Forum Settings
       
« Previous 1 2
Reply To Thread

Tactical Processor effects confirmed.Follow

#1 Aug 04 2011 at 10:10 PM Rating: Excellent
10 posts
So it's not the newest thing, but hey, we now have this to work with:
Camate wrote:
Prepare to have your minds blown as we have…yes…information about what the Tactical Processor does!

This attachment is a bit different than the standard attachment that, for example, gives XX% increase in attack, defense, etc. The automaton periodically checks the information around the automaton and decides its next move (if HP took a hit, it will cure, if TP has been saved up, it will perform a WS, etc.).

When using the Tactical Processor the frequency of these checks is increased and allows the automaton to respond to variations in its environment quicker and more precisely.

However, every time it performs a check, the overload value increases, so as the frequency of checks increases it will increase the possibility of an overload.

Epic sauce? Anti-climatic? Let us know!

Personally, I think that effect is sort of meh seeing as we couldn't even perceive any notable/consistent differences in AI with it on, but it's relieving to know what it finally does.

On the other hand though, maybe it would be good for making my automaton use a weapon skill before ~130% TP more frequently, though I'm not sure just how much that helps in the long run.

By the way, is this function of the automaton checking various conditions possibly also the difference among the various animators?

Edited, Aug 5th 2011 12:26am by PBlackII
#2 Aug 04 2011 at 10:40 PM Rating: Default
Seems sort of redundant, as to me they are just reiterating what it says in the item description.
____________________________
~Itege~Cerberus~
90PUP/90DRG/90NIN/90DRK
LordFaramir wrote:
Ya know, you're really similiar to me, whether you like that or not :p, but you need a catgirl display picture.

#3 Aug 04 2011 at 10:42 PM Rating: Default
Scholar
**
312 posts
i dont know whether to congratulate you on an interesting troll or tell you to wait a couple of days for a real friday thread :)
____________________________
99war 99mnk 96whm 99blm 90rdm 99thf 85pld 99bst 99brd 99drg
99blu 99pup 60sch 99dnc 99nin 99sam
50/50 glavoid shells
50/50 itz scales
6/75 orthrus claws
#4 Aug 04 2011 at 10:47 PM Rating: Decent
10 posts
What? This came from Camate and I didn't see any other threads in particular explaining this attachment. So I'm not sure how I'm trolling. Really though, this isn't too shocking, it explains why it's so difficult to test.
#5 Aug 04 2011 at 11:58 PM Rating: Excellent
itege wrote:
Seems sort of redundant, as to me they are just reiterating what it says in the item description.


well whoop the doodle doo we are back to square one.
____________________________
[ffxivsig]1637915[/ffxivsig]
#6 Aug 05 2011 at 2:23 AM Rating: Excellent
**
275 posts
Ok no, actually, I think I can wrap my head around this. Have you ever noticed how, for example, if the automaton has a maneuver up to use a job ability, and also has tp, or is a mage frame, it will never use the weaponskill and the job ability back to back but will have a chunk of time between the two?

Well, think of it like three concurrent timers- One where it takes in the situation and makes a decision, one where it acts using weaponskills or job abilities as tp/maneuvers allow, and the global spell casting timer. I'll call them I, A, and M, respectively.

So every time I comes up, the automaton gauges the current situation, and whether M is up or not, locking in what action it will perform next time A occurs. Now, if in the interval between I and A, something changes, the automaton would, in theory, act on out of date information.

The exact ramifications of this, though, are hard to tell, but if my understanding is right, we've had a hard time testing the tactical processor because we've not been looking at the right factors. We've been looking at the automaton to do things more quickly, when, perhaps, we should be looking for the automaton to reassess the situation more often, and thus make less incorrect decisions based on out of date information. The question is, then, how can we test to show that the tactical processor would be doing anything. I have a few ideas, but they're hard to pull off.

One possible test would be to take advantage of the soulsoother's much maligned na-spell priority. If I'm right, if you were to, say, take a poison potion in the interval between I and A, the automaton would still cure you. This would, obviously, take a lot of poison potions to get right, and very good timing. But, if someone were to go at it enough, they might be able to see how far apart I and A are without the tactical processor, then repeat the experiment with the processor attached, and compare the difference.

