Interrupt Build?

[Guest]

lol @ base 15 interrupting blows

 

I almost feel like we should have a wager going at this point.

[mazeltov]

I'd take it. Wouldn't make that much difference either way. If it worked by improving the total roll by 15%, then an average roll of 92 would still get something like 14 points from I.B.. Low rolls would be penalized more by the talent. Rolling high is already way overkill, so at least with +15 integer low rolls would be helped by the talent.

 

Though, as I suggested above, it's not that fixed I.B. would be overkill, but that the Interrupt mechanic itself is already overkill if specifically built for.

Edited April 18, 2015 by mazeltov

[Ruminate]

Take note of the following combat log:

 

I managed to get 8 hits off inbetween each bear hit when all of my interrupts are 100%. I attack approximately twice per second.

 

Now, take a look at this combat log:

 

 

I managed to get 6 hits off inbetween each bear hit when my interrupt chance is 0.

 

Does that sound good to anyone? If yes, then what would you be sacrificing to go from 0 interrupt chance to 100 interrupt chance?

[Guest]

hehe, I guess I'm more interested in the fact that the bear never hit you in the first example whereas he knocked 15% off your life with a graze in the second

 

EDIT: Also keep in mind that the "hit-per" metric that you're looking at here is going to be influenced by the interrupt rating of the weapon you're using. Flail is listed as "weak" (30, which is half the morning star and 60% of the great sword, estoc, etc).

 

@mazeltov: I just tested the phrase, Thick Grew Their Tongues, Stumbling O'er Words using Kana and it works as described. Combined with the Level 1 Frighten phrase, it dropped the enemy Concentration by 16 points (-10 base Con, -6 Con from the Resolve hit). Not sure why it's working for me, but I thought I'd let you know.

 

2ND EDIT: Just tested the wizard spell Expose Vulnerabilities and that worked too. Ninja fix in 1.04 maybe?

Edited April 18, 2015 by Achilles

[Ruminate]

hehe, I guess I'm more interested in the fact that the bear never hit you in the first example whereas he knocked 15% off your life with a graze in the second

 

I'm not sure why you're paying attention to the graze. The only thing relevant is the number of hits inbetween the bear's attacks.

If I set it up so that the bear has a 100% chance to miss, the same relevant scenario will occur: 6 hits without interrupts vs. 8 hits with full interrupts

[Guest]

Wrong, per the edit I made above.

 

And I still think you're missing point of interrupt: the significance is that he never hit you. On a long enough timeline, that bear would have died without ever hitting you. I've been repeating that for pages now.

[Ganrich]

If IB is a percentage then it won't be worth a Talent point IMHO.  If IB is a flat number then it could be a bit OP.  This is once it is fixed of course.  

 

@Ruminate -  I think a better example of whether Interrupt is worth it is giving both examples a set time (10 seconds or so) to see how many times the Bear gets an attack in that period of time.  Sure, number of attacks between each bear attack is relevant, but I only see 2 bear attacks in the first encounter, and I see 4 in the second.  1v1 interrupt isn't going to matter much anyway, but a party of interrupters that mix weak interrupts (.35 and .5 second) with strong interrupts (1 second) will have a more drastic outcome.  If you had a party member with a mourning star get an interrupt on that bear how many attacks would that barbarian land before the Bear attacked again?  If in 10 seconds the bear gets 3 attacks because interrupts slowed him, but in the second example got 4 then that is 1 less chance the bear had to do damage.  Number of attacks in between isn't everything.  

 

IF an interrupt party can cut the incoming damage by 1/3 to 1/2 via auto attack (EG using no per encounter abilities)... then that is silly good IMHO.  

[mazeltov]

@mazeltov: I just tested the phrase, Thick Grew Their Tongues, Stumbling O'er Words using Kana and it works as described. Combined with the Level 1 Frighten phrase, it dropped the enemy Concentration by 16 points (-10 base Con, -6 Con from the Resolve hit). Not sure why it's working for me, but I thought I'd let you know.

 

I just tried the Phrase again while attacking some villagers in The Black Hound, and can't replicate your success (they stay at 78 Concentration with the Phrase debuff applied). Thank you for testing it. Can you try Expose Vulnerabilities and Insect Swarm/Plague?

