notes on v1.10 (beta) Assassin Blade skills

[Crystalion]

If I am interpreting the way the minions are being set up correctly, the bladecreeper will have the AR of the caster plus the listed bonus.

That would be nice and certainly something to utilize (e.g. weapon switch to cast minion with lots of +AR). However, my admittedly error prone tests for AR/DR with Blade Sentinel didn't seem to mesh with that idea (I had bumped my test Assassin at one point to over 900 AR without seeing a clear decrease of the number of misses). I will get around to altering the Creeper if need be to test this again soon.

The size=3, I would look at being about the distance that you could closely space three gems dropped on the ground.

Ah, thanks for the interpretation. That would make sense, as it does seem to touch nearby monsters pretty effectively.

For most of the missiles with a NextDelay, it is to avoid having the effect deliver far too much damage.

Yes, I understand the need. It's just a pity they have long delay effects like Blade Sentinel etc. when long delays generate so much contention with a single bucket check. Fixing this isn't trivial, afaik, which is why I mentioned my lack of hope that they'd touch it at this late date.

Also you may want to consider modding the NextDelay factor to a higher value for your testing.  Or even putting it on a different missile altogether for the testing purposes.  Putting it on say firebolt with a NextDelay=250 might prove useful for certain testing.

I've been pretty successful, now that I know what to look for, in seeing this effect without mods. But such a blatent mod might make a believer out of Zath, eh? I rather prefer the idea of putting the effect on Holy Bolt, so I might possibly grant my hireling/minions psuedo-immunity. :D

[Zath]

Well, here are my test results:

I loaded up an L99 Assassin from nobbie's library in players 8 Nightmare. Assigned 1 skill point to Shock Web and Blade Fury, maxed out Fire Blast (increase Shock Web missiles), Blade Sentinel (for duration), Cloak of Shadows (helps testing), and Wake of Fire & Inferno (for damage). I raised my dexterity to 150 for a 95% chance to hit Zombies in the Blood Moor. Purchased a Gladius of Pestilence from Charsi with 50 poison damage over 5 seconds.

Test 1: Shock Web alone

Found pair of Zombies, blinded with Cloak of Shadwos. Spammed Shock Web at Zombies to see how much life it would lose.

Result: Could not hurt Zombie at all due to health regeneration (both subjects).

Test 2: Shock Web with Wake of Fire.

Found Zombie, blinded with Cloak of Shadows. Tossed Shock Web at Zombie and quickly cast Wake of Fire next to it, then kept spamming Shock Web.

Result: Dead Zombie (4 trials each on 3 subjects).

Test 3: Wake of Fire alone.

Found Zombie, blinded with Cloak of Shadows. Cast Wake of Fire next to it to see how fast it died.

Result: Zombie dies in practically the same amount of hits as in test #2 (4 trials each on 2 subjects).

Test 4: Shock Web with Wake of Inferno.

Found Zombie, blinded with Cloak of Shadows. Tossed Shock Web at Zombie and quickly cast Wake of Inferno next to it, then kept spamming Shock Web.

Result: Dead Zombie (2 trials each on 3 subjects).

Test 5: Wake of Inferno alone.

Found Zombie, blinded with Cloak of Shadows. Cast Wake of Inferno next to it to see how fast it died.

Result: Zombie dies in practically the same amount of hits as in test #4 (2 trials each on 2 subjects).

Test 6: Shock Web with Blade Sentinel.

Found Zombie, blinded with Cloak of Shadows. Tossed Shock Web at Zombie and quickly cast Blade Sentinel at it, then kept spamming Shock Web.

Result: Poisoned Zombie on first pass of Blade Sentinel (10 trials on 1 subject). I could also hear the re-poison sound effect on each subsequent pass of the Blade Sentinel.

Test 7: Shock Web with Blade Fury.

Found Zombie, blinded with Cloak of Shadows. Tossed Shock Web at Zombie and quickly cast Blade Fury at it, then kept spamming Shock Web.

Result: Poisoned Zombie on first hit of Blade Fury (10 trials on 1 subject). The re-poison sound effect also occured if I hit the Zombie with multiple Blade Fury missiles before the initial poison duration wore off.

Conclusion:

Testing was not rigorious since I didn't mod anything, but there is no evidence to suggest that the NextDelay of one missile has any effect on that of another's. I think you better go ahead with the mod and make a believer out of me, Crystalion.

