|
|
angelito
Honorable
Undefeatable Hero
proud father of a princess
|
posted November 08, 2006 05:57 PM |
|
|
Quote:
The question is: What beats?
There is no doubt that there are the best 2 level 6 units (Better than cyclop).
Good question. The main thing what is done wrong mostly, is to compare these creatures to other creatures of the same level. But it was mentioned from other posters many times before, this will never give a clue about their strength and usefullness. Nagas are the main force in middle game of Tower, while dreads never play an important role in the army of Necro (lords and skellies do!). That's the reason why for me, Efreet Sultan is the best level 6 unit. You hate to fight them as zone guards...u hate to fight them as guards in general. But they are so strong early week 2 with Inferno. Speed 13 and fireshield....besides archies, (hundreds of) demons, Sultans are the main power for Inferno. Cyclops are last among all level 6...no level 6 unit is less usefull than these 1-eyed crystal eating foos..
____________
Better judged by 12 than carried by 6.
|
|
russ
Promising
Supreme Hero
blah, blah, blah
|
|
posted November 08, 2006 06:39 PM |
|
|
Quote:
no level 6 unit is less usefull than these 1-eyed crystal eating foos..
What do they need crystal for, anyways? 
Aside from the crystal cost and the building order, unupgraded clops are the best deal for a level 6 unit. You can get 5 clops for every 2 knights!!!
|
|
bloucester
Famous Hero
MSN can go to hell
|
|
posted November 08, 2006 08:47 PM |
|
|
if you wanna know what the cyclops need crystal for, check heroes2 
|
|
tenzor
Adventuring Hero
|
|
posted November 09, 2006 01:05 AM |
|
|
Quote:
You didn't mention that death knights are harder to kill.. +10 life and +2 defense..
I calculated DK and Naga stats assuming not only the basic damage is doubled, including the curse advantages, and the conclusion I've reached is:
The damage of the Black Night is a bit lower than Nagas but it's alot harder to kill. While subtracting the damage bonus of the nagas from the DK's bonus, I can reach the conclusion:
DKs are 20% harder to kill, and +2 speed.
Nagas have no retalation.
The question is: What beats?
There is no doubt that there are the best 2 level 6 units (Better than cyclop).
Just for fun, I have created application simulating HOMM III battle 1 vs 1. It includes some units with their specialities and calculates battles with best tactic used (faster unit attacks against +2 defense of slower & slower does retaliation in round 1, faster attack slower + retaliation of slower, slower attacks faster + retaliation of faster in round 2 etc.).
I must admit that I was wrong if I said that naga queen would beat dread knight in fight 1 vs 1. Depending of number of units (1 vs 1 is not the same as 1000 vs 1000 ), dread knights will win from 54.5 to 61 percent of fights. Their success is based on speed - if nagas would be faster (e.g. by haste), dread knights have no chance. In case of equal speed, they would have only little chance to win (around 15%). This is simply given by more total damage per enemy hitpoints in 1 round caused by nagas and no simulation is needed to confirm it.
But let the speed of dread knights and naga queens unchanged. I tried put them agains various enemies and the results are interesting.
round 1: 100 dreads/q.nagas vs 750 royal griffins
Poor griffins, they cannot use the unlimited retaliation here. Dread knights won 64% of fights while naga queens only 27%.
round 2: vs 650 swordsmen
Dread knights won less than 0.5% of fights, queen nagas 91%! Great victory for nagas..
round 3: vs 3000 pikemen
Teoretically, dread knights should be better here as pikemens damage would be halved in case of cursing. But they won less than 10%, while nagas 61%.
It seems that naga queens are better than dread knights in case that they fight against 1 stack of slower units. But units faster than nagas & not faster than dread knights are much easier work for dreads than nagas.
Also in fights agains units faster even than dread knights, they are still better choice than queen nagas.
And one tip in the end: dread knights are slightly better in higher numbers. I think that this is caused by fact that more units means more rounds of combat -> more chance to death blow or cursing enemy..
