Targetting

[Myth]

May I begin a discussion containing a (or a few) suggestion(s) regarding how the targeting system is set to work?

 

This involves naming its filters, so I'm asking just in case it would be considered spoiler material.

Edited September 25, 2015 by Myth

[Grido]

Assuming you refer to combat, anything in or around the target section on this page would be fine. If you mean something else then I'm not sure what you refer to by "filters".

[Syrian]

that info is horribly inaccurate and lots is unstated 

[Myth]

Targeting uses filters to determine who hits who. First and most obvious filter is creature strength. Weak hits the weaker, strong hits the stronger. But in case of same strength, other filters are applied to determine who gets hit. Those are the ones I'm referring to.

Also, what Sy said. See here:

 

http://magicduel.invisionzone.com/topic/16717-inaccurateoutdated-information-profile-and-skills-principles-weapon-sets/

 

:p

[Grido]

Assuming you refer to combat, anything in or around the target section on this page would be fine. If you mean something else then I'm not sure what you refer to by "filters".

 

that info is horribly inaccurate and lots is unstated 

I'm aware, hence the now bold/redded "or around" bit. I can update the info there, I just don't know enough about combat to comfortably rewrite it.

 

 

I personally have no issue with discussions along these lines, so long as it's more general rather than using actual creature examples.

[Myth]

We've done the tests, came up with some conclusions. Examples might be required at some point in the discussion, not sure about it. I just have some nasty beef with the way it's coded :p

 

So is it alright to name the other filters?

[Ary Endleg]

Why not? Nobody felt upset when I was talking about a lot of in-depth combat details in Combat Talk last year.

[Myth]

WARNING: information in this particular reply is apparently false ('tis what happens when the basis is wrong to begin with)

 

 

Very well, then.

 

Targeting follows a few simple filters, one of which I find unnecessary.

 

It begins logically:

- weak/strong checks creature total strength (as opposed to the description which states that only attack and defense matter)

- dying checks creature VE

 

But these checks are bound to fail at some point, specifically when values are equal. When this happens, Target Dying moves on to compare IDs, which is perfectly understandable. Someone needs to get hit, so let's pick the youngest... but then, Target Weak/Strong move to compare CIDs. Why?

 

By my understanding, a CID is a design number. In no way does it represent a creature's strength. A creature is as strong or weak as it's understood, therefor used in a correct or incorrect manner. This filter, in my opinion, needs to be left out. I would go as far as suggesting it replaced with creature level, or number of creature auras, or move on to compare VE.. and so on. But that would complicate matters.

 

Instead, I suggest it be changed to ID, same as Target Dying. You can't go wrong with that, since IDs are unique, so there would be no need for a 3rd filter (4th, if you count the fact that the creature needs to be alive to be targeted). There can even be a logic behind it. RL: younger = stronger; older = weaker. MD undying vampiric style: younger = weaker; older = stronger. Take your pick, come up with more, whatever. :P

 

(Disclaimer - Information above was attained through lots of tests. Thank you, testers! :P)

Edited September 25, 2015 by Myth

[Chewett]

Sure it works like that?

>> total strength (as opposed to the description which states that only attack and defense matter)

 

Meaning?

 

>>Multiple "filters"

 

How many tests were these? I hate to burst your bubble but...

Edited September 25, 2015 by Chewett

[Myth]

Same strength cases are difficult to produce, but there are 6 creatures which can be used to test this. Two, by two, by two.

 

These 3 cases were considered, and all 3 acted in the same manner.

 

An anomaly did come into play with Target Weak/Strong. After the CID filter, I assumed an ID filter would go in effect. Mostly it did, but in some cases, not so much.

 

We had a fight where same strength same CID lead to higher ID getting attacked by Target Strong, then another fight (in which another creature type was added, lower strength - rest were same creatures from previous) which lead to lower ID being targeted. That gave me a "random" impression on the 4th filter used for Strong/Weak.

Sure it works like that?