[Occhidiangela]

1. My Blade Sentinel with 2 Envy Jewels in the hat works wonders on players 8 for the Assassin Honor_Shoulder. If I could only get 4 more, two in a shield, two in a piece of armor, I'd be giggling me booty off. :) Your comment in re Shock Web explains quite a bit, but it you shock web after they turn green, I think it works out OK. Shock webbing before Bommeranging seemed to make me miss, now I begin to understand the why. Or, maybe I am just a spazz. :P

2. Blade Sentinel and its oddball pathing.

No kidding.

I think there is a max range wraparound or maybe some oddball feature that goes with it.

Sometimes I aim past the target crowd (I collect crowds first, then Boomerang to get max bang per toss) and the Blade Creeper/Boomerang goes just where I aim it, filling all and sundry with green love. Sometimes I aim past the target, and Boomerang goes behind me, directly away from where I am aimed the Boomerang. My hypothesis at the moment is that either I am aiming past an obstruction, and it is bouncing back pre launch, or there is a range limit that wraps around behind me (that seems less likely to me). Or, I am a spazz.

Aye, Boomerang does not do well going past rocks and trees and such.

I will take another look later today if I get a chance to mess with that.

[Occhidiangela]

Pathing.

If I place my cursor about 1/3 of the screen width away, the Boomerang just heads toward the target and comes back. If I start to get a bit further away, it seems that the Assassin slings the blade. That makes it first go backward, away from the initially selected target, and then stop and reverse itself and track toward the target. There seems to be a range limit of about 1/3 to 1/2 of the screen's width, more testing in different visual modes probably warranted.

Boomerangs do not go over thngs, they go arond them, and keep seeking the target up to max range, or the shorter than max range target point that your cursor selected. For damned sure will not go through bars/cages in the Act I Catacombs.

Boomerangs go around nearly everything: small tables, stackes of boxes, trees, chests, swamps terrain. Try throwing a Boomerang over a swamp square in the Black marsh, and it will go around it.

They don't go over anything as best as I can figure, they go around it, besides terrain that is "level ground."

Will look at more in Act II when I get there, but I expect to see the same.

[Occhidiangela]

Act I, Inner cloister.

I am armed with Vidalla's Bow. Sentinal must have a To hit calc, as you say, since I see the Lightning effect of the bow trigger on some passes versus the monsters, and on other passes, it does not trigger.

This suggest to me that your earlier assumption in re the missile annulment may have been an issue of Blade Sentinal AR, or perhaps not, hard to say.

This Assassin found Rixot's Keen this past weekend, but I have as yet to see proof that crushing Blow is passed along via Blade Sentinal the way it is on Blade Sentry.

Given that Shield and Sentry do, though, my brain suggests that logic (I know, each skill has its own neat little features, that CB should be carried along.)

[Crystalion]

Well, here are my test results

Very interesting. Our results disagree, so let's explore the apparent differences...

I'm using v1.10s unmodded. The last test I did, single player, after reading your post:
(non-hack) clvl 72 Assassin (I killed off my merc by dying--don't worry, I saved myself for later restore)
removed +skills items. pumped wake of fire to slvl 20 (shock web at slvl 1).
shock web too rough for normal zombies so tried players 8--still too rough.
moved on to canyon of magi (still players 8) and saw reasonably slow damage by shock web vs. brute type monsters.
cast one wake of fire to assure myself that monsters are totally wiped out by it in rapid order.
found another brutish monster and laid shock webs down on him. cast a wake of fire (which spammed); kept casting shock webs on the monster. Monster survived (and wasn't much hurt) as I expected.

So, are you using v1.10s?
Have you tried a non-hack character? (I doubt this matters--you could try a -act 5)
Were you casting the shock webs first and making sure you saw the monster satisfactorily hit by them?
Have you tried normal? (or, apparently, only nightmare difficulty? which I've not tested in at all, btw)
Have you tried not maxing out shock web missiles? (not required, so best saved for later testing)

I'll try nightmare and see if that makes a diff and edit back...
edit: players 1 nm zombies effect observed as expected; players 8 more difficult (as my weenie level 20 WoF hurts but doesn't kill them) to discern but apparently also as expected.

p.s. no harm in your testing it, but Blade Fury and Wake of Inferno wouldn't be expected to show any effect, as they don't have a NextDelay.

edit p.s.: procedural notes: I'm either not letting the monster move or I'm making sure shock webs cover the ground the monster is moving over. Once I see "immunity" to a single Wake of Fire (and it ends and I keep the shock webs going) I then spam a bunch of them (this never makes any difference, the monster is always immune to all of it, but I thought I'd describe what I'm doing to try to avoid making a movie of it).

