure what the relative defensive strength of a civilian is set to but I have seen it posted that city improvements have a defensive strength of 16. So if the engagement generator randomly chooses a city improvement for the bomber to engage, then the bomber gets 3 bombs at a hit of 8 to try and destroy the improvement at a 16. If the bomber fails, you get the all to familiar “Bombing Run Failed” message. If the bomber succeeds in destroying the improvement on the first of three bombs, then the other bombs just fall as duds and have no other effect.

Units that are defending in the city act as fortified defenders and always heal 100% by the next turn even if a barracks is not present. This means you have to kill them in one turn or start over in the next turn as if nothing has happened. I did not observe any combat promotions for the defenders as a result of any of the engagements (thank goodness for that bit of good fortune) so at least the defenders did not get stronger in each turn when they just stood there and took the bombing.

The RESULTS are fairly interesting:

In total there were 86 bomber missions (or Sorties) flown against the 3 small towns.

On the average it took 32 turns to defeat a small town that was initially defended by one or two spearman defenders. The air attack forces included one fighter and then one bomber increasing up to two bombers at the end. The fighters were always out of bomb range and were only used to illuminate the target for observation during each turn.

The bombing was not 100% continuous because of interruptions by barbarians and vicious 1 warrior assaults from other directions.

The enemy civs were defiant up to the end and refused to negotiate a peace treaty that included 1 technology even when they would respond to envoys.

In total, 258 bombs were dropped on the cities and their defenders and 93 hits were scored for a wounding hit success rate of 36%. Just under 50% of these wounds were fatal while the other half fully healed before the next bombing wave could arrive on target.

Only 20 of the 86 sorties resulted in kills of defenders or civilians in the cities and this was a 23% success rate. We should recognize that one bomb out of three would kill a civilian, but all the defenders took 3 bombs to get successfully smackoed just because the way the experiment was designed. An equal number of civilians were killed compared to the number of military defenders.

Workers, scouts, and settlers are immune to bombardment so I think they get totally ignored in the combat figures even when they are present in the city.

A total of 42 of the bomber sorties failed to hit anything at all and generated the dreaded “Bombing Run Failed” message. This was a 49% total failure rate.

There were three engagement sequences for the three towns, but they were all very similar, so here is just one example using Berlin including reference notes:

We are the Iroquois
Germany demands tribute of 42 gold (entire treasury and our only unique tech)
Germany is delined and Declares war
Settler Produced from Berlin -- first visibility over the city of Berlin by recon
Pop 1 working on a spearman -- 2 hits on spearman 1
Pop 1working on a spearman --2 hits on spearman 1
Pop 1 spearman produced -- bombing run fails
Pop 1 working on a spearman -- 1 hit on spearman 1
Pop 1 spearman defender killed -- 3 hits kills the spearman 1
Pop 1 spearman produced -- 2 hits on spearman 2
Pop 1 working on a spearman -- bombing run fails
Pop 2 working on a spearman -- bombing run fails
Pop 2 working on a spearman -- 2 hits on spearman 2
Pop 2 working on a spearman -- 2 hits on spearman 2
Pop 2 second spearman appears -- 2 hits on spearman 2
Pop 2 working on an archer -- distracted by barb horseman
Pop 2 working on an archer -- killed 1 citizen
Pop 1 working on an archer -- rebase to deal with barbs
Pop 1 working on an archer -- killed barb conscript horseman in open tundra
Pop 1 working on horseman -- rebase back to attack
Pop 1 working on horseman -- killed archer in the open
Pop 2 working on horseman -- bombing run fails
Pop 2 Germans declare war on China also -- 3 hits kill spearman 2
Pop 2 Regular warrior runs into city -- bombing run fails
Pop 2 Regular warrior runs out of city -- bombing run fails
Pop 2 still working on horseman -- bombing run fails
Pop 2 working on horseman -- bombing run fails
Pop 2 working on horseman 2 -- I am distracted by barbs again
Pop 2 working on horseman 2 -- horseman runs out and kills China Warrior
Pop 2 working on horseman 2 -- bomber kills 1 point horseman in the open
Pop 3 working on horseman 2 -- I am distracted by barbs again
Pop 3 working on horseman 2 -- bombing run fails
Pop 3 working on horseman 2 -- 3 hits kill spearman 3, 2nd bomber kills civilian
Pop 2 switch to spearman emergency -- 3 hits kill spearman 4, 2nd bombing fails
Pop 2 working on spearman -- 1 citizen killed, bombing run fails
Pop 1 working on spearman -- archer runs in from outside, fails, fails
Pop 1 working on spearman -- 3 hits kill archer, 2nd bombing run fails
(2 runs against roads, 1 hit) -- Russia allies with Germany and declares war on us
(2 runs against roads, 1 hit) -- China Archer waltzes into undefended Berlin