The other thing, depending on if my idea about this is correct and a possible exception to the rule, is that this might explain why valoredge shield bashes so infrequently, and so often too late. It may be that the mob has to be readying a TP move when I occurs, to trigger a shield bash the next time A comes up. Now, depending on just how the fine details of this go, it could be that the tactical processor may well have at least one valid purpose, that of improving valoredge's shield bash use timing.

Now, this is still subject to some unknowns. Are I and A completely separate timers, or a set cycle? If they are one cycle, does the tactical processor shorten the overall I to A to I again sequence, or merely the distance between I to A, possibly lengthening the distance to the next I.
Or, if they are on separate timers, does this mean that the automaton performs multiple I checks per A cycle?

Personally, I find all these questions interesting, if perhaps mostly academic.

EDIT: The delay between ranged attacks by sharpshot would work like this as well, functionally replacing M when using sharpshot instead of a mage frame. So at each I it checks if it has maneuvers for a job ability attachment, if it has tp for a weaponskill, or if it's ranged attack cooldown is available. A few simple tests then would determine what the priority is, in that case.

Edited, Aug 5th 2011 4:27am by VonCrown
#7 Aug 05 2011 at 3:41 AM Rating: Good
Avatar
**
275 posts
That's kind of how I'm looking at it, VonCrown.

If that's how the Automaton determines its next action, perhaps that is also why it seems to wait a considerable amount of time after 100% TP before using a weaponskill. This is merely conjecture, but what if it waits until it's in an idle stance before running the Check Function? Auto-Attack might not be on a timer, but more of an on-rails thing where it just attacks at a frequency. A higher attack frequency might be interrupting it from performing a Check Function, hence it waits a touch longer to perform a WS.
#8 Aug 05 2011 at 4:36 AM Rating: Decent
10 posts
I completely agree with VonCrown. I think the easiest way to test the theory is this:

Use soulsoother frame equipped with flashbulb and have a poison potion ready. Use the poison potion, deploy on the mob, and record the difference in time from the use of flashbulb and the use of poisona. That interval of time would be the frequency the automaton checks conditions.

That or instead of soulsoother for poisona, use eraser attachment on a melee frame so that way we test two JA timers.

Edited, Aug 5th 2011 6:40am by PBlackII
#9 Aug 05 2011 at 5:09 AM Rating: Good
Avatar
**
275 posts
PBlackII wrote:
I completely agree with VonCrown. I think the easiest way to test the theory is this:

Use soulsoother frame equipped with flashbulb and have a poison potion ready. Use the poison potion, deploy on the mob, and record the difference in time from the use of flashbulb and the use of poisona. That interval of time would be the frequency the automaton checks conditions.

That or instead of soulsoother for poisona, use eraser attachment on a melee frame so that way we test two JA timers.

Edited, Aug 5th 2011 6:40am by PBlackII


I don't think it would be that simple though. The automaton can do multiple tasks per check it seems, or it's checking far more frequently than we think. My guess is that it works by queueing tasks based on the feedback when it checks and follows until it reasseses again. This seems most likely if the true function of the attachment is to have it check more frequently.
#10 Aug 05 2011 at 6:04 AM Rating: Decent
10 posts
I just tested the eraser/flashbulb and the time of use between the two was consistently around 5 seconds with and without the Tactical Processor/Ice Maneuvers. This does not appear to affect back to back usage job ability attachments, so I think it confirms Aeria's suspicions that it does multiple tasks per check.

Perhaps looking at automaton job ability attachments isn't the right approach either, instead look at what we know what it should affect according to Camate's post: cure according to hp assessment, or weapon skills according to TP; the latter would be easier to test due to tactical switch.

I will see how fast it will use the same weapon skill in a row, I feel like it won't generate much results, but hey, it's something.

Edited, Aug 5th 2011 8:10am by PBlackII
#11 Aug 05 2011 at 7:06 AM Rating: Good
*
173 posts
Again, please link, people.

____________________________
PUP90
Selka - Bismarck server
#12 Aug 05 2011 at 8:15 AM Rating: Decent
PBlackII wrote:
I just tested the eraser/flashbulb and the time of use between the two was consistently around 5 seconds with and without the Tactical Processor/Ice Maneuvers. This does not appear to affect back to back usage job ability attachments, so I think it confirms Aeria's suspicions that it does multiple tasks per check.