[Crystalion]

Ruvanal wrote:

Also you may want to consider modding the NextDelay factor to a higher value for your testing.  Or even putting it on a different missile altogether for the testing purposes.  Putting it on say firebolt with a NextDelay=250 might prove useful for certain testing.

Zath went to the trouble of doing some tests and didn't get my results, so, aside from repeating my tests and making some suggestions/question for him, I've taken Ruvanal's suggestion...

I modified fire bolt and cold bolt to have NextDelays. I didn't observe any difference in a quick test. So I also gave them NextHit values of 1 and this seemed to activate the usage of NextDelay (thus possibly answering the question I raised for the two missiles with NextDelay but no NextHit).

Giving Firebolt a NextDelay of 25 and Coldbolt 250 I observed the following with a v1.10s -act 5 level 33 Sorc in Normal Players 8 Blood Moor (one skill point put into Cold Bolt and none for FB, as you are started with a +1 FB staff)...

Repeatedly firing Firebolts at a monster every other one would pass harmlessly through.
Repeatedly firing Coldbolts at a monster the first hits and the next ten seconds (roughly--I didn't time it) worth miss.
During that period (post coldbolt) the firebolts also all miss.

Ergo, the NextDelay effect isn't tied to a particular missile as far as I can tell (that is, it appears to be a "global" effect per target).

I will now go and do a PvP test with 3 players (with the mod) and edit report back...
edit: same results for a 3rd player... i.e. I have a PvP "victim" and use a hostile Sorc to Cold Bolt them after which I switch to another hostile player, a Necro, who finds his Teeth and Poison Novas won't harm the victim until the 10 seconds are up.

[Zath]

If you can reproduce those results in a modded game, then that's good enough for me. Please do not be offended by my doubts of your "personal experience". I just like to be thorough, and I would have asked for a more conclusive test regardless of who was reporting this bug.

[Crystalion]

If you can reproduce those results in a modded game, then that's good enough for me.

Ah, but what if I've accidently used "protomatter"?! Much better for someone else (aka anyone else) to be able to reproduce the phenomena.

Please do not be offended by my doubts of your "personal experience".

"Pay no attention to that man behind the curtain!" :D

I just like to be thorough, and I would have asked for a more conclusive test...

Actually, I feel the same way. In particular it niggles at me that you couldn't reproduce my results. An explanation for this, however simple the cause, is certainly worthwhile in principle even if there isn't some other quirk we've not uncovered.

For example, I still am not satisfied with my understanding of Blade Sentinel AR vs. DR, after hours of thinking about it, discussion here and mods/tests. Just because I don't have unlimited time to study a problem doesn't mean it is "solved", so I have to be careful when ignoring an oddity for convenience not to forget and think there is no oddity (another example would be that I'm not satisfied that we understand "passive blocking" from increased block rate in v1.10).

edit:...I wrote:

casting the shock webs first and making sure you saw the monster satisfactorily hit by them?

and I think it is worth noting that shock web is really sucky for it's targeting. A monster can be "completely surrounded" by shock webs and not actually be hit by them. Helpfully, if the monster is not immune to lightning, they will show a "lightning hit" animation once a second on their body if they are being hit. My best guess as to your inability to reproduce my results would be this targeting misfeature.

Incidentally, I've now done the test with lightning *immune* zombies (i.e. they take no damage from the shock web) and they nonetheless are still granted NextDelay "immunity" (by shock web) to Wake of Fire (et al). A helpful trick I often use, fyi, that I used in this latest test, is to turn off the monster (in this case zombie) regen.

[relight]

I think Wake of Fire and Wake of Infero doesnt have Next Delay, thats why Zath's tests didnt go right.

I'm not sure though, but it's not on Crystallion's list... (the one in his previous post)

[Crystalion]

I think Wake of Fire and Wake of Infero doesnt have Next Delay, thats why Zath's tests didnt go right.

I'm not sure though, but it's not on Crystallion's list... (the one in his previous post)

This is why I warned the list should be annotated--"wake of destruction (and "maker")" is, in fact, Wake of Fire. I provided the list anyway as a quick and dirty aid to those who aren't able to read or modify the .txt files directly.

edit: sorry for the confusion... but I'm really not qualified to be the one to provide such annotation.

[Nechtán]

Actually, if the current lightning strike bug is a manifestation of this bug, it is 100% certain that it exists.