[Guest]

Yeah, posted a 2nd edit with the result of Expose Vulnerabilities. It works.

 

That phrase only lasts a few seconds. Are you sure your getting good before and after comparisons? Not doubting you, just trying to help.

 

Before:

After: 

Edited April 18, 2015 by Achilles

[Ruminate]

Wrong, per the edit I made above.

 

And I still think you're missing point of interrupt: the significance is that he never hit you. On a long enough timeline, that bear would have died without ever hitting you. I've been repeating that for pages now.

 

Why would a long enough timeline matter? Interrupts in Pillars of Eternity are not the same as interrupts in other games. An interrupt in PoE does not reset the attack, it only delays it.

 

Here is the same test with a morning star, naked, with +27% attack speed:

 

 

11 hits inbetween 4 bear hits at 100% interrupt chance.

 

 

7 hits inbetween 4 bear hits at 0% interrupt chance.

 

11 hits vs. 7 hits inbetween 4 bear hits. I'll ask again: does that sound good to you? What would you sacrifice in order to go from 0 interrupt chance to 100%?

[Guest]

Yes, it absolutely does for reasons I've been consistently providing for pages now. Happy?

 

Long enough timeline matters because...as you've clearly shown...you can (despite your claim) tie up an enemy until it is dead. It doesn't matter where or not it resets: indefinite delay is the same. damn. thing.

 

"This game doesn't do precisely the thing I'm showing it do in this screenshot". Really?!

 

EDIT: Oh, btw, I'm not sure what you're doing with the cheat codes to keep the bear from hitting you, but turn it off and post scenario number 2 again.

Edited April 18, 2015 by Achilles

[Infares]

hehe, I guess I'm more interested in the fact that the bear never hit you in the first example whereas he knocked 15% off your life with a graze in the second

 

EDIT: Also keep in mind that the "hit-per" metric that you're looking at here is going to be influenced by the interrupt rating of the weapon you're using. Flail is listed as "weak" (30, which is half the morning star and 60% of the great sword, estoc, etc).

 

@mazeltov: I just tested the phrase, Thick Grew Their Tongues, Stumbling O'er Words using Kana and it works as described. Combined with the Level 1 Frighten phrase, it dropped the enemy Concentration by 16 points (-10 base Con, -6 Con from the Resolve hit). Not sure why it's working for me, but I thought I'd let you know.

 

2ND EDIT: Just tested the wizard spell Expose Vulnerabilities and that worked too. Ninja fix in 1.04 maybe?

That, and an "interrupt build" tends to reference your whole team rather than simply a single member. You get two barbs and two blast wizards vandalizing a whole bigass clump of melees and you're the only one hitting people.

[Infares]

 

@mazeltov: I just tested the phrase, Thick Grew Their Tongues, Stumbling O'er Words using Kana and it works as described. Combined with the Level 1 Frighten phrase, it dropped the enemy Concentration by 16 points (-10 base Con, -6 Con from the Resolve hit). Not sure why it's working for me, but I thought I'd let you know.

 

I just tried the Phrase again while attacking some villagers in The Black Hound, and can't replicate your success (they stay at 78 Concentration with the Phrase debuff applied). Thank you for testing it. Can you try Expose Vulnerabilities and Insect Swarm/Plague?

Are you confirming that the debuff shows on the enemy's popup? Some chanter phrases have small aoe and if they start before you're in range they won't apply to a particular enemy during its duration.

[KDubya]

 

Wrong, per the edit I made above.

 

And I still think you're missing point of interrupt: the significance is that he never hit you. On a long enough timeline, that bear would have died without ever hitting you. I've been repeating that for pages now.

 

Why would a long enough timeline matter? Interrupts in Pillars of Eternity are not the same as interrupts in other games. An interrupt in PoE does not reset the attack, it only delays it.

 

Here is the same test with a morning star, naked, with +27% attack speed:

 

 

11 hits inbetween 4 bear hits at 100% interrupt chance.

 

 

7 hits inbetween 4 bear hits at 0% interrupt chance.

 

11 hits vs. 7 hits inbetween 4 bear hits. I'll ask again: does that sound good to you? What would you sacrifice in order to go from 0 interrupt chance to 100%?