Perhaps looking at automaton job ability attachments isn't the right approach either, instead look at what we know what it should affect according to Camate's post: cure according to hp assessment, or weapon skills according to TP; the latter would be easier to test due to tactical switch.

I will see how fast it will use the same weapon skill in a row, I feel like it won't generate much results, but hey, it's something.

Edited, Aug 5th 2011 8:10am by PBlackII


Flashbulb a bad thing to test it with, the automaton automatically uses it within seconds of a light maneuver going up or it's timer coming up. Tact should effect conditional abilities like eraser. Where it doesn't activate unless you have a status effect.

Shield bash would be the best thing to test it with in my opinion becuase it only effects decision making. And there's only a handful of things it has to decide, when to weapon skill, what spell to use within the combination of maneuvers, and when to use by condition only Job abilities.(Eraser, shield bash)

So you guys should be testing things like shield bash, eraser, etc. See if responds to a shield bashable tp move or -ga spell quicker then it normally would etc.

PS Yes black I know you also tested eraser.
____________________________
If cookies were milk and milk was cookies. Would Oreo's still be america's favorite brand of milk?
#13 Aug 05 2011 at 9:05 AM Rating: Good
10 posts
Right, no complaints there; as for now, I'm tired. I did this if it's worth anything to anybody—not enough to satisfy me with any answer with details to this thing.

So basically what I did was get automaton past 100% TP, deployed (and stayed close to it), waited for it to use a weapon skill, immediately use tactical switch, and then record the amount of time it took for it to use a weapon skill again. My rationale is that within every check, the automaton will see if it has enough TP to use a weaponskill; each time it uses a weapon skill, it checked to see if it met the conditions to be able to do so. Ice maneuvers with the tactical processor equipped would decrease the difference in time between weapon skill use.

I recorded two numbers: the difference in time from when both weapon skills were readied, and the difference in time that both were used (as you can see, they aren't more than a second in difference). It appears though that without maneuvers, the difference in time is usually 8 seconds, while with any amount of maneuvers, it's around 5 seconds. Again, this is not conclusive data, need more numbers to verify.

The mob I did this on were defoliate trees in Bhaflau Thickets because I needed something that would not die within a couple weapons skills and melee, but not cause the automaton too much harm.

Also, I did not do 3 maneuvers to the extent of the others, for it was getting similar results as with one or two ice maneuvers and caused me frustration with one overload consuming so much time. Somebody with Kenkonken or something can test that.

Finally, this is data I collected, this isn't conclusive at all to me though; somebody else please do something similar to help get better data:

No Tactical Processor Equipped
Readies
· 6 seconds
· 8 seconds
· 5 seconds
· 8 seconds
· 6 seconds
· 9 seconds
· 11 seconds
· 8 seconds
· 8 seconds
· 11 seconds

Uses
· 5 seconds
· 8 seconds
· 6 seconds
· 8 seconds
· 6 seconds
· 8 seconds
· 11 seconds
· 8 seconds
· 8 seconds
· 11 seconds

Tactical Processor Equipped 0 Ice Maneuvers
Readies
· 8 seconds
· 8 seconds
· 6 seconds
· 5 seconds
· 14 seconds
· 8 seconds
· 11 seconds
· 5 seconds
· 16 seconds
· 10 seconds

Uses
· 7 seconds
· 8 seconds
· 7 seconds
· 5 seconds
· 13 seconds
· 7 seconds
· 11 seconds
· 5 seconds
· 16 seconds
· 11 seconds

Tactical Processor Equipped 1 Ice Maneuver
Readies
· 7 seconds
· 5 seconds
· 5 seconds
· 9 seconds
· 4 seconds
· 11 seconds
· 5 seconds
· 15 seconds
· 5 seconds
· 5 seconds

Uses
· 8 seconds
· 5 seconds
· 6 seconds
· 10 seconds
· 5 seconds
· 12 seconds
· 5 seconds
· 14 seconds
· 5 seconds
· 5 seconds

Tactical Processor Equipped 2 Ice Maneuvers
Readies
· 5 seconds
· 10 seconds
· 6 seconds
· 10 seconds
· 5 seconds
· 5 seconds
· 6 seconds
· 11 seconds
· 10 seconds
· 7 seconds

Uses
· 5 seconds
· 9 seconds
· 7 seconds
· 11 seconds
· 5 seconds
· 6 seconds
· 5 seconds
· 11 seconds
· 9 seconds
· 6 seconds
#14 Aug 05 2011 at 9:45 AM Rating: Good
Avatar
**
275 posts
I thought of some things we need to consider.

