| Author | Caravan Problem |
|---|
https://www.lordswm.com/warlog.php?warid=487517622
sigh...........i felt like the demon had 4 morale.............hell horses feel even worse than cave demons |
i found the caravan's probability of triggering a special ability is much higher than players. feels like cheating. stun, knock shooting, knock back, and fear attack are just one after another.
another annoying thing is dealing with apparition. when i ambushed a caravan of necro using DE, i found i had 90% chance missing the fist hit to apparition and >50% to miss the first two hits. when i use necro to ambush, i found my apparition has great chance to be hit by the first 2 consecutive attack, and seems follow the formula of 2 hits -> 1 miss -> 2 hits. |
#82 yea it does feel like you're on the short end of the stick........
https://www.lordswm.com/warlog.php?warid=477770547
i think this special ability triggering has been around for a long time............this was what happened when i was level 7..........my swordsmen NEVER got a turn off and only retaliated once i think |
I found two of my caravans with very similar troops. I lost vs them.
first one:
https://www.lordswm.com/warlog.php?warid=487405261<=-1
cyclops: 48
orc chiefs: 21
orcs : 24
ogre magie : 32
ogres: 8
hob gobs : 54
gobs: 4
second one:
https://www.lordswm.com/warlog.php?warid=487462835<=-1
cyclops: 44
orc chiefs: 28
orcs : 44
ogre magie : 10
ogres: 18
hob gobs : 78
gobs: 10
differences between the caravans :
cyclops: -4
orc chiefs: -7
orcs : +20
ogre magie : -22
ogres: +10
hob gobs : +24
gobs: +6
From my point of view, this is a decrease in difficulty.
I'll try to find some caravans, with even closer stacks, but it won't be easy. |
#80
I know it is probably not 7% anymore - but also the decrease is bigger than 5%. Can't figure yet the exact numbers but it feels like win -> 15% increase and loss -> 10-12% decrease for me. |
#82, #83
† Triggering special abilities may trigger during any attack. The probability of triggering depends on HP of damager and target.
If DHP ≥ THP, then
Probability = 25% + 3% * (DHP/THP);
else
Probability = 25% - 3% * (THP/DHP)
https://www.lordswm.com/help.php?section=32 |
| interesting............so in the demon battle there was a 37% chance of fear attack on my monks on their first attack, by the way is the probability counted after damage is counted or before the attack? |
Before the attack, but here's the rest of the details I didn't quote:
where:
» DHP is the total remaining HP of the damaging stack;
» THP is the total remaining HP of the target stack;
» Trigger probability lays within the interval of [5%;75%];
» The moments to calculate the probability differ depending on the ability type.
-For Knocking shot and Stun, the probability is calculated using HP's of stacks before the actual attack.
-For Blinding brilliance, Fear attack, Knockback, Mayhem, Shadow whip and Thunderbolt, the calculation occurs after the attack and before target's retaliation;
» Blinding brilliance, Knockback, Mayhem, Shadow whip and Thunderbolt can also trigger on retaliation strikes. If blinding brilliance triggers on retaliation, it interrupts any attacker's actions supposed to occur after their strike (e.g. such abilities as Strike and return, Triple strike etc.) |
| I meant after the attack before retal. |
85
I don't understand why you keep saying these your-imagination-rates. If so, why people feel it gets harder, and losing a lot continuously? it is stupid. |
@90:
Because not everyone has the same feel. i.e, my post #84.
The problem may be that players are mixing up their feels with what a "feel-less" algorithm is processing.
Here is a fictive exemple.
As a DE with shrews and assaillants, you meet a wizard caravan, with an average number of grems, magis and golems.
ex1:you lost to it. Next time you meet very few golems but a lot of grems/magis. Statisticaly, it is weaker. But the reduction was all on golems. Since there is less golems then what there should be, you end up with more grems/ magis than before.
ex2:you win it. Next time, you meet a very huge amout of golems, but fewer grems/magis. It is statisticaly stronger, but as DE it will be very easy.
This is something DE will experience the most, because the shrew+assaillants combo can shred an incredible amout of slow melee units.
Killing 500 modern golems as a DE is a joke. Killing them with say necro is very hard.
