Page 1 of 1

Sighting distance and ai surrender

Posted: Thu Apr 05, 2001 1:15 am
by john g
Perhaps someone else has noticed this and can explain what is going on.

I set up a test scenario to see which was the best towed AT gun of 1943, the 88flakAT or the 17 pdr AT gun. Using the same starting positions for them and their targets the results of the test should have pointed out a winner. Strangely, the test just gave me new questions to ponder.

The 88's were able to see and fire further than the 17pdrs even when they were given the same crew experiance and officer ratings, is there a modification to the sighting distance for a unit based on its rangefinder value or fire control value? Or perhaps a larger gun (sitting higher off the ground) has a larger visual horizon?

The other puzzling aspect of the test was the point values at which the opposing force surrendered.

The defending AT guns were about 1000 points, all bought from the German list; the attacking force was approx 9000 points of T34 m42's. The Soviet side surrendered with as few as 4 tanks destroyed and always surrendered to the 88's with fewer losses. The scores of the 88's ranged from about 440-120 to 1600-280 while the scores of the 17pdrs were more like 2500-600 to 4400-800. Is there a built in psychological test in the ai to cause it to surrender when it knows it is clearly just there getting slaughtered? Why would it surrender when it has only lost 4 tanks out of nearly 4 regiments?

In back of the AT guns were neutral vp hexes totalling 750 points set as the movement objective of the Soviet side, but they never continued the game long enough to make it there. When I set up the test I assumed the 9-1 advantage would give the Soviets enough tanks to blow through the AT gun line and destroy all the guns, unfortunately they don't continue past about 13 tank losses to the 88's or 30 lost to the 17pdrs with even lower losses in some of the battles.

At least some of the AT guns survived all the battles.

If anyone has any answers, I would be glad to hear them. It has me stumped as to which towed AT gun rules 1943.
thanks, John. :confused:

Posted: Thu Apr 05, 2001 1:52 am
by Antonius
I have always assumed the AI surrenders when a given fraction of its units is destroyed or routed AND it controls no victory hexes.

As units can retreat/rout because other units in the same formation are destroyed, especially when it happens in the same turn, your AT guns can in effect rout the entire ennemy force by taking out a fraction only of the attacking tanks.

If you want your test to last longer, you should give the AI one vicory hex from the start so it won't surrender.

Posted: Fri Apr 06, 2001 11:08 pm
by john g
Originally posted by Antonius:
I have always assumed the AI surrenders when a given fraction of its units is destroyed or routed AND it controls no victory hexes.

As units can retreat/rout because other units in the same formation are destroyed, especially when it happens in the same turn, your AT guns can in effect rout the entire ennemy force by taking out a fraction only of the attacking tanks.

If you want your test to last longer, you should give the AI one vicory hex from the start so it won't surrender.
With losses as few as 4 tanks out of almost 4 regiments? I know from experiance that the ai will surrender when it has broken and has no victory hexes, but I always thought the break point was 75% losses.
Here are some of the scores the at guns acheived.

88 flak AT
1660-189
495-136
1668-532
930-191
1421-273
1515-336
1454-137
1167-286
1527-206
1408-126

17pdr AT guns
3132-844
3125-656
2698-559
1922-832
4487-515

Based on these numbers the 88 flak is the better AT gun, it managed decisive victories several times when the 17pdr never did. The primary differance is lower number of guns killed due to the larger crew size, several of the 88's were down to 3-4 crew when the Soviets surrendered, a 17pdr would have been destroyed if it lost that many crew.

This resolves nothing about the ability of the 88 to spot enemy units further away with identical crew.
thanks, John.

Posted: Sat Apr 07, 2001 1:16 am
by ZinZan
Unless spotting distance is based on number of crewmembers; that would mean a higher vhance of seeing at greater range. This would make sense (of a sort).

Posted: Sat Apr 07, 2001 1:40 am
by Redleg
To a certain extent, you were also testing German troops against British troops unless you had crewed both guns with troopers of one nation or the other.

Surrendering is also affected by the quality of the troops concerned.

Posted: Sat Apr 07, 2001 7:28 am
by McGib
Dont quote me on this but I believe that the AI will also surrender when/if a certain percentage of units are routed/retreating. Next time you try the 88's try another country, even fellow germans.

Posted: Sun Apr 08, 2001 12:39 am
by john g
Originally posted by Redleg:
To a certain extent, you were also testing German troops against British troops unless you had crewed both guns with troopers of one nation or the other.

Surrendering is also affected by the quality of the troops concerned.
The At guns were all bought from the German list with a German A0 unit. I even went so far as to edit all the exp and leaders the same, and another time used the change unit to another type in the deploy screen so that the units would be the same. My next test is to start messing with the 17pdrs rangefinder and fc to see if that is what is allowing the 88 flak to see further.
thanks, john.

Posted: Sun Apr 08, 2001 12:45 am
by john g
Originally posted by McGib:
Dont quote me on this but I believe that the AI will also surrender when/if a certain percentage of units are routed/retreating. Next time you try the 88's try another country, even fellow germans.
Since I set this up as a scenario instead of a quick battle, the at guns were always fighting the exact same opfor. The Soviet 43 army should not be surrendering at less than 20% losses like it was here.
thanks, John.

Posted: Sun Apr 08, 2001 12:57 am
by Paul Vebber
Send me your test scenario and I will see if I can figure it out

Posted: Sun Apr 08, 2001 8:59 pm
by Spike
Hmm, I did a similar test just with 88s vs T34s after reading this and the Soviets won rather handedly at 7-1 odds.. (alot less points though, which may make all the difference) I'm not sure these tests will be worth anything soon though.. once 5.0 is out the special opfire will be changed a bit, and large attacks should work a bit better. No problems surrendering either.. just follow the advice from Antonius and give them a victory hex at the beginning.. also you may want to make the VH's higher valued, I always assumed the AI wasn't willing to lose tons of equipment unless the VH's were worth more than what they lost. (hopefully that was clear!) :)

Posted: Sun Apr 08, 2001 9:45 pm
by Wild Bill
The advice of Antonius is probably the best way to handle it. I've begun assigning one victory hex to the AI, setting it a 20 points and causing it to be "hidden" in a corner until the last turn of the scenario.

This seems to keep the battle going to the end. This is a long standing problem stemming back to SP3 and it has never been corrected...WB