1) Does the introduction of a maneuver force the automaton to reasses the situation on the spot vs waiting for its natural check timer?

2) Does the check function work at all times or only during Deploy / attacking?

3) Does Tactical Processor raise the burden floor or raise the amount of burden of each maneuver?

As far as Shield Bash goes, I believe that it's actually programmed to react, as the automaton wouldn't know when a mob would use a move and thus wouldn't be able to queue it.

Edited, Aug 5th 2011 11:46am by AeriaAllslove
#15 Aug 05 2011 at 10:56 AM Rating: Decent
***
1,270 posts
Truth be told, i wouldnt be suprised if the only thing that worked was the overload increases aspect.

It wouldn't suprise me in the least if the other half that is supposed to decrease reaction time did nothing at the moment, similar to using a scanner on a black mage frame...
#16 Aug 05 2011 at 1:43 PM Rating: Good
**
490 posts
Assume the tactical processor is just like a faster processor for your computer. Faster tics in exchange for potential for overheating. The auto makes decisions on set intervals where an 'if' statment would be required. (Shield bash comes to mind) Let's slow everything way down and assume the autos standard processor tic is every second (its probably every few microseconds realistically.) Say a mob uses a tp move at 3.5 seconds, the bash will occur at 4, after the move goes off. Tactical processor could shorten the tic time, just like higher end processors do, allowing it to make faster decisions, and now bash at 3.7 or 3.6, possibly stunning the move.

More tics per second would create less 'down time' where burden can decay, making overloading easier.

It would probably only apply to times decisions need to be made, curing, weaponskilling, bashing... and now, a possibly easy way to test it, dispelling. Can we clock how long it takes to dispel a buff (once used by the mob) with and without the processor equipped? The resulting time might be impossible to calculate by eye or with the log. It also might be hard because monsters may use the tp move closer to, or further from the auto's set tic. It could use the move at 3.1 or 3.6, and the auto would still only dispel on its 4 tic, creating a natural variance. The tactical processor should shorten the average of that variance.

(Posting from phone, apologies for any grammar/autocorrect/nonsensical stuffs)
#17 Aug 05 2011 at 4:27 PM Rating: Good
**
275 posts
I think my poison potion test was a bit misunderstood, it's already been pretty well confirmed that the tactical processor doesn't actually reduce the forced interval between job abilities/spells. What I meant is, using the example tested with flashbulb, see if without a tactical processor, if you use a poison potion very quickly before you expect a spell/ability, it might still use the spell/ability it had previously planned, if you used the poison potion in the interval between when it checks, I, and when it acts, A. What I meant to look for was how soon before the spell do you have to use a poison potion for it to -not- cast poisona, but rather a different spell, while you are poisoned, because it hasn't 'noticed' the poison yet at the moment it casts. If we could roughly work out the longest interval between using a poison potion and having a non-poisona spell/action taken, we could figure out the length of the delay between the automaton deciding what to do, and doing it. Then, one would add the tactical processor, and repeat.

Think of it less like the automaton doing things more often, and more of reducing the delay between when it checks and decides and when it acts, causing it to act 'appropriately' more often.
#18 Aug 05 2011 at 5:39 PM Rating: Good
**
610 posts
So Von, Basically you're proposing that the reason the automaton casts poisona over cure when you're in dire need of a cure, is because the automaton already decided to use the spell poisona before your HP fell into the red? The decision was just waiting on the cast timer for the next spell.

I guess the counter to this is that if the tactical processor was equipped, it might have re-assess that you needed a cure instead of a poisona before it cast poisona right?
#19 Aug 05 2011 at 6:33 PM Rating: Good
**
490 posts
Jayheart wrote:
So Von, Basically you're proposing that the reason the automaton casts poisona over cure when you're in dire need of a cure, is because the automaton already decided to use the spell poisona before your HP fell into the red?