 

 

 

In your test is how did you adjust/control your interrupt chance?

 

Are they such that in the first test every hit has to interrupt and in the second test every attack can not interrupt?

 

To compare the two scenarios you'd need to go until a set number of hits occur in both cases, as in count the number of bear attacks you receive before you deiver ten hits on the bear or whatever number you want but it has to be the same for both cases.

 

In your test if the bear is killed with 11 hits, then the bear is still alive and attacking in the seven hit case.

[Ruminate]

Yes, it absolutely does for reasons I've been consistently providing for pages now. Happy?

 

Long enough timeline matters because...as you've clearly shown...you can (despite your claim) tie up an enemy until it is dead. It doesn't matter where or not it resets: indefinite delay is the same. damn. thing.

 

"This game doesn't do precisely the thing I'm showing it do in this screenshot". Really?!

 

EDIT: Oh, btw, I'm not sure what you're doing with the cheat codes to keep the bear from hitting you, but turn it off and post scenario number 2 again.

 

I have 3 might, 18 constitution, 100 perception, 100 resolve on one screenshot.

I have 3 might, 18 constitution, 3 perception, and 100 resolve in the other.

If I normalized my stats, I wouldn't be interrupting 100% of the time. Not even half the time, or even a quarter of the time.

A normalized stat spread wouldn't show the difference between 0% interrupts and 100% interrupts.

 

But if it makes you happy, here is a test using "normalized" stats.

 

This first screenshot shows a stat spread of 19 might, 19 dexterity, and 10s in everything else:

 

 

5 hits and 4 grazes from the morning star. 9 attacks until the bear dies.

 

This next screenshot shows a stat spread of 19 dexterity, 19 perception, and 10s in everything else:

 

 

5 hits, 6 grazes, and 2 misses from the morning star. 13 total attacks, but it wasn't enough to kill the bear.

 

The only thing you can extrapolate is the effect might has on grazes. Might lessens the impact of grazing hits because a graze does not multiply your damage by 0.5. Grazes subtract 50% from your damage bonus. Going from 100% to 50% is a 50% reduction, but going from 150% to 100% is only a 33% reduction. Throw in some DR and its why I was grazing for 6 damage in one screenshot, but grazing for 11 in the other. 11 is almost double of 6, even though 19 might only adds 27% bonus damage.

 

But this thread isnt about how important might may or may not be. Its about how important interrupt may or may not be.

Edited April 18, 2015 by Ruminate

[Guest]

Here's the test I want you to run:

Scenario 1 - Interrupt 100% of the time

Scenario 2 - Interrupt 0% of the time

 

 

I extrapolate that the bear will die in Scenario 1 and that you will die in Scenario 2

 

[KDubya]

Ruminate -

 

Your method of guaranteeing either interrupts or not looks like it'd work fine. As you say we want to see the effect of interrupts not the randomness of damage.

 

Do grazes interrupt for less time than hits and crits?

 

If grazes and crits have different durations on the interrupt, you could add enough accuracy to guarantee a crit and run the test until a set number of hits is obtained. This would give you the absolute best case, perfect world scenario of every hit is a crit and interrupts. That could then be compared to a 0% interrupt scenario to see the absolute best difference between the two.

[Guest]

KDubya, I disagree. The point of an experiment is to gauge the effect of changing one variable to see how it performs against a control. He has no control and he changes 6 variables in each run. If he want to determine the effectiveness of interrupt, he needs to change interrupt (and nothing else).

 

The attack roll only determines whether or not an interrupt occurs. I think there was mention of a penalty for a graze, and 25 is added to the roll for a crit. Otherwise it is only the base interrupt (determined by Perception) + d100.

 

Duration is determined by the interrupt rating of the weapon (0 for blunderbusses, 60 for morning stars).

[Ruminate]

Here's the test I want you to run:

Scenario 1 - Interrupt 100% of the time

Scenario 2 - Interrupt 0% of the time

 

 

I extrapolate that the bear will die in Scenario 1 and that you will die in Scenario 2

 

 

How can you extrapolate that i will die in scenario 2 when I already gave you a scenario above where I killed the bear with 19 might and 0 interrupts? You're not making any sense.

