RELIC wake up and fix damn RNG please!

ok Please Post New Topic "Tank armor too high Reduce it please"

On Balance Sub forum please

RNG not Bug form game but Form Bad luck Player

why cant you see the truth ?

Why can't you?

What Relic CAN fix is the number of hills that are unmarked on the map making it extremely hard to use at guns and assault guns. There is already penetration, missing, and LoS. I think there are enough obstacles in the tank warfare department already - there doesn't need to be elevation problems. Another person who posted about CoH3 also agreed with this point. Elevation is poorly done in the game. If it was supposed to be in, there should be CLEAR hill markings on the map.

-can you explain more about law of large numbers ? with example . i didn't get it

The law of large numbers is one of the basic laws of probability theory. What it states, is that if the series of independent random events is long enough (in thousands), then the distribution of results will closely resemble the probability distribution that was used to generate each random value. The larger the number of rolls, the closer these two distributions are going to be.

Mind, that it doesn't work for small numbers. So for example 20 rolls is definitely not enough for this to work. But at 2000 rolls, the distributions should be close enough to create a fair game. And thanks to automatch elo system, it even adds up between subsequent games.

So how does it work in coh2? Coh2 performs enourmous number of rolls each game. Each small arms shot is a rng roll, including every round shot by HMGs and other automatic weapons. In case of bigger guns, there are rolls for scatter, penetration and crits. This creates enough random events for the law of large numbers to work.

How does it effect players? Coh2 uses uniform distribution from 0 to 1 to create its random events, which means that after whole game, a histogram of all numbers rolled should have pretty much equal amount of numbers in all buckets. Here is a histogram showing an example for 10000 random numbers generated using exactly the same distribution as coh2 uses:

Show Spoiler
Becouse set of numbers rolled by each player resembles the initial distribution, then the distributions of numbers rolled by these players must be very close as well. What it means is that during the whole match, you and your opponent will get lucky pretty much equal amount of times. In effect, a better player will win in the end.

Rng is what breathes life into this game. This isn't starcraft or similar where the outcome is purely your actions. Personally I find those rts's too lifeless. Here in coh It's about choosing engagements and taking calculated risks. The skill in this game isn't just about tactical deployment but also analysing situations on the fly when unexpected results happen and testing your ability to adapt. No, it isn't always fair, sometimes the dice roll against you, but it's still your responsebility to deal with it, fight back and win regardless. In the long run as people pointed out it is fair since the dice will sometimes roll in your favor or you can take actions to mitigate the bad rolls. But the rng is callous, fair and unbiased. In my opinion this is what a wargame should be and why I love coh1/2 despite it's faults.

Most of the bad rng is removed from the game, and it's probably in its best state of all time when it comes to balance.

Instead of focusing on those two shots that missed the tank and blame rng gods you should rather be reflecting if there was anything else you could have done to secured the kill or larger egnagement. The beauty of coh is that outcomes aren't 100 % predictable despite having fought similar engagements a hundred times. If you can't get comfy with that then you should find a different game that plays more to the rules you like.

Most of the bad rng is removed from the game, and it's probably in its best state of all time when it comes to balance.

Instead of focusing on those two shots that missed the tank and blame rng gods you should rather be reflecting if there was anything else you could have done to secured the kill or larger egnagement. The beauty of coh is that outcomes aren't 100 % predictable despite having fought similar engagements a hundred times. If you can't get comfy with that then you should find a different game that plays more to the rules you like.

I can't emphasis this 2 statements even more.

A bad occurrence on a 3 flip coin case is rare, but not impossible to happen. You feel like shit when you roll that natural 1 (5%/10% as showned by Nosliw) but i'm sure you won't be complaining when you roll a nat20. The beauty of the game is that outcomes are unknown, but the result can be tilted in your favour through your skill. IF you think the result of a game is decided on a single dice roll, maybe you weren't even close to win that game or there were plenty of other scenarios or variants you are not taking into account which would had made that 50-50 roll a 99-1 in your favour.

This doesn't even show the complete picture of how bad "RNG" was a couple of years ago.

Examples
Show Spoiler

Bad RNG was: 1% death crit on vehicles. Explosive crits. Skill planes and every kamikaze plane tells a story. Abandoned vehicles. Flamer crits. Chasing snipers on retreat (target tables were not a thing). Ramming tanks. Really bad map balance with auto win spawn positions. Vehicle crits. Blizzard. Killing planes. Snares. B4. Old armor RNG bombs. Long, long etc.

