Tasty if true

[Megalou]

I played a game with lots of mounted units gallopping around killing each other for a long time. I noticed, or rather got the distinct feeling, that units who have to take many steps (within their turn) before attacking had a lesser chance of winning. I recalled a popup from Civ2: "Sir, these forces are tired. If they attack..."
I haven't read anything about this and I've read quite a lot. Eerie...

PS The terrain on the defending units' tiles didn't seem to be the cause, nor rivers. Possibly, the defending units could have been fortified (with 25% defense bonuses). Logically, far away enemy units are perhaps statistically more likely to be fortified. But sometimes the defenders were units that couldn't have been fortified, since they were footsoldiers crawling into my territoriy trying to reach my resources.

 

[satchel]

I've never noticed the effect you mention, and I don't think it is real. However, it is easily testable under controlled conditions. I don't have access to the game right now, so I cannot. But it would be quite simple for you to create a scenario with (say) 100 cavalry and 100 riflemen. According to the combat calculator, if both units are veteran and the rifleman is fortified on grassland, the cavalry should win 34% of the time. Take your 100 cavalry and attack the rifleman on your first move, and record the number that wins - it should be roughly 34%. Then repeat, except attack on the cavalry's second move. Finally, try attacking on its third move. See if the results vary by a statistically significant amount. Report your results here and become famous!

 

[TheNiceOne]

I think its fairly safe to say that we now have discovered every factor that goes into battle resolution, and number of squares moved (or anything similar) does not play any part - only attacker's attack rating, defender's defense rating, and all defense bonuses for terrain, fortifying and radar towers.

And I have often attacked an enemy city where I start within my borders, move two squares within enemy borders and then attack (with cavalry) - and I've never noticed that those units lose more than normal. This is a good tactic btw, because units than are at least three squares away are not considered by the AI to be a threat, so it will not reinforce the city.

 

[Megalou]

Umpf. I made the test suggested by sachel. (Not quite the same because I don't know how to disable promotion in the editor.) I set up one scenario where I let AI cavalry attack my regular riflemen from one square away, then one where they attacked from three squares away. Testing with a total of 400 riflemen for each scenario, 4% more riflemen survived in the second scenario. Obviously I have let my imagination run away with me.

Silly thread this. It should be reported. Thanks for the answers, though.

 

[Megalou]

Note: the cavalry in the test always attacked because they were on a small island.

 

[warmonger]

Megalou - just out of interest could you do this test the other way round? that is your cavalry attack the AI rifleman. Afterall.... I certainly wouldn't expect the AI horseys to get tired.....

 

[MadHatter]

Another interesting variation is to attack the cavalry with the riflemen in one move and compare that to the riflemen attacking the cavalry in their second move (include a road to move on). This I daresay may be more interesting because when the cavalry attack on their second move, they still have more than a full movement point left, the riflemen (technically) have 2/3 of a movement point left. Will this affect the riflemen's success rate?

 

[satchel]

Originally posted by Megalou
Silly thread this. It should be reported.

Not at all! you are a very good sport - thank you for doing the test!

 

[Chieftess]

Well, I did a test, the number of cav 1 tile away was less than the ones standing next to the riflemen. I used 50 reg units in each stack, on grasslands.

2 tiles away (I moved the unit back 1, and foward twice)

33 Cav survived (67%)
8 promoted
11 at full health (including 3 vets)
7 at -1 hp (including 3 vets)
15 at -2 hp (including 2 vets)

31 AI riflemen survived (62%) 6 became vets, but doesn't account for ones defeated.

1 tile away:

32 Cav survived (64%)
2 promoted
7 at full health (including a vet)
6 at -1 hp
19 at -2 hp (including a vet)

29 AI riflemen survived. (58%) (one was promoted to elite, which could have obscured the results). 8 were promoted, but doesn't include ones that were promoted, then lost the next round.

adjacent tile:
39 Cav survived (78%)
2 at full health (including a vet)
11 at -1 hp
25 at -2 hp

29 AI survived, (58%) 2 of which were promoted (again, can't account for ones promoted then defeated).


-1hp = 2/3, 3/4, 4/5
-2hp = 1/3, 2/4, 3/5

 

[JAWiseman]

Megalou-
I have noticed the same as you. It destcintly feels like when units move more than one square they perform worse than when adjacent. In fact, I have begun to move these units "into position" for an attack, rather than take the perceived "penalty". This was all done based on my perception, and not on any facts...

It appears that Chieftess test confirms a higher survival rate for adjacent attacks? She never concludes her post.

 

[MadHatter]

Wow, Chieftess, I wish I had as much free time to play Civ as you apparently do...

 

[Megalou]

Yes JAWiseman, 78% was a high number. But there were a lot of those cavs that had lost two hit points. Interesting that you had the same feeling, though. I might do my scenarios with me as the attacker since you all are interested. But not right now. (Check the posting time.)

It's an eerie fact that we tend to think of the game as more "alive" than it could possibly be. (Sometimes I even get the feeling that the computer punishes me for mere stupidity.) Possibly, the constructers could conceal a few rules to encourage this feeling, but my basic assumption now is that someone should have discovered before if units did in fact become tired before the end of their movements.

Also, I think it would be a poor rule, because the ability to calculate how far a unit can reach is one of the main skills of the game and should not have any downsides.

Thanks!

 

[Megalou]

Originally posted by Megalou
Umpf. I made the test suggested by sachel. (Not quite the same because I don't know how to disable promotion in the editor.) I set up one scenario where I let AI cavalry attack my regular riflemen from one square away, then one where they attacked from three squares away. Testing with a total of 400 riflemen for each scenario, 4% more riflemen survived in the second scenario.

This description was messed up. What I meant to say was that 4% more riflemen survived when attacked from one square away, That was a coincidence for sure, but didn't support my theory in any way.