This is something the algorithm is probably not taking into account, leading to the previous complains. |
91:
Im with you. Formula should consider faction, so faction without no retaliation troops like knight or elfs would have bigger chances.
But now back to changes: I have a proof admins changed formula at least for DE, I will show it on my previous example.
1. necro caravan:
5372 Skeletal Bowmen
190 Infected Zombies
302 Zombies
I won and next was:
2. necro caravan:
5524 Skeletal Bowmen
69 Infected Zombies
504 Zombies
I won again and next was:
3. necro caravan:
7292 Skeletal Bowmen
247 Infected Zombies
278 Zombies
Raise between caravan 2 and 1 is significantly lower than between 3 and 2. (I would even say that caravan 2 was easier than caravan 1:)
I have one more example: a row of won wizard caravans where I dont see significant raise, so I won even 5 wizard caravans of same type in a row and 1st and last have almost the same difficulty. You can compare
1. wizard caravan
49 Genies
52 Lorekeepers
25 Magi
1250 Golems
2. wizard caravan
43 Giants
59 Lorekeepers
23 Magi
1332 Gargoyles
3. wizard caravan
70 Genies
0 Lorekeepers
182 Magi
1461 Enchanted Gargoyles
4. wizard caravan
64 Senior Genies
0 Lorekeepers
102 Magi
1708 Enchanted Gargoyles
5. wizard caravan
44 Senior Genies
38 Lorekeepers
48 Magi
926 Modern Golems
I won all of them. |
According to TG formula:
180 Vampire counts are much more stronger than
(243 Vampires + 311 Ghosts + 70 Apparitions + 82 Infected zombies + 186 Zombies)
Well, this doesnt seem real:) |
Lets do some theorycraft about those 3 necro caravans in your post #92.
Assuming that the tg formula has been changed at the same time as the AI, it all happened the 14/02.
the 2 first necro caravans have been encoutered on the 13/02. Which means the small gap between 1 and 2 is still following th 7% rule, but the big gap between 2 & 3 is with the new one.
In my opinion, there is too much dynamic balance between 1&2 or 2&3 to conclude anything.
What i find interesting is the difference between 1&3.
Troop count in 3 compared to 1:
infected zombies: 130%
zombies: 92%
skellies: 135%
The zombies pose a little problem. But i would estimate the average increase to about 30%.
Now if we substract the 7% increase from 1 to 2, we end up with a ~21% increase with the new rule. Three times more than the old one.
Remember this is a bunch of theory coupled with estimations and rounding. Nothing anywhere close to accuracy. |
94
Lets do some theorycraft about those 3 necro caravans in your post #92.
Better analyze gap between 1&2 for dynamic balance.
Assuming that (quite obvious assumption):
1. 1&2 are before the change
2. The increase difficulty was 7% (before the change)
3. The number of stack of a type aren't considered in the dynamic balancing
4. In dynamic balancing IZ (tier 2 upgrade) > zombie (tier 2 basic) > skellis (tier 1 upgrade)
Since skellies are 103%, Infected 36% and zombies 167% than:
- an infected is the equivalent of 115% zombies (it can be as high as 130% but non above)
- a zombie is 1,000% a skellie (it can be as low as 800% but non much less)
The above value are the result of some tries in excel (total increase difficulty 7%).
If I apply those value to 2&3 I get an increased difficulty of 14% (changing slightly the above value it can go up to +25%).
So assuming that:
A) dynamic balancing work in the same way before and after the change
B) caravan difficulty wasn't reseted with the change
Difficulty increase if you win has doubled (or even more).
I think that assumption A is WRONG!
Because:
programming an easy way to store each player current difficulty for each type of caravan I would store a single number:
army strength (as used in rage and bloodlust special abilities)
the change to the IA seems to include:
- special tread level for low number stack (1-2 creature stack, retail takers)
- reevaluation of tread level of stack (probably by change in creature's strength)
- different option for blocked shooters (walk away)
- use of non spell for casters when a spell would be overkill (both hero in caravan and spell caster creature)
So what actually happened is that each thief has reached a certain strength level for his/her caravans BUT the strength values of many creatures were changed with HUGE variation in actual numbers of creature.
