IDEAS FOR RYZOM

Sorry -- I was incomplete in my statement.  I meant that humans want to see patterns and do even if the data is random.  You are perfectly correct that testing for autocorrelation needs more sophisticated action.

Where did you get the 0.31 figure from?

Something i don't understand in your answers or that you don't understand in the said problem is the following. This is not a problem about :
- Damage versus mob hp
- Spell cost versus regen
- My resistances versus mob's magic against me
- Chicken versus crossing the road
- Etc.

Just this : My spell versus chances of resists from the mob. Also not partial resists (they are annoying too but at least i still do damage), not damage mitigation (i know it's stupid to use acid against a kipee). Full resists.

The plus side of melee is the stanza accurate attack that gives a bonus to hit. We don't have that in magic (and unless you prove me wrong, "concentration" avoids interruption /failures and not resists).
Just a hint, by the way. I heard (that's a solid source of info, indeed) that intelligence stat improves greatly the chances of spell penetration (i.e.: avoid a full resist). So it "would" be a good idea to level healing a lot further than elemental magic. Truth or false ?

Intelligence raises your sap level. The highest quality Magic Amplifier you can use is your Intelligence + 10.

All branches of magic give you access to raising Int at trainer.

For most mages, Int is raised first if an upgrade is available. The Magic Amplifier Q limits the usable level of spells, so buying spells higher than the Q of your amp will not work well. You have to get a higher Q amp first.

Our answers all said (aside from the digressions into other matters): "That's the way it is.  It's not a bug."

Higher level spells will reduce resists.  Higher success rates for casting decrease the problem of resists.  It might be that higher intelligence reduces the chance of resists, but it would be hard to separate that from the other factors since most people increase the power of their spells in parallel with increased intelligence.

On the other gripping hand, if you and I are together online at some point, we could test that by setting up identical stanzas and then seeing if my casting (with max intelligence) had fewer resists than yours (with less than max intelligence, I'm assuming). 

I pledge that when you want to do that test, I will work with you to do it.  Only when we have scientific measurements of this matter will it be able to determine if your other question has meaning.

 -- Bittty (associate in the New Trykoth Council of Sciences)

Xatokar, do you actually realize that your answer contains absolutely zero relation to my question ? If you have no anwer but you like reading to your own wisdom, open a text file and enjoy yourself.

Thanks Bitty on the other hand for saying something actually helpful.
But for what i can see, spell level (QXX of spell) doesn't change anything and it may lower success rate if the spell is brand new. Skill level does it, well afaik.

Any time for tests, anyway.

Hiya -

The issue the OP is trying to bring to the table is regarding 2 things both of which have outlined in the original post, then clarified later, and lastly re-clarified again... none of which you have adressed at all or have come remotely close.

Spell resistance is completely independent from magic protections.
You inflict 50(100) points of magic damage on xyz -> 50% magic protection
You inflict 50 points of magic damage on xyz, although using a spell that can hit for 100 and mob is still alive -> partial resist (rare)
Xyz resists your spell -> full resist

Mobs can have protection on some kinds of magic and resistance to others. That means it's not always the best idea to use a spell that does full damage.

As a new player you only got 3 types of elemental magic available. There might be mobs that are resistant and/or protected from all 3 kinds. Anyway it's most important to find the right magic type against an opponent.

At Level 125 you will get 4 more types of elemental magic. Those help a lot because it's much more likely one of seven spells is effective against a given mob. With a few exceptions of mobs that are protected against all kinds of magic an elemental mage at 125+ should have no trouble hitting creatures 10 level above his own level (maybe 10-20% partial or full resists). Creatures 25 Level above the player's level can still be killed easily, quickly getting worse if the difference gets bigger.

Gingos are nasty creatures. Their presence in front of Yrkanis makes it the hardest starting place for new players. They resist a lot, they got above average hp, they blind you (making you unable to hit, dodge or cast spells).

Here's a simple autocorrelation experiment: Low-level mage (lvl 36; jungle resist 65; lakes resist 41) attacks sluggish psykopla (lvl 9) with acid 1 (lvl 3 spell) cast from an enchanted daggar. This is one continuous battle! (Tough little plant! :p )

Homin hits 457 times and is resisted 41 times.
Psykopla hits 210 times and is resisted 307 times.

What follows is the correlogram data showing the autocorrelation function at different lag values. Correlation values can range from -1 to 1, with 0 meaning no correlation, 1 being a perfect correlation, and -1 being a perfect negative correlation. The values for lag 1 show how correlated adjacent (one apart) casts are, values for lag 2 show how correlated casts two apart are, etc.

Homin attacking plant:

The autocorrelation at lag 1 is -0.090
The autocorrelation at lag 2 is -0.037
The autocorrelation at lag 3 is 0.043
The autocorrelation at lag 4 is -0.038
The autocorrelation at lag 5 is -0.065
The autocorrelation at lag 6 is 0.018
The autocorrelation at lag 7 is -0.034
The autocorrelation at lag 8 is 0.047
The autocorrelation at lag 9 is -0.034
The autocorrelation at lag 10 is 0.020

Plant attacking homin:

The autocorrelation at lag 1 is -0.036
The autocorrelation at lag 2 is -0.047
The autocorrelation at lag 3 is 0.000
The autocorrelation at lag 4 is -0.022
The autocorrelation at lag 5 is 0.020
The autocorrelation at lag 6 is -0.026
The autocorrelation at lag 7 is 0.032
The autocorrelation at lag 8 is 0.023
The autocorrelation at lag 9 is -0.024
The autocorrelation at lag 10 is 0.052

We see very low values centered on zero, indicating that no autocorrelation is apparent in this dataset. :)

Here's the actual data set. 1 means a hit, and 0 means a resist. :)

Homin attacking plant:
11111011111111111111111111111101111111011111111101111110111110111 11111111111111111111111111111111011111111111010111111111101111111 11111111111111111011111111111101111111011011110110111111111111011 11111111111111011111111011111111101111111111111111101111111111111 11111111111111101111111011111111111111111111111111111111110111111 01111111111111101101111111111111111101111101110111110110111111111 11110110111111111111111111011111111111111111101011101111111111111 1111101111111111110111111111111111110111111

Plant attacking homin:
01001000110000111101001100000000010000010011000010010010100010000 10100111101110001000100010011101101000100100011001010101001111010 01011000100001111001011100000011111001110001110100101111011000101 00100010110001111100100110010000001101000001000100000110111010000 11101000011000101011101100001000000000000111010110111001001011110 10110000100000001010000101001001110110000001101001000101000001100 00010111001100010011101001001001100001100010110000100010100101000 00100010101001110010100100110011010000100100101100110001101010

Hopefully this doesn't screw up your web browser! haha