 

 

Ruminate -

 

Your method of guaranteeing either interrupts or not looks like it'd work fine. As you say we want to see the effect of interrupts not the randomness of damage.

 

Do grazes interrupt for less time than hits and crits?

 

If grazes and crits have different durations on the interrupt, you could add enough accuracy to guarantee a crit and run the test until a set number of hits is obtained. This would give you the absolute best case, perfect world scenario of every hit is a crit and interrupts. That could then be compared to a 0% interrupt scenario to see the absolute best difference between the two.

 

 

Thats gonna be difficult to test, especially since a string of crits, even at 3 might, will floor the bear. But I'll try to think of something.

[Ruminate]

KDubya, I disagree. The point of an experiment is to gauge the effect of changing one variable to see how it performs against a control. He has no control and he changes 6 variables in each run. If he want to determine the effectiveness of interrupt, he needs to change interrupt (and nothing else).

 

The attack roll only determines whether or not an interrupt occurs. I think there was mention of a penalty for a graze, and 25 is added to the roll for a crit. Otherwise it is only the base interrupt (determined by Perception) + d100.

 

Duration is determined by the interrupt rating of the weapon (0 for blunderbusses, 60 for morning stars).

 

So I should do what you did on page 2? Take a lone combat log with 2 interrupts, then tell everyone how important interrupts are because you assume the spider would have gotten a hit if you didn't interrupt?

[Guest]

 

Here's the test I want you to run:

Scenario 1 - Interrupt 100% of the time

Scenario 2 - Interrupt 0% of the time

 

 

I extrapolate that the bear will die in Scenario 1 and that you will die in Scenario 2

 

 

How can you extrapolate that i will die in scenario 2 when I already gave you a scenario above where I killed the bear with 19 might and 0 interrupts? You're not making any sense.

Because you moved the goalpost of your own experiment. You changed all of your stats in between rounds, therefore you aren't testing the effectiveness of one thing. 

 

It's called putting your thumb on the scale.

[Guest]

 

KDubya, I disagree. The point of an experiment is to gauge the effect of changing one variable to see how it performs against a control. He has no control and he changes 6 variables in each run. If he want to determine the effectiveness of interrupt, he needs to change interrupt (and nothing else).

 

The attack roll only determines whether or not an interrupt occurs. I think there was mention of a penalty for a graze, and 25 is added to the roll for a crit. Otherwise it is only the base interrupt (determined by Perception) + d100.

 

Duration is determined by the interrupt rating of the weapon (0 for blunderbusses, 60 for morning stars).

 

So I should do what you did on page 2? Take a lone combat log with 2 interrupts, then tell everyone how important interrupts are because you assume the spider would have gotten a hit if you didn't interrupt?

Non sequitur. Go look it up.

[mazeltov]

2ND EDIT: Just tested the wizard spell Expose Vulnerabilities and that worked too. Ninja fix in 1.04 maybe?

 

1.04 fix confirmed. Interrupters rejoice!

 

I.B. appears to still be broken. 

[Ruminate]

 

 

Here's the test I want you to run:

Scenario 1 - Interrupt 100% of the time

Scenario 2 - Interrupt 0% of the time

 

 

I extrapolate that the bear will die in Scenario 1 and that you will die in Scenario 2

 

 

How can you extrapolate that i will die in scenario 2 when I already gave you a scenario above where I killed the bear with 19 might and 0 interrupts? You're not making any sense.

Because you moved the goalpost of your own experiment. You changed all of your stats in between rounds, therefore you aren't testing the effectiveness of one thing. 

 

It's called putting your thumb on the scale.

 

 

You do realize there is a 97 point stat difference between 3 might, 19 dex, 18 con, 100 perception, and 100 resolve vs. 3 might, 19 dex, 18 con, 3 perception, and 100 resolve, right?

You do realize that letting a fight run its course with those stats is meaningless, right?

[Guest]

Designing Experiments for Dummies

 

Scroll down to the section on variables.

Read about dependent variables and independent variables.

Have a revelation about how your poorly designed "experiment" is a complete mess because you build 97 point stat swings into your test runs.

Stop posting in this thread.

Edited April 18, 2015 by Achilles