Mod testing proved that using a spear attack that takes a long time to hit (when chilled or otherwise) the skill works 'reasonably fine', but with a fast javelin (and some animdata.d2 modifications for a 3 frame base speed) every other attack shorts out right away.

Which makes me wonder: those missiles with a 4 frame delay between hits, if they are already slow and have SM cast on them, can they actually hit twice?

Which makes me wonder #2: Nova, it slows down until it is stationary at its maximum range, I pray to god it has a NextDelay above 4 or it may actually hit twice at max range.

Which makes me wonder #3: Why can Firestorm hit more than once? I understand why Twister will never do so, but the firestorm fire 'pierces' in a sense too, so shouldn't it also be affected by this bug?

Which makes me wonder #4: When a freezing arrow pierces, won't the bug immunize the target to either the freeze splash or the arrow?

[Crystalion]

Mod testing proved that using a spear attack that takes a long time to hit (when chilled or otherwise) the skill works 'reasonably fine', but with a fast javelin (and some animdata.d2 modifications for a 3 frame base speed) every other attack shorts out right away.

Sorry, outside my expertise.

Which makes me wonder: those missiles with a 4 frame delay between hits, if they are already slow and have SM cast on them, can they actually hit twice?

Actually the way I discovered this "bug" was because I was trying to understand how I might make Blade Sentinel more effective (which includes the "standard" question: what does Slow Missiles do to this missile?). Depends on the data, but it is entirely possible there are SM profits to be made. These kinds of "ah ha" tricks can go a long time without discovery even when the underlying mechanism is known (e.g. the "laser" was published ages ago, but how many people know you can use Chill Armor to obliterate those annoying Magma Demon type monsters?).

Which makes me wonder #2: Nova, it slows down until it is stationary at its maximum range, I pray to god it has a NextDelay above 4 or it may actually hit twice at max range.

Player Novas appear to have NextDelay of 4.

Which makes me wonder #3: Why can Firestorm hit more than once? I understand why Twister will never do so, but the firestorm fire 'pierces' in a sense too, so shouldn't it also be affected by this bug?

Firestorm doesn't have a NextDelay.

Which makes me wonder #4: When a freezing arrow pierces, won't the bug immunize the target to either the freeze splash or the arrow?

AFAIK neither Freezing Arrow or its splash missiles have NextDelay.

[Guest]

i copied the skill column and also added/translated the missing/confusing ones.

if the part about blade sentry attacking untill it scores a successfull hit is true, then i guess the same applies to multi-shot: you stand next to a monster and shoot with 10% chance to hit and level 9+ MS you'd hit almost every time.

about fistoftheheavensbolt: when i fixed it(is it only to me that HB's never hit anything?) and didn't have any NextDelay the HB's would usually hit the target i hit with FoH directly, so this might only effect the missile if it's a sub-missile?

[Zath]

Aha! It all makes sense now. If NextHit=0, then NextDelay is simply applied as an "initial delay". This would explain the behavior of both Fists of Heaven's Holy Bolts and Diablo's lightning. The 3 frame delay for Diablo's lightning attack is precisely why players can avoid it completely if they stand really close to him. The diablight missiles have a rated speed of 13.3 yards per second. With a 3 frame initial delay, that means they won't damage anything standing within 1.6 yards of Diablo.

[Crystalion]

i copied the skill column and also added/translated the missing/confusing ones.

Nice table. The two zeros are, iirc, actually blanks (Ruvanal et al: isn't there a diff in some columns between "don't care" blank and zero?). Your two annotation question marks are, I believe, indeed Assassin charge-up skills released.

if the part about blade sentry attacking untill it scores a successfull hit is true, then i guess the same applies to multi-shot: you stand next to a monster and shoot with 10% chance to hit and level 9+ MS you'd hit almost every time.

Interesting thought. Would depend on the rate of travel of the missile I'd think, relative to its NextDelay. That suggests if the missile is too fast to benefit, then Slow Missiles might help.

If it helps in your and Zath's thoughts on figuring out FoH and Diablight: I have done a number of tests where I left NextHit blank but had a NextDelay of 25 (lightning, inferno1/2, fire bolt) with no apparent effect (i.e. no missiles missing, ever, even if adjacent).

[Guest]

The two zeros are, iirc, actually blanks

yep.

0 = blank = less than 1

[Zath]

How about modding the NextDelay for diablight to a very high value and see if that will make players immune to it from a farther distance?