From a programming standpoint, there needs to be a timer the auto uses to check on variables. This is almost MANDATORY in games for AI. Generally in AI programming, a programmer will us some kind of timed loop to check/react. It's probably the same for autos. It's probably the same for mobs. The timer itself is unknown in this case. We know job timers, but we don't know at what intervals the auto gets it's on status update. Probably 10-40 times a second or so (for a ballpark estimate.. really, it could be 2, for terribly slow updates, or 100+ for a well-informed machine. I'd guess at least 10+ a second, which would still allow JAs to appear to hit at exactly 30 seconds, ect..)

It most likely checks all of these and more at every interval.

(in no particular priority order)

Do I have enough TP to use a WS?
Is my master injured enough to require curing?
Am I injured enough to require curing?
Does my master require a status removal?
Do I require a status removal?
Do I have any job abilities that I can use, and are the required maneuvers up?

After whatever has the top priority is decided, it then queues that action and performs it either instantly, or at it's next interval.

The Tactical Processor (since it's well... a processor) probably shortens those intervals. This is probably why people have sometimes observed more occasional (but not constant) stunning of TP moves, or more occasional WSing as soon as it reaches 100tp, ect, ect..

Because of the nature of when that processing interval tics, it's never going to be constant, like I said in my above example.

Based on the arbitrary number 4 again, it could reach 100% TP at 3.1, 3.6, 3.8 and wouldn't WS until 4. If that tact processor (arbitrary) reduced it by .2, you would then see it more consistently (but not constantly, as it could > 100% at 3.1 and not go off for another .7 (instead of) .9) go off sooner.
#20 Aug 06 2011 at 6:14 AM Rating: Good
Avatar
**
275 posts
I took Sharpshot out last night after work just to see if I could eyeball any changes in Machinette's behavior equipped with Tactical Processor rather than without it.

Observations:
1) Whenever I threw up a Light Maneuver, Flashbulb would go off before my animation completed, rather than wait a few seconds for it to happen.

2) Likewise, whenever I gave a Wind Maneuver (and the JA timer was ready) for Barrage Turbine, Mach wasted no time in using it (As in, Wind Man, Deploy, Barraged right away), again instead of waiting a few seconds.

3) Buffoon's Collar and Cirque Farsetto +1 combined seems to cover the burden accumulation penalty enough where you can still operate Maneuvers normally. Even using two winds back to back, I never Overloaded all night.

4) Mach was more prone to WS as soon as TP was 100% rather than wait a few swings for 120, 130, etc. Enemy HP varied each time. Without TacPro installed, Mach liked to wait until about 1/2 HP, even when TP was over 100% and appropriate Maneuvers were present.

This small test reminds me of another Tactical Processor test I saw some time ago, before we were told what it does (though I can't seem to find the thread any more). The Tester at the time determined that Eraser and other JAs seemed to go off sooner.

With this small amount of data, I can hypothesize the following:

1) Without TacPro, Check Function (CF) happens at already short intervals.

2) With TacPro, CF happens even faster.

3) I believe the CF occurs at all times, even when not Deployed. This is evidenced by still being able to Overload easily with TacPro while Retrieved / Not-Attacking (Pre Overload-supression gear). The Automaton may simply be checking HP, MP, TP, and possibly WSs to be ready to act when Deployed.

4) CF adds a negligible amount of Non-Elemental burden per check. However, quickly repeated CFs when TacPro is equipped can be too much too fast, especially with repeated same-element Maneuvering.

5) The introduction of a Maneuver forces the Automaton to run CF with the new Maneuver in consideration. This is evidenced with Maneuver-dependent Job Ability attachments firing upon use of the Maneuver if the JA timer is ready.

6) TacPro may prioritize Maneuver-dependent Job Ability attachments above other considerations. As mentioned above, my Flashbulb attachment fired almost immediately with TacPro vs. the few seconds it usually takes without it.

----------

Obviously, there is still much more testing to be done, but from what I can tell, Tactical Processor does seem to be doing its job. I'll be playing with the Stormwaker series this afternoon to see if I can get some data about it's spell casting behavior.
#21 Aug 06 2011 at 3:07 PM Rating: Good
Late to thread but let me throw this out to see if i'm in line with current thinking after this bit of info from SE.