>> total strength (as opposed to the description which states that only attack and defense matter)

 

Meaning?

 

>>Multiple "filters"

 

How many tests were these? I hate to burst your bubble but...

 

Meaning this:

 

"Strong creatures
The creature will target the creature with the highest total amount of attack and defence attributes."

 

Isn't true. Reported and proven. It goes for all stats. :p

 

 

Now, burst my bubble, because the tests were many. :p

[Chewett]

What If I said you were taking too few battles and drawing too many conclusions from that?

By that I mean (in the stats world) you are overfitting your data.

"Strong creatures
The creature will target the creature with the highest total amount of attack and defence attributes."
 
Isn't true. Reported and proven. It goes for all stats. :P

Nope. Its not all stats.

[Myth]

On which part in particular? I'm quite sure the 1st filters for weak/strong are inaccurate compared to their in-game descriptions. So, the CID/ID part?

[Chewett]

Of what you call "filters", there is only ever one level, for everything. All the rest are undefined.

If its target weakest, its based on power+attack+defence only.

If its dying, its based on vit percent only.

Let me give you some code if you dont believe me:

[php] case 'weak':
asort($sortcr_power, SORT_NUMERIC);
$key = key($sortcr_power);
$ret[] = $creatures[$key];
break;[/php]

If the php sorting functions are deterministic, there is some determinism. But otherwise there is no further order imposed. Hence why I was saying you are overfitting your data as you are drawing conclusions where there are none.

If there are two creatures with the same power, then its undefined precisely which one is chosen, and depends entirely on how PHP sorted the array. Same with dying, as is everything else.

Hence, there is only one filter so anyone who is learning combat, please remember this (asthir please update your above posts by putting a warning). I suggest next time you run more tests, although clearly you have stated above that your tests disagreed (which surely led you to the conclusion your ideas about multiple filters was wrong?)

[Myth]

Then might I suggest a second filter, to make it possible to determine by players, and not PHP's mood?

[Rophs]

What does each target choice do if there are two creatures with equal need to be targeted? (same strength, both dying, etc) (does not apply to random)

 

What SHOULD they do?

[Chewett]

So as asthir pointed out creature Id will give a total ordering of what creatures will be hit.

So as a simple thing you could just say it's based on the first sorting (as explained) then the creature id.

I think it makes sense that the final sorting order is on creature id as you need to preserve a total ordering to make it deterministic.

Or we could make it random. In combat sometimes you falter, slip or make the wrong move. It would mean that you couldn't predict what it would hit, might make it more interesting and also potentially stop people tryin to predict what will happen precisely.

I like it not being random but think we could probably add more sorting orders to make it more interesting. Like surely a creature with less vitality is weaker than one with more, even if they have the same power attack defence.

[Rophs]

Order by position in riutal would make a lot more sense than order by id. id order would make two otherwise identical creates have one which is sometimes strictly better than the other. Order by ritual position would also be a lot easier to understand and realize without knowing beforehand what would happen after seeing the results

[Myth]

I'm still confused as to how that sorting actually happens. Two battles with the following ritual:

 

Attacker: 2 Water Guardian I + 2 Dark Archer

vs.

Defender: 1 Water Guardian I + 3 Dark Archer

 

In Target Strong (damage, so as to not alter strength), WGs from both sides get hit. Now I get it, they're stronger. But, highest ID is picked on the attacker's team.

 

Target Weak, which went after the DAs, in the fight above, hit the lowest ID on the attacker's side and the highest on the defender's side. So why did the happen?

 

After removing a creature type (the WG) from the battle, both Target Weak went for high IDs. After adding a creature a creature type for the attacker's side (2 Hollow Warriors), Target Strong went for the lower ID from attacker, and the highest from defender...

 

I don't see anything deterministic about this >_<

 

You say it's based on the first sorting, then creature Id. Are you referring to CID here, or the creature's actual ID? If it's the latter I can't find a logic behind it. It feels as if it depends on how many types of creatures are present on a team, and that makes no sense to me.

