Item Search
     
BG-Wiki Search
+ Reply to Thread
Page 1 of 12 1 2 3 11 ... LastLast
Results 1 to 20 of 231

Thread: Cast and Recast     submit to reddit submit to twitter

  1. #1
    An exploitable mess of a card game
    Join Date
    Sep 2008
    Posts
    13,258
    BG Level
    9
    FFXIV Character
    Gouka Mekkyaku
    FFXIV Server
    Gilgamesh
    FFXI Server
    Diabolos

    Cast and Recast

    Gradually shifting away from the Data Compilation thread into multiple threads so as to better organize the information. Post Haste related stuff here and link in the OP. It will combine the ease of posting new data, but maintain organization.
    Haste

    General
    Abilities
    Magic
    Equipment
    Main/Sub/Ranged/Ammo

    Helm

    Neck

    Ear

    Body

    Hands

    Rings

    Back

    Waist

    Legs

    Feet

  2. #2
    An exploitable mess of a card game
    Join Date
    Sep 2008
    Posts
    13,258
    BG Level
    9
    FFXIV Character
    Gouka Mekkyaku
    FFXIV Server
    Gilgamesh
    FFXI Server
    Diabolos

    Slow

    General
    Abilities
    Magic
    Equipment
    Main/Sub/Ranged/Ammo

    Helm

    Neck

    Ear

    Body

    Hands

    Rings

    Back

    Waist

    Legs

    Feet

  3. #3
    An exploitable mess of a card game
    Join Date
    Sep 2008
    Posts
    13,258
    BG Level
    9
    FFXIV Character
    Gouka Mekkyaku
    FFXIV Server
    Gilgamesh
    FFXI Server
    Diabolos

    Cast Speed

    General
    Abilities
    Traits
    Magic
    Equipment
    Main/Sub/Ranged/Ammo

    Helm

    Neck

    Ear

    Body

    Hands

    Rings

    Back

    Waist

    Legs

    Feet

  4. #4
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Casting Speed remains difficult to quantify (fraps anyone?), but I think Recast has been pretty much nailed. As far as I can tell, Recast is a floating point operation that uses a bunch of low-bit-size variables.
    https://www.bg-wiki.com/bg/Spell_Recast

    If anyone sees anything to add, well, it's a wiki.

  5. #5
    TSwiftie
    Join Date
    Oct 2005
    Posts
    1,920
    BG Level
    6
    FFXI Server
    Fenrir

    Quote Originally Posted by Byrthnoth View Post
    Casting Speed remains difficult to quantify (fraps anyone?), but I think Recast has been pretty much nailed. As far as I can tell, Recast is a floating point operation that uses a bunch of low-bit-size variables.
    https://www.bg-wiki.com/bg/Spell_Recast

    If anyone sees anything to add, well, it's a wiki.
    When it comes down to Spell Casting Speed and Attack Speed, there are a few problems I've seen.

    -There is innate variance. For example, casting one spell over and over again will yield different completion times.
    -There's a high probably and some evidence that there are "tiers" in which this is visible. Easiest example is my Slow II testing which showed no difference from 1 and 2 merits, but did show a difference on the third merit.

    I think general testing in these areas are great, but it would take a big breakthrough to get precise testing like we have with recast values. That said, I think there might be a way to do this by analyzing the data that is sent to/from server. If I recall, attack animation speed varies based on attack delay.

  6. #6
    Failed Sex Ed
    Join Date
    Aug 2007
    Posts
    2,391
    BG Level
    7

    The different completion times for cast speed are probably due to internets lag. I have always seen cast times to be consistent +-2 frames except on rare occasions. All of my testing has been with cures.

  7. #7
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    I do like Shaddix' method for testing Casting Time the best, as it halves the possible lag and doesn't require fraps.

    As a review, I think:
    1) Get two players, one with a known and one with an unknown quantity of Casting time gear and Cure.
    2) Gear swap one of the two characters so that they have less than max HP.
    3) Make a script that casts Cure 3 on the lowered HP guy with both characters (via Multisend?)
    4) Run the script, note which one cures for HP.
    5) Repeat steps 2-4 until you're pretty sure which one casts faster more often. (idk how reliable this would be, as I've never actually done this)
    6) Try and move the "known" casting time guy closer to the "unknown" guy.
    7) When they go 50/50, you've found the Fast Cast value.