I think I discovered several things that could be bugs or logic flaws that make the bombing results much worse than expected:

First, is the presence of the Palace as an improvement in the capital cities that were used in this test. I think the Firaxis code may be targeting the palace, but making it impossible to hit. I bombed the undefended pop 1 towns on several occasions and the software let me bomb the town but never scored a hit in 8 different sorties with 24 bombs dropped at 3 per sortie. Since the Palace is present in the mix it adds a wild card that may be exempting anything from getting hit on a bombing run where the random number generator selects the only city improvement that is present. When you try to bomb terrain with nothing present that can be hit, then you get the no bombardment allowed symbol.

The implication of this wildcard is that in a pop 3 city there is one citizen that can never be hit by bombardment and there are some city improvements that cannot be destroyed due to overlaps in the improvement placement code. This may be artificially increasing the bombardment resistance of the cities and towns. If the Pyramids, Sun Tsu’s and/or Hoover Dam is present on the continent with the target civ, then each city and town would have the granary, barracks, and/or hydro plants as un-destroyable improvements that would just absorb bombing hits as unplanned extra defensive strength.

Wonders may also be factored into the miss rate since my guess is that wonders can never be hit by bombardment but are probably still included in the calculation of building count for the city or town.

Note, that I need to run this test on some simple cities with no improvements to see if undefended, pop 1 cities can still be targeted but not hit by any bombardment as well.

Third, there are some units that cannot currently be hit by bombardment (workers, settlers, scouts, explorers, catapults, cannons, etc.) When these units are present in the city or town they could be factoring into the miss rate.

Fourth, it would be impossible to kill veteran or elite military units in any sort of 1-on-1 engagement rate with bombers and these units would always be restored to full health in the next turn whenever they are defending a city. This is because the maximum hit rate of a bomber is 3 and the hits point of veteran and elite units would exceed that number.

Based on the hit and miss rates in this experiment, I would say that the number of bombers required to reduce a town to an easily captured level is roughly equal to the number of citizens plus the number of improvements times two - plus a factor for the defender strength. Since the defenders will fully heal in each turn they must be reduced in one turn or you have to deal with the process of plinking off the improvements

If you include the number of bombers to kill 2x the civilians and 2x the improvements then:

One additional bomber will kill a regular spearman defender 1/3rd of the time.
Two additional bombers will kill a regular spearman defender in almost every case.
Two additional bombers will also kill a veteran spearman defender in most cases.
Two additional bombers will kill an elite spearman defender in 2/3rds of the cases.

As you increase the strength of the defenders, the hit/miss ratio will shift downward to where it will require one additional bomber for each expected hit point of rifleman defender and about 5 additional bombers for each 4 expected hit points of Infantry defender.

It may not be even mathematically plausible to use bombers against mechanized infantry units defending cities. I think the math says 3 or 4 bombers per hit point, or 12 to 16 bombers for every 1 mech infantry unit and it is key to note that half that number spread over two turns would statistically have little or no chance of success.

All of these numbers assume the town is located on flat ground and they should be roughly doubled for cities of pop 7 or greater located on hills.

Strong defenders effectively shift the bombardment impact onto the civilian improvements and civilian population because there is a uniform probability that the bombardment will hit the civilians instead of the defender. When there are 3 civilians and 1 improvement located in the town with 1 defender, then 4 out of 5 bombing missions will target the non-military targets if all else is taken equally.

Other observations about bombardment:

AI settlers just get confused and shut down when the military units around them are being engaged in battle. When you bomb a spearman that is escorting a settler, you can kill the spearman if you have enough bombard hit points. You cannot hit the settler since he/she is currently exempt from bombardment. If you kill the spearman, the settler just freezes up and stands there. If you knock the spearman down to one hit point of health in territory that is not neutral or friendly to the spearman, it will retreat back to its own territory and the settler will be confused and just stand there in many cases.

Interesting results. I have a few questions and comments, though.

* You seem to suggest that a bombing run counts each idividual improvement, pop point and defender when deciding which will be attacked. I believe this is not the case. Rather, the PRNG simply decides if it will attack unit/improvement/pop point. I remember Soren mentioning that the percentages were roughly 50/25/25. For example, in your case (1 reg defender, 1 improvement, 3 pop point) the chance of hitting is not 20/20/60 as you suggest, but rather 50/25/25.