[Chewett]

You say it's based on the first sorting, then creature Id. Are you referring to CID here, or the creature's actual ID? If it's the latter I can't find a logic behind it. It feels as if it depends on how many types of creatures are present on a team, and that makes no sense to me.

No. My above post was indicating how is perhaps _could_ work. Please refer to my post http://magicduel.invisionzone.com/topic/16814-targetting/?p=167957 explaining to you why your "filter" idea is wrong.

Do you understand that if the strength/vit/whatever is the same, is it undetermined how the sorting is finished?

[Ary Endleg]

This needs to be sorted out. If everything is equal then some parameter needs to have final say who goes first in combat and in auras. ID seems to be a good solution but it would be better if ID isn't used.

 

It needs to be unique parameter, not random one, it should be possible to tactically use it. ID can't be tactically used because it's fixed.

 

I like Rophs' idea with slot number. Slot numbers are unique and you can tactically use it, meaning it's up to your decision where you want to place creature.

 

There's only small problem with slot numbers, each side has it numbered from 1 to 6. But that's not the problem if we give attacker in same slot priority over defender, this principle is already used in some other combat cases (auras for example).

 

So it could go in this order Att1, Def1, Att2, Def2..... Def6. OR Att1, Att2, Att3.... Def6.

[Chewett]

>>This needs to be sorted out.

Why? While I can probably guess a couple reasons, its always good to hear your reasons. Not just hearing demands to have X fixed :)

>>But that's not the problem if we give attacker in same slot priority over defender

Attacker priority over defender? Isnt this discussion about how a target is chosen, e.g. An aramor using attack strong is only targetting one of six, so attacker/defender precedence doesnt mean a thing surely?

[Ary Endleg]

Because currently we have no idea who will be targeted first, which aura will go first nor which creature will act first if all parameters are same.

 

It's not just a targeting issue, it's also order of attack issue, when more creatures have same init, the first one to act is the one that php/db sorts first, usually by id order. (if I remember correctly Asthir told me once about some anomaly to this id rule, I told him he's hallucinating :p :)) )

 

In order of attack issue attacker over defender priority would be needed. For targeting it wouldn't be needed I guess, although I have no idea how certain abilities are coded, but some do stuff on both sides of combat like lifesteal and martyr.

 

Also, if you have multiple auras and their sources have same init stat, which aura is applied first? :p

 

It's an extensive problem as you see, it originates from the fact that ordering algorithm doesn't consider possibility of it's parameters being equal.

[Chewett]

This thread is about targetting Ary, If you want to discuss other areas please make a new topic :)

[Myth]

Aaah... sorry, missed the "could".

 

But yes, it needs to be sorted out simply because it's a game system, and it makes no sense to have PHPs internal scripting mechanisms determine how a game system works.

 

Moreover, this system was at some point meant to be understood, and this becomes impossible when the parameters are no longer a part of the game, but the program it was built in instead. (unless you have knowledge of how that program works, of course)

 

I like the idea referring to multiple such filters. It makes sense that lower VE would imply lower strength. Slots can also contribute to this, if combat is thought to happen as shown: 3 creatures in front, 3 at the back, making the 3 at the back more protected = stronger, in order of slot number perhaps. (4, 5 then 6, followed by 1, 2 and then 3) And yes, slot number is indeed unique and can act as a final filter.

 

 

(off topic: and yes, this logic does extend to creature action sorting, so then a new topic it is; because the order simply can't be determined in same initiative situations, turning it into a luck based game; aura order CAN be determined, but based on a factor which can't be acquired easily by the other team: "hey, may I know your creature's [insert factor] so I know what to hit you with and win?" :p)

P.S.: Random factors already exist: there's random damage, target random, target multiple, some aura effects... don't think we need more randomness around those :p

[Ary Endleg]

Well I simply pointed out other areas that have similar problems which originate due to same thing. Hence it seems only logical to mention them too since they could all benefit from same fix.