Of course, this is only 1 vs 1 fight and as Angelito said, Nagas are main offensive force of Tower while dread knights are useful but not essential for necros.
|
|
HoMMage
Hired Hero
Unundead
|
|
posted November 09, 2006 04:35 AM |
|
|
Interesting... you should put them up against Wolf Raiders sometime.
Anyway, the best unit is any unit in the hands of a Hero who has at least one Phoenix/Archangel, expert Air and expert Tactics
|
|
bloucester
Famous Hero
MSN can go to hell
|
|
posted November 09, 2006 10:14 AM |
|
|
|
speed means victory in higher number of somewhat equal units. in cases of champions vs dread knight vs war unicorns, the faster always wins. ( or the death blow )
|
|
russ
Promising
Supreme Hero
blah, blah, blah
|
|
posted November 09, 2006 06:32 PM |
|
Edited by russ at 18:32, 09 Nov 2006.
|
btw, now that I think about it - most heroes have armorer and offense. Lets say both heroes have equal stats, armorer and offense. Who will do better?
Also, #1 combat spell is expert shield. It is so good that sometimes I have a very hard time deciding whether I should cast shield or implo my enemy's level 7. Who will do better if both heroes have expert armorer, offense and earth? (i.e. they both cast shield).
|
|
Ecoris
Promising
Supreme Hero
|
|
posted November 09, 2006 09:03 PM |
|
Edited by Ecoris at 21:06, 09 Nov 2006.
|
Damage calculation
I thought that the damage calculation formula was common knowledge.
Base damage (BD): #creatures * damage range (i.e. 10 beholders have a BD of 30-50)
Total damage: BD*(1+a+b+c...)*(1-q)(1-r)(1-s)... , where a,b,c,... are damage bonuses as decimal numbers, while q,r,s,... are damage reductions. Thus, anything that gives a bonus to damage is additive, but anything that gives a penalty is multiplicative. Taking the integer part may reduce damage calculated by this formula. Damage cannot be reduced to less than 1. (this paragraph has been quoted on several occasions by various members AFAIK and I have seen it on other websites).
The dread knight's double damage is just treated as a damage bonus of 100% (i.e. 1.00) in the formula. As one can see this means that it will at most result in 2x damage (namely when there are no other damage bonuses).
How exactly BD is calculated is unsure however. If we take the the 10 beholders as an example again BD is not calculated like described: A random value in the damage range is selected (i.e. 3, 4 or 5), then it is multiplied by the number of creatures in the stack (which would result in either 30, 40 or 50 points of base damage). At least I have tested that two beholders may have a base damage of ANY value from 6 to 10, not just the even ones. Perhaps a random value in the total damage range (30 - 50) is selected instead, but that's really not of great importance.
Offense vs. Armorer
I remember having read a long discussion about this in another thread. If there are no other damage bonuses/reductions than offense and armorer (assumed to be at expert level) the total damage is BD * 1.30 * 0.85 = 1.105 * BD. In other words, an increase of damage of about 10%. However things are not as simple as this because multiple damage bonuses will all be reduced by the total damage reduction. On the other hand multiple damage reductions lessen their total effect.
Lets take an example, why not 100 dread knights vs 100 naga queens for a change  .
Base damage is not relevant in this discussion only the total modifier (i.e. the term (1+a+b+c...)*(1-q)(1-r)(1-s)...).
1. Let us assume that the heroes have the same attack and defense. The dread knights will have a damage bonus of 25% and this is the only damage modifier.
2. Assume that both heroes have offense and armorer at expert level, the damage modifier (I'll call it M) is (1 + 0.30 + 5*0.05)*0.85 = 1.3175. The result of offense/armorer is +6.175 percentage points to damage compared to situation 1. (without the dread knights' higher attack skill it was +10.5%).
3. Same as 2. but now the dread knights deal "double damage". M = (1 + 1.00 + 0.30 + 0.25)*0.85 = 2.1675. Compared to situation 2. the damage has been increased by 65 % (not 100% as the description "double damage" suggests).