[Ruvanal]

Aha! It all makes sense now. If NextHit=0, then NextDelay is simply applied as an "initial delay". This would explain the behavior of both Fists of Heaven's Holy Bolts and Diablo's lightning. The 3 frame delay for Diablo's lightning attack is precisely why players can avoid it completely if they stand really close to him. The diablight missiles have a rated speed of 13.3 yards per second. With a 3 frame initial delay, that means they won't damage anything standing within 1.6 yards of Diablo.

How about modding the NextDelay for diablight to a very high value and see if that will make players immune to it from a farther distance?

For this missile I point to the missiles.txt fields of
Param2=30
*param2 desc="Minimum range"

Which if I am remembering correctly will translate to about '5 yrds' game scale. Or about the distance of his outstretched hands which can reach over a characters head.

My results have been similar to Crystalion's with things like inferno missiles having a NextDelay=150, no failure to hurt the monsters if the NextHit=0.

The two zeros are, iirc, actually blanks (Ruvanal et al: isn't there a diff in some columns between "don't care" blank and zero?).

In some columns, yes there is a difference, but here these are zeros. It will depend on the data type that is being used for the field. In most cases the data type will be for one of the numeric forms (BYTE, WORD, DWORD, signed versions of these) in which case the blanks are zeros. If the data type is that of a string variable though, a blank would be an End of String symbol ('\0') while a 0 would be a zero with an End of String symbol ('0'+'\0') which in turn would need to parsed to actually mean a zero value.

Your two annotation question marks are, I believe, indeed Assassin charge-up skills released.

Actually they are not. They just appear to be a couple of unused missiles that the design team was playing around with during the development of the LoD game. Probably they were placeholders for the assassin skills till some of the other support work could be completed. There are quite a few unused missiles in the files, for example look at missile ID=178 'electric throwaxe' that I have found nothing that uses it. There is though a partial reference to this missile in the files that indicates that they may have considered using it as missile form for the one handed hammers (never enabled in the final release of the game v1.00).

QUOTE 
if the part about blade sentry attacking untill it scores a successfull hit is true, then i guess the same applies to multi-shot: you stand next to a monster and shoot with 10% chance to hit and level 9+ MS you'd hit almost every time. /QUOTE

Interesting thought. Would depend on the rate of travel of the missile I'd think, relative to its NextDelay. That suggests if the missile is too fast to benefit, then Slow Missiles might help.

The missile velocity and size would be very important here. With the speed of 24 and size of one, I don't think that there would be more than 2 checks on the larger sized opponents before the missile was out of range.

Which makes me wonder: those missiles with a 4 frame delay between hits, if they are already slow and have SM cast on them, can they actually hit twice?

IIRC from some testing long ago, yes a slowed missile like this can harm more than once. But it would require the target to be either rather large or stationary in most cases to actually achieve this result even with a slowed missile.

Which makes me wonder #2: Nova, it slows down until it is stationary at its maximum range, I pray to god it has a NextDelay above 4 or it may actually hit twice at max range.

Actually the Nova skill has a constant velocity of 24. What you are seeing is just a perceptual illusion that makes seem like the missile is slowing at the end. Probably this also is coinciding with the switch of one of the frames of the graphic animation.

Actually the way I discovered this "bug" was because I was trying to understand how I might make Blade Sentinel more effective (which includes the "standard" question: what does Slow Missiles do to this missile?).

It will also help if you understand just what it is that you are looking at. The missile blade creeper has a velocity of zero. It is basically being 'dragged' around by the monster bladecreeper (<- note no space) that is one of the assassin pet monsters that she summons. It is this creature that is the whirly thing that you see moving about, not the missile itself (which has no graphics).

[Zath]

Ruvanal:

Isn't the conversion factor 1 yard = 2/3 tiles, such that 30 tiles would be equal to 20 yards? I've always interpreted that minimum range as the minimum distance Diablo's lightning attack will reach regardless of other modifiers. Even if we ignore the conversion factor, I think you would have to be a lot closer than 5 yards to avoid Diablo's lightning attack. This of course, still leaves us with the question of what the significance of NextHit=0 with NextDelay!=0 is <_< Maybe NextHit just determines how often a missile will check for a hit? A combination of NextHit=0 and NextDelay!=0 would then imply the NextDelay check is not performed until the frame in which a hit or mutiple hits has occured passes? (i.e. you can hit a target multiple times in one frame, after which the NextDelay limit is applied)