Since many creature's strength were lowered (assumption) this result in a general increase of creature number. |
Better analyze gap between 1&2 for dynamic balance.
I avoided this because it makes no sence.
Since we have no clue of what the real value of each creature is, and neither do we have a clue on what stack actually got a raise or decrease in 1&2 due to a lot of troops disparity, each of the creature value span is infinite. This leads to an infinite raise% span between 2&3.
My way of doing it is not accurate at all, but this is not much more than a random guess.
So assuming that:
A) dynamic balancing work in the same way before and after the change
B) caravan difficulty wasn't reseted with the change
Difficulty increase if you win has doubled (or even more).
I think that assumption A is WRONG!
Because:
programming an easy way to store each player current difficulty for each type of caravan I would store a single number:
army strength (as used in rage and bloodlust special abilities)
I hardly see how this statement contradicts A, since it probably always worked this way, before and after the various tg changes.
So what actually happened is that each thief has reached a certain strength level for his/her caravans BUT the strength values of many creatures were changed with HUGE variation in actual numbers of creature.
Now this is interesting, and could be spotted with caravan 4&5 of that same type. |
I avoided this because it makes no sense.
It seems to make no sense, but since there are only 3 type of creatures moreover only ONE type of creatures has an increase of about 7%. So with my hypothesis (1-4) you can actually compute the relative strength of the three creature's type (not exactly, but near).
My way of doing it is not accurate at all, but this is not much more than a random guess.
This way doesn't guarantee accuracy BUT it's useful to compute a "new" range of difficulty increase in case you won. Tinkering with the possible value I found out that after the change (with my assumption) the increase difficulty in example 2&3 is 14% or above but even if you change your coefficient it doesn't go less than 12% but it can go over 20%.
The point was to express assumption A) with I think is false.
I hardly see how this statement contradicts A, since it probably always worked this way, before and after the various tg changes.
Yes and No :)
What happened in my "conclusion" is that the program for computing a combination of troop for a specific caravan works in the same way BUT the coefficient used in the program (=strength of creatures) has changed SO the dynamic balancing algorithm has changed (since this include the coefficient)!
Let me explain with an example.
Suppose you want to understand if a caravan difficulty increased and by how muck and you have only 2 type of troops (I'm simplifying): Skellies and Zombie
- From previous caravan you have found out that dynamic balancing exchange 1 zombie for 2 skellies. This is independent of the fact that the server store a single number for caravan strength or keep single creature number
- The change done by admin (only for example sake) while leaving zombie strength equal to the strength of two skellies BUT both get -5% (example, because judged too slow in ambush)
- Confronting two caravans after the change still 2 skellies = 1 zombie BUT the first caravan after the change will have 5/95 "more" troops.
The problem is that only confronting two caravans' strength (same type) one before and one after the change can the changed be analyzed.
Any analysis of caravan's strength increase with both before (or both after) the change will mostly conclude that caravan strength has simply jumped up. Dynamic balancing with many creature will hide the actual change and if most creature have only a slight strength adjustment it will make the analysis much more difficult.
If I'm right:
after 50-100 ambushes (enough to encounter all or mostly all type) the apparent strength increase of caravan if you win the previous will return to 7% (and loosing to -5%).
Theories are useless if they don't predict the future, ask a physics :D |
for triplebuzze:
In post 91, you make a very good point. DE wants single big stacks of melee troops that they can hack down using shrew and lizards. Other factions like elf or knight look at 500 modern golems in one stack and start to cry because it's impossible to kill with efk/guards. The algorithm does not take this into account and treats everyone the same way, ignoring faction strengths and weaknesses. |
I have one more example, which is more simpler
1. knight caravan:
429 Monks
637 Recruits
217 Farmers
I won and next was:
2. knight caravan:
435 Monks
737 Recruits
258 Farmers
I won again and next was:
3. knight caravan:
530 Monks
744 Recruits
1476 Farmers |
to vidlak666
What's the time of these three battles. I post this because there's a significant increase in caravan strength for DE players after valentine's update. If the first two was before 2-14 and the third one was after 2-14, it can be explained. |