Other have reported that small cities with no improvements show a larger number of "bombing failed" messages, which is believed to be coming from the fact that the bombers try to hit an improvement or pop point that isn´t there to be hit. Your test confirms this, too.

* When the bombardment targets a unit, you suggest that the bombardment is compared to the unit´s defence strength. I don´t think this is the case, but it´s worth more testing. This would mean that a unit in mountains is harder to hit than a unit on grassland, but my own experience suggest that the chance is equal. I could be wrong, of course.

* I haven´t seen units heal to full strength if there is no barracks present in a city. Units heal 2 points inside a city, so a reg unit would heal to full strength even without a barracks, but an elite with 1 hp should not heal to full strength after one round. Did you change this in the editor?

* Bombers and fighters are not calculated in the overall strength comparison, so you are right the AI players underestimate your force.

Ouzh, don´t have time to read all this (but I will later today!), only skimmed it!

Thank you very much, cracker, for all that work!

I`m looking forward to understanding bombardment a little better after I carefully read your great posts!

Hurricane, thanks for posting and participating this discussion:

Originally posted by Hurricane
* When the bombardment targets a unit, you suggest that the bombardment is compared to the unit´s defence strength. I don´t think this is the case, but it´s worth more testing. This would mean that a unit in mountains is harder to hit than a unit on grassland, but my own experience suggest that the chance is equal. I could be wrong, of course.

Click to expand...

I am fairly certain that this works with the defensive strength of the unit because just in the first passes the results of bombing archers (D=1) versus spearman (D=2) would show significantly different results. With bombers hitting at 8 strength and a rate of fire of 3. Regular Archers are almost always killed with one sortie. Regular Spearmen are rarely killed with one sortie and in fact there are regular occurances of only 1 hit or only two out of three hits.

I think I am interpreting the engagement rules correctly in that once the RNG determines randonmly that the Spearman is the target, then the rate of fire is used to determine the number of shots and the bombard strength is used against the defense strength modified by any terrain or city defense bonuses.

* I haven´t seen units heal to full strength if there is no barracks present in a city. Units heal 2 points inside a city, so a reg unit would heal to full strength even without a barracks, but an elite with 1 hp should not heal to full strength after one round. Did you change this in the editor?

Click to expand...

I did not change this in the editor and I am fairly certain that fortified defenders fully heal in the city in just one turn if their damage was incurred in the defense of the city. If you move a wounded defender into the city it still takes several turns to heal but if you have an elite defender that takes 4 hits of damage during a defens of the city, then WALA the next turn that elite unit is back to full 5 point strength.

On a side note, I agree that defenders should earn combat promotions but have to relay a funny story that was frustrating:

"Overwhelming stack of attackers approachs evil zulu city early in the game. zuly city is defended by one regular Impi and one regular warrior.

First regular archer attacks impi with a hit point sequence of:
win lose win lose lose die -- Remaining 1 point reg impi is promoted to 2/4 veteran
Second regular archer attacks impi with sequence:
win lose lose die -- remaining 1 point impis is promoted to 2/5 elite.
Third attacker is regular warrior attacks impi with sequence:
lose win lose die -- remaining 1 point impi vanishes behind warrior.
Fourth attacker is regular warrior attacks zulu warrior with sequence:
win lose lose die -- remaining 2 point warrior is promoted to 3/4 veteran.
Fifth attacker is regular warrior attacks zulu warrior with sequence:
lose lose lose die - 3 pt zulu warrior is promoted to 4/5 elite.

Attack stack is depleted:

Scavange 1 horseman from nearby frontier city to hurl at warrior in desperation:

horseman can reach the city by road at absolute limit of movement, attack sequence goes:

win, win, win, win zulu warrior dies,
horseman is victorious and captures last zulu city.ne zulu city respawns about 6 tiles further north and 1 archer, 3 spearman and a settler rush defiantly out .

I am keenly following the thread on lethal bombardment and appreciate the empirical data and explanations submitted by cracker. Well done. I also appreciate the academic arguments put forth by beeblbrox on the same subject, in another thread. The discussions are both educational and entertaining.
I am going to chime in with my two marks worth. I have been playing since Civ1, and have had a different opinion of artillery with each release. I am going to leave my view on lethal bombardment of ground units until the end of this post.