4. No double damage but expert shield cast on nagas.
M = 1.55 * 0.85 * 0.70 = 0.92225. Notice this. According to the descriptions the A/D gives +25%, offense gives +30%, armorer gives -15% and expert shield gives -30%. The unknowing player would guees that these modifiers would result in +10%, but because of the formula the base damage is actually reduced by 7.775%.
One could go on and think of even more examples, but the above ones show how that the resulting effect of the various damage modifiers can easily be misjudged.
You can read more about offense/amorer and stone skin/shield here:
Stone Skin vs Shield
How do the offence and armorer skills work exactly?
Do take care though as many wrong comclusions are put forth in these threads.
I can't find the last thread I was thinkig of (it's one of those "one pick: offense or armorer") but it's a good thread as well. IIRC a member posted something like "why not just have both".
|
|
tenzor
Adventuring Hero
|
|
posted November 09, 2006 10:35 PM |
|
|
Quote:
Interesting... you should put them up against Wolf Raiders sometime.
Double attack is not implemented yet But due to very high damage/hp ratio of Raiders and fact that they are faster than nagas, I think that it would be really better to have the dread knights on my side instead 
|
|
bloucester
Famous Hero
MSN can go to hell
|
|
posted November 10, 2006 04:10 PM |
|
|
well done, ecoris!
i was thinking of posting sth similar to what you have done... but yours is pretty good.
|
|
tenzor
Adventuring Hero
|
|
posted November 12, 2006 01:13 AM |
|
|
Quote:
speed means victory in higher number of somewhat equal units.
...
And therefore you would choose 100 zombies instead of 100 obsidian gargoyles as you wrote in one of your previsou posts. No comment 
|
|
bloucester
Famous Hero
MSN can go to hell
|
|
posted November 12, 2006 11:51 AM |
|
|
|
hahaha, i'm happy someone is reading my posts. you had to choose the other crappiest level 2 unit to as a counter argument. no matter, i go with the zombies this time. try 100 zombies vs 100 gargoyles in one on one combat. the gargoyles win! so speed means victory! however, i still prefer to have a hundred zombies in my town.
|
|
tenzor
Adventuring Hero
|
|
posted November 12, 2006 02:56 PM |
|
|
Quote:
hahaha, i'm happy someone is reading my posts. you had to choose the other crappiest level 2 unit to as a counter argument.
Of course, zombies
Quote:
no matter, i go with the zombies this time. try 100 zombies vs 100 gargoyles in one on one combat. the gargoyles win! so speed means victory! however, i still prefer to have a hundred zombies in my town.
Just FYI, according to simulation, gargoyles wins only 32,5% of battles vs zombies
But Heroes are not about 1 vs 1 stack combats. Zombies are usefull only in towns. Gargoyles are useful everywhere. They are fast, flying, defensive - good tactical units. You may use them for town defense as well as for blocking enemy archers, defending your owns etc. As Angelito mentioned, they are 1-hex wide (most of flyers are 2hex) so they can get everywhere. They cannot be ressurected but also sucked by vampire lords and this is really useful.
|
|
bloucester
Famous Hero
MSN can go to hell
|
|
posted November 13, 2006 11:00 AM |
|
|
|
well, tenzor, please tell me more about this simulation, because i really don't think that the gargoyles win only 30 % of the time. try the real thing - buy a hundered of them both and you do the combat with heroes with equal stats and abilities ( two lvl 1 clerics for example )
|
|
Ecoris
Promising
Supreme Hero
|
|
posted November 13, 2006 12:55 PM |
|
Edited by Ecoris at 12:58, 13 Nov 2006.
|
Could you make a new thread like "1 vs 1 combat tests" instead. This discussion is becoming off-topic. (I know that I've contributed to this myself).