Essentially, Tact. Pro.(TP) takes the flowchart-ish AI of the automaton and shortens the time between each step. And in classic SE style, to balance the TP having a mildly beneficial effect, its use sharply raises the chance of overload if you don't have items to help manage it?

If my take is correct on what you guys are saying, even knowing what it does wont have me equipping it anytime soon.
____________________________
[ffxisig]10000[/ffxisig]
----------------------------------
One of my server's brightest minds... wrote:
Not to make too extreme a comparison, but Rog is like Nelson Mandella...


Quite possibly the DUMBEST thing i've ever read.
#22 Aug 06 2011 at 5:19 PM Rating: Good
Avatar
**
275 posts
spiritreaverdiablos the Irrelevant wrote:
Late to thread but let me throw this out to see if i'm in line with current thinking after this bit of info from SE.

Essentially, Tact. Pro.(TP) takes the flowchart-ish AI of the automaton and shortens the time between each step. And in classic SE style, to balance the TP having a mildly beneficial effect, its use sharply raises the chance of overload if you don't have items to help manage it?

If my take is correct on what you guys are saying, even knowing what it does wont have me equipping it anytime soon.


Not so much reducing the time between steps in a flow chart, rather how often it re-draws the flow chart based on new data. I'm still thinking of how to play with the Stormwaker to figure out how it will affect spell casting.

Edited, Aug 6th 2011 7:19pm by AeriaAllslove
#23 Aug 06 2011 at 7:08 PM Rating: Decent
**
490 posts
Their description was pretty cut and dry. The auto updates itself with relevant info every so often. Tactical Processor reduces that interval. Fin.

Now, let's find a way to actually test it and find a use for it.

Brainstorm testing ideas.
#24 Aug 06 2011 at 8:57 PM Rating: Good
AeriaAllslove wrote:
spiritreaverdiablos the Irrelevant wrote:
Late to thread but let me throw this out to see if i'm in line with current thinking after this bit of info from SE.

Essentially, Tact. Pro.(TP) takes the flowchart-ish AI of the automaton and shortens the time between each step. And in classic SE style, to balance the TP having a mildly beneficial effect, its use sharply raises the chance of overload if you don't have items to help manage it?

If my take is correct on what you guys are saying, even knowing what it does wont have me equipping it anytime soon.


Not so much reducing the time between steps in a flow chart, rather how often it re-draws the flow chart based on new data. I'm still thinking of how to play with the Stormwaker to figure out how it will affect spell casting.

Edited, Aug 6th 2011 7:19pm by AeriaAllslove


Ok that's pretty interesting and i can see how tests run before had a hard time puzzling out its exact function. Still think i'll skip it for now, but still nice SE finally gave answer to the questions of what it does and how.
____________________________
[ffxisig]10000[/ffxisig]
----------------------------------
One of my server's brightest minds... wrote:
Not to make too extreme a comparison, but Rog is like Nelson Mandella...


Quite possibly the DUMBEST thing i've ever read.
#25 Aug 08 2011 at 10:40 AM Rating: Good
***
1,270 posts
johnnykrysys wrote:
Their description was pretty cut and dry. The auto updates itself with relevant info every so often. Tactical Processor reduces that interval. Fin.

Now, let's find a way to actually test it and find a use for it.

Brainstorm testing ideas.



Except people tried to test it for 6 years and found nothing. Nowing what it does now doesn't exactly make it easier to test. That fact no one has been able to parse a difference in 6 years already means there is no 'use' for it.

I am a firm believer that this attachment does nothing of use...not saying its broken, just that you will never see a noticable/parseable difference.
#26 Aug 08 2011 at 1:03 PM Rating: Excellent
**
480 posts
Exactly. One of the first things people tried was seeing if Shield bash reacted more quickly and could actually stun TP moves, instead of doing it way later. It didn't increase its performance in a noticeable way, so if TP really reduces check times... It's not enough to be noticed, thus not enough to be worthwile.
____________________________
Frejan - Ragnarok
SCH 99 PUP 99 WAR 99 RDM 49 NIN 49 SAM 49
Windurst Rank 10, ZM+CoP: The Last Verse, ToAU: Eternal Mercenary, WotG Champion of the Dawn.
« Previous 1 2
Reply To Thread

Colors Smileys Quote OriginalQuote Checked Help

 

Recent Visitors: 1 All times are in CDT
Anonymous Guests (1)