PRD and Weapon Swapping

Started 3 Sep 2019
by Cadebrennus
in Ask the Team
"Another example would be a card system, where all possible solutions are in a bag and every random roll draws a card making that particular card unavailable for subsequent draws until the bag is empty. We're using PRD for crits / procs and since this change the card system for defensive rolls"

Does this reset the RNG (the "card draw system" every time you swap from Main Hand to Two Hand and vice versa? Because myself and many other players have noticed a very high miss rate when swapping back and forth, not even considering bag swapping. I weapon swap A LOT during one fight (including bag swaps), even against mobs.
Tue 3 Sep 2019 11:26 AM by Sepplord
https://forum.playphoenix.online/viewtopic.php?f=38&t=11065&start=10#p84111

if i understand it right, this should be the answer to your question
Tue 3 Sep 2019 11:29 AM by gruenesschaf
No. However, due to streak avoidance when filling the bag and you obviously only swapping once you hit something, it's kind of more likely to then draw a defensive result.
The end result of streak avoidance is basically as if the bag was filled as a person would do it and not actual rng, ie if it were coin flips instead of having really long streaks as probability would predict, it will have much shorter streaks. I think at 50% the max possible streak is capped at 5.

Also PRD when refering to the different rng systems used here has nothing to do with defensive rolls and is not the card / bag system.
Tue 3 Sep 2019 2:02 PM by gotwqqd
Personally I think the bag should be refilled once it was lowered to some predetermined percentage...maybe 30-40%
Tue 3 Sep 2019 2:28 PM by Cadebrennus
gruenesschaf wrote:
Tue 3 Sep 2019 11:29 AM
No. However, due to streak avoidance when filling the bag and you obviously only swapping once you hit something, it's kind of more likely to then draw a defensive result.
The end result of streak avoidance is basically as if the bag was filled as a person would do it and not actual rng, ie if it were coin flips instead of having really long streaks as probability would predict, it will have much shorter streaks. I think at 50% the max possible streak is capped at 5.

Also PRD when refering to the different rng systems used here has nothing to do with defensive rolls and is not the card / bag system.

I thought a flat out miss was unaffected by the opponent's defense? As far as I know there's no skill (block/parry/evade) that affects miss rate. The only thing that comes close is Merc's Dirty Tricks but that procs a fumble, not a miss. In addition, taking into consideration that both parties have enchanted armor/weapons, I believe that many styles, dependent on their bonus to hit reduce or can flat-out eliminate miss chance.

During my crazy weapon swapping I see "you miss" messages far in excess of defense messages. This is on mobs and players. Next time I'm on I'll take screenshots and turn on my logging.

I don't swap a lot just for poisons and procs, I also swap a lot because I like the initial start of a chain to start with a cap speed weapon, since the GR sucks at the beginning of a chain. Then, I swap to the slowest possible weapons for the follow-ups which are typically 0.88-0.93 GRs, and they count much more for big hits, while swinging at cap speed due to the previous weapon. Once I finish a style chain (or it's interrupted in some manner) I go back to the cap speed weapon to start another chain. Basically, this is why I'm seeing A LOT of misses, because I swap nearly every swing.
Tue 3 Sep 2019 2:43 PM by gruenesschaf
Cadebrennus wrote:
Tue 3 Sep 2019 2:28 PM
I thought a flat out miss was unaffected by the opponent's defense? As far as I know there's no skill (block/parry/evade) that affects miss rate. The only thing that comes close is Merc's Dirty Tricks but that procs a fumble, not a miss. In addition, taking into consideration that both parties have enchanted armor/weapons, I believe that many styles, dependent on their bonus to hit reduce or can flat-out eliminate miss chance.

During my crazy weapon swapping I see "you miss" messages far in excess of defense messages. This is on mobs and players. Next time I'm on I'll take screenshots and turn on my logging.

I don't swap a lot just for poisons and procs, I also swap a lot because I like the initial start of a chain to start with a cap speed weapon, since the GR sucks at the beginning of a chain. Then, I swap to the slowest possible weapons for the follow-ups which are typically 0.88-0.93 GRs, and they count much more for big hits, while swinging at cap speed due to the previous weapon. Once I finish a style chain (or it's interrupted in some manner) I go back to the cap speed weapon to start another chain. Basically, this is why I'm seeing A LOT of misses, because I swap nearly every swing.