Edit: On a second glance of the last few pages I see that it's been so for a quite a while.
|
|
bloucester
Famous Hero
MSN can go to hell
|
|
posted November 13, 2006 01:02 PM |
|
|
|
well, we are discussing tower units against other units...
|
|
tenzor
Adventuring Hero
|
|
posted November 14, 2006 08:55 AM |
|
|
Quote:
well, tenzor, please tell me more about this simulation, because i really don't think that the gargoyles win only 30 % of the time. try the real thing - buy a hundered of them both and you do the combat with heroes with equal stats and abilities ( two lvl 1 clerics for example )
The application uses the calculations from HOMMIII to simulate the 1 vs 1 combat. Best possible tactic and neutral morale and luck are used here. Best possible tactic means that faster unit attacks first in 1st round but slower has +2 to defense, then retaliates. In 2nd round, faster attacks, slower retaliatrs, slower attacks, faster retaliates etc. In case of equal speeds randomly choosen side attacks first in each round.
I can spent 1-2 hours by tracking 100 combat results in HOMM III but I don't consider this as a good idea. Simulation (if correct calculations are used there) will do 5 000 combats in one minute and the results are much more accurate than in case of 100 combats.
I can upload the simulation to web if you want (and open new thread ) so you will have a chance to test it
By the way, simple calculation of damage/enemy hp (=enemy units killed) in gargoyles vs. zombies combat:
gargoyles (1st round, against def +2): 12.5 (100*2.5/20)
zombies (1st round): 13.0625 (88*2.5*0.95/16)
gargoyles (2nd round): 20.9 (87*1.1*2.5/20 + 65*1.1*2.5/20))
zombies (2nd round): 22.5625 (2* 76*2.5*0.95/16)
So you can see even from 1st round calculations that the zombies will win more than 50% of combats.
|
|
angelito
Honorable
Undefeatable Hero
proud father of a princess
|
|
posted November 14, 2006 11:00 AM |
|
|
But this simulation doesn't show how it would work in "real" game. Gargs are faster, zombies can't get moral. So the gargs will always use the "wait...attack...fly away.....wait..attack.." tactic. And I really doubt the zombies will win ANY fight.
Because with your simulation, the sprites won't win any fight vs any unit I bet..
____________
Better judged by 12 than carried by 6.
|
|
bloucester
Famous Hero
MSN can go to hell
|
|
posted November 14, 2006 01:26 PM |
|
|
please upload that for me. not that i can't get myself, but i want to have your version.
meanwhile, consider this
malcom lvl 1 druid 0 attack 2 defense basic wisdom basic eagle eye
eleshar lvl 1 druid 0 attack 2 defense basic wisdom basic intelligence
terrain type - desert
100 zombies agaisnt 100 obsidian gargoyles
positon of both stacks - middle slot
battle
G - wait
Z - defend
G - go forward
G - wait
Z - defend
G - hit 11 z dead, 14 g dead
G - hit 10 z dead, 10 g dead
Z - hit 11 g dead, 9 z dead
G - hit 8 z dead, 10 g dead
Z - hit 9 g dead, 7 z dead
G - hit and...
and then gargoyles get morale bonus and everything goes to hell.
the zombies would have won this time, though.
so it's finally settled - obsidian gargoyles cannot even beat zombies, so they are the most useless lvl 2 unit.
you'd say - a gargoyle beat a marksman, but think of the marksman severely weakening enemy stacks from afar, before the hand to hand combat...
gargoyles suck
|
|
Ecoris
Promising
Supreme Hero
|
|
posted November 14, 2006 03:09 PM |
|
|
Quote:
The application uses the calculations from HOMMIII to simulate the 1 vs 1 combat. Best possible tactic and neutral morale and luck are used here. Best possible tactic means that faster unit attacks first in 1st round but slower has +2 to defense, then retaliates. In 2nd round, faster attacks, slower retaliatrs, slower attacks, faster retaliates etc. In case of equal speeds randomly choosen side attacks first in each round.
When a unit takes a defensive stance its defense value increases by a certain percentage value (I can't remember whether it's 20% or 30%) not just by 2. You should implement this. This also points out another problem: The hero's defense value is added in a "real" combat.
In case of a speed tie the same unit would go first in each and every round when they had started hitting each other.
I second Angelito's opinion that the gargs would deploy the strike-and-return tactic which is a great advantage.
Such tests can't prove that a unit is "the most useless lvl 2 unit"; tank beats flyer, flyer beats shooter, shooter beats tank, and so on.
|
| |
|
|
This thread is pages long: 