I was gonna search for that cpt sprice game but instead i found this.
Show Spoiler

Insert BANEmemehere: "You think you got a bad RNG roll? You merely adopted RNG. We were born in it. We were mold by it. We didn't see stupid death crits removed till meme kamikazes planes were removed".


For a more serious response, watch this.

The law of large numbers is one of the basic laws of probability theory. What it states, is that if the series of independent random events is long enough (in thousands), then the distribution of results will closely resemble the probability distribution that was used to generate each random value. The larger the number of rolls, the closer these two distributions are going to be.

Mind, that it doesn't work for small numbers. So for example 20 rolls is definitely not enough for this to work. But at 2000 rolls, the distributions should be close enough to create a fair game. And thanks to automatch elo system, it even adds up between subsequent games.

So how does it work in coh2? Coh2 performs enourmous number of rolls each game. Each small arms shot is a rng roll, including every round shot by HMGs and other automatic weapons. In case of bigger guns, there are rolls for scatter, penetration and crits. This creates enough random events for the law of large numbers to work.

How does it effect players? Coh2 uses uniform distribution from 0 to 1 to create its random events, which means that after whole game, a histogram of all numbers rolled should have pretty much equal amount of numbers in all buckets. Here is a histogram showing an example for 10000 random numbers generated using exactly the same distribution as coh2 uses:

Show Spoiler

Becouse set of numbers rolled by each player resembles the initial distribution, then the distributions of numbers rolled by these players must be very close as well. What it means is that during the whole match, you and your opponent will get lucky pretty much equal amount of times. In effect, a better player will win in the end.

Thanks for explain

What Relic CAN fix is the number of hills that are unmarked on the map making it extremely hard to use at guns and assault guns. There is already penetration, missing, and LoS. I think there are enough obstacles in the tank warfare department already - there doesn't need to be elevation problems. Another person who posted about CoH3 also agreed with this point. Elevation is poorly done in the game. If it was supposed to be in, there should be CLEAR hill markings on the map.

Full agree and I am working on it.

Show Spoiler

RNG issue is more about timing and critical moment. There are moments when having bad RNG isn't critical at all and other when it does.
You don't really care on having bad RNG on a unit you wouldn't have kill even with RNG favors. But you do care of RNG when it is about what the OP describes.

BTW, what the OP describes as example wouldn't have occure on Coh1 thanks to defletion damage. Dice Rolling for 100% damage or 100% deflection isn't the best design choice and shows till today how hard it make for the balance team to set the cursor for each unit. Dice Rolling for let's say 50 to 70% of the damage but you still have 50 to 30% damage guaranteed is in my opinion better in terms of balance.

RNG issue is more about timing and critical moment. There are moments when having bad RNG isn't critical at all and other when it does.
You don't really care on having bad RNG on a unit you wouldn't have kill even with RNG favors. But you do care of RNG when it is about what the OP describes.

BTW, what the OP describes as example wouldn't have occure on Coh1 thanks to defletion damage. Dice Rolling for 100% damage or 100% deflection isn't the best design choice and shows till today how hard it make for the balance team to set the cursor for each unit. Dice Rolling for let's say 50 to 70% of the damage but you still have 50 to 30% damage guaranteed is in my opinion better in terms of balance.

That might or not be true because you end up having chip damage been more effective than intended when you achieve critical mass.

I don't think deflection damage should be something which applies to all sources but it could be specific to a certain type of unit (say heavies and AT guns). Problem is that you are adding another layer of variables in the balance box. BTW, this is not something which hasn't been tested in the past. IS2 original design had deflection been part of it's core, and that didn't work well for it.

I've always been in the camp of 'non-turret tanks should have deflection damage' to separate them from turreted ones..

I have the feeling it is reducing it. From my perspective it is much easier to balance Churchill tanks than IS2 or KT. Because it is less about %pen/armor but damage a unit can sponge before retreating. Once all AT units have a mínimum dps guarantee per hit, you can adjust their dps and lifepool, %Pen is just there to provide an extra damage, what is called critical hit in many games and can turn tides as RNG factor.

Now about the poke damage you are refering to, I don't see it as a bigger problem. a volley of 10 bazooka dealing flat out 30% of their damage (but with a really low chance to pen) or a volley of 3 (actual) bazooka over 10 that deal damage the target, the result is the same in term of damage.
On the first case, if one of them crit, its additional damage
On the second case, if a 4th crit, that the same amount of damage added.

What about the idea of certain units, probably assualt guns having the ability to switch ammo types. Say between AP for deflection damage and full damage on pen but medicore penetration stats and HEAT or APCBC for larger pen and full damage, but no deflection damage.

Full agree and I am working on it.

Thanks for the confirmation. I really appreciate it.