Assuming level 50 and both enchanted It's only affected by attacker weapon spec, enemy defensive bonus styles and your to hit bonus styles. It's just in the same single combat result roll as other defensive results like block, evade etc. and this roll uses the card / bag system
Tue 3 Sep 2019 9:25 PM by Cadebrennus
gruenesschaf wrote:
Tue 3 Sep 2019 2:43 PM
Cadebrennus wrote:
Tue 3 Sep 2019 2:28 PM
I thought a flat out miss was unaffected by the opponent's defense? As far as I know there's no skill (block/parry/evade) that affects miss rate. The only thing that comes close is Merc's Dirty Tricks but that procs a fumble, not a miss. In addition, taking into consideration that both parties have enchanted armor/weapons, I believe that many styles, dependent on their bonus to hit reduce or can flat-out eliminate miss chance.

During my crazy weapon swapping I see "you miss" messages far in excess of defense messages. This is on mobs and players. Next time I'm on I'll take screenshots and turn on my logging.

I don't swap a lot just for poisons and procs, I also swap a lot because I like the initial start of a chain to start with a cap speed weapon, since the GR sucks at the beginning of a chain. Then, I swap to the slowest possible weapons for the follow-ups which are typically 0.88-0.93 GRs, and they count much more for big hits, while swinging at cap speed due to the previous weapon. Once I finish a style chain (or it's interrupted in some manner) I go back to the cap speed weapon to start another chain. Basically, this is why I'm seeing A LOT of misses, because I swap nearly every swing.

Assuming level 50 and both enchanted It's only affected by attacker weapon spec, enemy defensive bonus styles and your to hit bonus styles. It's just in the same single combat result roll as other defensive results like block, evade etc. and this roll uses the card / bag system

Why the excessive misses then vs enemies that aren't using defensive styles such as fleeing player Casters or mobs?
Wed 4 Sep 2019 11:50 AM by gruenesschaf
Cadebrennus wrote:
Tue 3 Sep 2019 9:25 PM
Why the excessive misses then vs enemies that aren't using defensive styles such as fleeing player Casters or mobs?

Don't sum up main and offhand misses, as all kinds of streak avoidance the card / bag approach handles main and offhand separately, ie there is one bag for mh attacks and one for offhand attacks.


Here a test output from when we switched over to the combined table / bag approach for defensive rolls.

Separate = individual rolls, not doing the rolls further down when a defensive roll succeeds (evade roll succeeds -> bail out as evade, no block / parry / miss whatever check follows)
Combined = All chances squashed into a single roll

RNG = .net default rng
PRD = https://dota2.gamepedia.com/Random_distribution with failure counter on the attacker per weapon hand
Table = the bag / card system, drawn by / on the attacker per weapon hand

First percentage is the actual percentage of the total hit attempts
Second percentage is the percentage from all hits "reaching" that stage (for hits it would be total swings - all defensive results, for miss it would be total swings - all defensive results except miss etc, up the check order) and should pretty much be the nominal chance.

Streaks show the longest streaks of a given result, defensive means all defensive types combined (not the sum of defensive streaks but the longest defensive streaks consisting of any defensive result, including miss)

Separate RNG is "the default" and what it was in alpha, Separate Table is what we had for a while and Combined Table is what we have since sometime before launch.

This is with a nominal 34% evade chance and 14% miss.

Separate, RNG
- 0/1000000 guard blocks, 0%, 0%
- 0/1000000 dashing blocks, 0%, 0%
- 0/1000000 dashing parries, 0%, 0%
- 338255/1000000 evades, 33.83%, 33.83%
- 0/1000000 parries, 0%, 0%
- 0/1000000 blocks, 0%, 0%
- 92567/1000000 misses, 9.26%, 13.99%
- 569178/1000000 hits, 56.92%, 100%
- Streaks - Hit: 21 | Defensive: 15 | DB: 0 | DP: 0 | Guard: 0 | Evade: 12 | Parry: 0 | Block: 0 | Miss: 5