In my opinion, Civ3 treats cannons and such realistically in some ways: bombard range, need for additional defensive unit to avoid capture, low movement. However, Civ 3 is absurd in other ways: lack of lethal bombardment for ships, immunity of workers, settlers, aircraft, inaccuracy of bombardment. The implementation of bombardment in Civ3 is often just contrary to the historical record. Gunfire SINKS ships- that is the whole point of a battleships guns. Sure, the guns will miss a significant fraction of the time and register non-lethal hits, too. But sometimes big guns will sink a ship with a few well-placed salvoes. Land-based artillery should be able sink ships, as well, which is the point of coastal artillery. With some non-zero probability, all artillery should be able to sink a ship. I just think the probability must be at least non-zero - it's contrary to history other wise. A significant concentration of artilery fire should sink a ship within range. On rare occasions, a single shot. Remember the Hood? It happens. Similarily, a number of cruise missiles should sink a ship.

Bombers shopuld be able to sink ships, as well, with some non-zero probability. we have to open our minds here and think about what we mean by "bomber". If we just think of B-17s, or big level bombers in general, we're too narrow. Civ3 is too narrow with having only bombers and fighters. Where does the Ju-88 two-engine bomber fit in? Or the B-25? The Ju-88 was an effective ship-killer in WW2, both as a dive bomber and torpedoe plane. Similarily, numerous U-boats were sunk by Allied aircraft, including even large seaplanes dropping depth charges. Does that count as bombing? The Tirpitz was sunk by a high-altitude British four-engine level bomber, although after several unsucessful attempts. Level bombers have missed ships ALOT - think of the B-17s in the Pacific. However, bombers CAN sink ships, especially merchant ships and transports. The Luftwaffe closed the Channel to British ships in 1940-42 so that no large ships were allowed in the Channel. Fw-200 four-engine bombers sank ships in mid-ocean in the convoy battles. Remember the Arizona? One bomb fashioned from a battleship shell exploded in the magazine and we know the rest. OK, it was a single-engine carrier-based bomber, probably what would be classed as a fighter in Civ3, but does it matter? Sometimes bombing sinks ships. One shouls be aable defend one coasts with guns and bombers that can at least stop and sink unescorted transports.

What about ground units? Bombing and artillery have not been shown to destroy infantry, armor, etc. as a fighting force in any significant historical scenario. This is true even when unprecedented amounts of artillery have been used, or large numbers of bombers. Modern units survive bombardment. In the real world, bombardment of enemy combat units has come to serve to wear the enemy down though constant harassment, attrit them, and de-moralize them, making them easier targets for the assault by your own ground forces. In this way, Civ3 is pretty realistic, in my opinion. Still, I can not accept that enough bombardment does not destroy certain units. I can not believe that suffcient fighters/bombers/cannons do not destroy a spearman unit in open terrain. Or destroy an ancient age city with ease. Or sink galleys. Maybe a "relative era factor" would make sense, to avoid weird battle results. Ancient era units can not be considered as having the training and experience to remain cohesive under bombardment. Sufficent bombardment should render them "destroyed" as far as the game is concerned. This reminds me of the old Civ games when a phalanx on defense sometimes took out a battleship. Yeah, like that could happen.

What I do think is that bombardment accuracy should increase, so that more hits register and the effectiveness of artillery/bombing is incressed. Also, a 'strategic bombing mode' should be allowed to specifically bombard city improvements, population, workers, etc. That was the point of the Allied bombing campaigns against Germany and Japan. Similarly, a "tactical mode" to target combat units. Targeting combat units in cities would result in "collateral damage" to improvements, population, etc, SOMETIMES, not most of the time. Ships, cannons and aircraft would be vulnerable targets, like in real life. Remeber, the Battle of Britain was primarily a campaign against RAF airfields. Pearl Harbor? Ships and aircraft in a port city DESTROYED by aircraft. Little collateral damage.

Enough for now,

Great work, Cracker, I'm impressed.

I've also been following this discussion rather closely. I certainly appreciate the historical argument. But as Cracker's tests indicate, the way bombardment is implemented now, even with the lethal ability, it takes an insane stack of artillery/bombers to actually take a city. So from a gameplay standpoint, what harm is there in allowing artillery to actually destroy some "individual" units?

I wholeheartedly agree with the artillery bombardment of ships, and believe that by giving cannons on up lethal sea bombardment, we can really correct a significant flaw in the "Coastal Fortress".

Homegrown
Flogging will continue until morale improves.
Now playing: Pharaoh | Zeus |Civ III | Sim Golf | Zoo Tycoon

Are you ready to take the challenge?
Monthly contests for many of your favorite games.

Defensive value according to version of the game.

Original version: building=4 citizen=4
1.16 version: building=8 citizen=8
1.17 version: building=16 citizen=16
1.21 version i didnt try it.
I think its better to lower def value from 16 to 8 into civ3 editor, otherwise there is way too much bombardement fail.