Separate, PRD
- 0/1000000 guard blocks, 0%, 0%
- 0/1000000 dashing blocks, 0%, 0%
- 0/1000000 dashing parries, 0%, 0%
- 339919/1000000 evades, 33.99%, 33.99%
- 0/1000000 parries, 0%, 0%
- 0/1000000 blocks, 0%, 0%
- 92145/1000000 misses, 9.21%, 13.96%
- 567936/1000000 hits, 56.79%, 100%
- Streaks - Hit: 6 | Defensive: 9 | DB: 0 | DP: 0 | Guard: 0 | Evade: 7 | Parry: 0 | Block: 0 | Miss: 3

Separate, Table
- 0/1000000 guard blocks, 0%, 0%
- 0/1000000 dashing blocks, 0%, 0%
- 0/1000000 dashing parries, 0%, 0%
- 340000/1000000 evades, 34%, 34%
- 0/1000000 parries, 0%, 0%
- 0/1000000 blocks, 0%, 0%
- 92400/1000000 misses, 9.24%, 14%
- 567600/1000000 hits, 56.76%, 100%
- Streaks - Hit: 9 | Defensive: 5 | DB: 0 | DP: 0 | Guard: 0 | Evade: 3 | Parry: 0 | Block: 0 | Miss: 1

Combined, RNG
- 0/1000000 guard blocks, 0%, 0%
- 0/1000000 dashing blocks, 0%, 0%
- 0/1000000 dashing parries, 0%, 0%
- 339414/1000000 evades, 33.94%, 33.94%
- 0/1000000 parries, 0%, 0%
- 0/1000000 blocks, 0%, 0%
- 91554/1000000 misses, 9.16%, 13.86%
- 569032/1000000 hits, 56.9%, 100%
- Streaks - Hit: 30 | Defensive: 14 | DB: 0 | DP: 0 | Guard: 0 | Evade: 11 | Parry: 0 | Block: 0 | Miss: 6

Combined, Table
- 0/1000000 guard blocks, 0%, 0%
- 0/1000000 dashing blocks, 0%, 0%
- 0/1000000 dashing parries, 0%, 0%
- 340000/1000000 evades, 34%, 34%
- 0/1000000 parries, 0%, 0%
- 0/1000000 blocks, 0%, 0%
- 92000/1000000 misses, 9.2%, 13.94%
- 568000/1000000 hits, 56.8%, 100%
- Streaks - Hit: 5 | Defensive: 4 | DB: 0 | DP: 0 | Guard: 0 | Evade: 3 | Parry: 0 | Block: 0 | Miss: 2

Combined, PRD
- 0/1000000 guard blocks, 0%, 0%
- 0/1000000 dashing blocks, 0%, 0%
- 0/1000000 dashing parries, 0%, 0%
- 340951/1000000 evades, 34.1%, 34.1%
- 0/1000000 parries, 0%, 0%
- 0/1000000 blocks, 0%, 0%
- 91702/1000000 misses, 9.17%, 13.91%
- 567347/1000000 hits, 56.73%, 100%
- Streaks - Hit: 4 | Defensive: 15 | DB: 0 | DP: 0 | Guard: 0 | Evade: 8 | Parry: 0 | Block: 0 | Miss: 4
Wed 4 Sep 2019 3:05 PM by Cadebrennus
gruenesschaf wrote:
Wed 4 Sep 2019 11:50 AM
Cadebrennus wrote:
Tue 3 Sep 2019 9:25 PM
Why the excessive misses then vs enemies that aren't using defensive styles such as fleeing player Casters or mobs?

Don't sum up main and offhand misses, as all kinds of streak avoidance the card / bag approach handles main and offhand separately, ie there is one bag for mh attacks and one for offhand attacks.



Perhaps I wasn't clear in my earlier explanation. I am not accounting for offhand, nor does it matter for weapon swapping MH. I was talking about a simple switch (not bag switch) from MH to 2H and back from 2H to MH. Even that swap makes the "new" swing miss much more often than just staying with a weapon and not swapping from MH->2H and vice versa.
This topic is locked and you can't reply.

Return to Ask the Team or the latest topics