    You could use a script to automate steps 2-5, really.

  8. #8
    TSwiftie
    Join Date
    Oct 2005
    Posts
    1,920
    BG Level
    6
    FFXI Server
    Fenrir

    Quote Originally Posted by Byrthnoth View Post
    I do like Shaddix' method for testing Casting Time the best, as it halves the possible lag and doesn't require fraps.

    As a review, I think:
    1) Get two players, one with a known and one with an unknown quantity of Casting time gear and Cure.
    2) Gear swap one of the two characters so that they have less than max HP.
    3) Make a script that casts Cure 3 on the lowered HP guy with both characters (via Multisend?)
    4) Run the script, note which one cures for HP.
    5) Repeat steps 2-4 until you're pretty sure which one casts faster more often. (idk how reliable this would be, as I've never actually done this)
    6) Try and move the "known" casting time guy closer to the "unknown" guy.
    7) When they go 50/50, you've found the Fast Cast value.

    You could use a script to automate steps 2-5, really.
    It's definitely a unique way of testing, but it seems you're adding more lag variables, not removing them. Casting on two players at once doesn't always fire at once. (Best example off the top of my head is when I did Kraken BC while dualboxing 2 BLM. The BLM who fired first always varied between me + my mule.)

    @ Shaddix; I haven't done much FRAPS testing on spell cast times. I'm just not a big into data that doesn't involve concrete values. (Slow testing taught me how incredibly annoying it is.) If you say there's little variance though, it's possible I'm just wrong. As for Attack Delay animations, I won't budge on that being just too difficult!

  9. #9
    Failed Sex Ed
    Join Date
    Aug 2007
    Posts
    2,391
    BG Level
    7

    The testing with two characters is the only way I could figure out how to get around the issue of cast animation "tiers". Casting time isn't tiered, but whenever you are casting a fast spell like cure, you only see a change in speed every x% of additional fastcast. It is still however, casting faster server-side, but the animations will be identical.

    It does seem a bit shaky testing with two characters like that. I am not sure if you would be able to identify a 1% FC difference with that method. I will test and see if loq vs no loq is easily verifiable doing it that way. If so then it would at least be accurate to 2%. Maybe I can figure out some combination of gear to test a 1% difference.

  10. #10
    Salvage Bans
    Join Date
    Jul 2009
    Posts
    964
    BG Level
    5
    FFXI Server
    Leviathan

    Quote Originally Posted by Byrthnoth View Post
    5) Repeat steps 2-4 until you're pretty sure which one casts faster more often. (idk how reliable this would be, as I've never actually done this)
    Not very. You can't tell the difference in a lot of cases even with known different fast cast values. Whether that is because the difference is just too small or because fast cast/cure cast does in fact have a limited precision (what everyone calls breakpoints) I was not patient enough to do thousands of casts to find out.

  11. #11
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Quote Originally Posted by Kirschy View Post
    It's definitely a unique way of testing, but it seems you're adding more lag variables, not removing them. Casting on two players at once doesn't always fire at once. (Best example off the top of my head is when I did Kraken BC while dualboxing 2 BLM. The BLM who fired first always varied between me + my mule.)
    Yeah, that's why you'd have to repeat the test multiple times. As long as there is some relationship between Fast Cast and casting time, it would come out in the average. The question is how long you'd have to make your script. Obviously you'd need to do a control where they have a known and equal casting time to confirm that the script itself wasn't adding a substantial difference, and correct for it in the script if it was.

    I say you're removing lag variables, because:
    1) I look like I start casting -> Server -> Client -> I look like I'm finishing casting and analyze it via FRAPs
    2) I look like I start casting -> Server -> Cure HP is assigned -> Client

    So you're cutting out one step of lag and measuring the relevant quantity (when the spell lands according to the server).

    If you felt like doing this with something like Snapshot, you could do a similar thing with two rangers (equal delay/variable snapshot with one known) and level 0 monsters. Whoever kills the monster shot first. Same idea.

  12. #12
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Well, myth busted and looks like Turtle was right.

    No Fast Cast on either player, both BLM/WHM casting Cure 4: 18 Main, 4 Mule <--- Okay, so probably not the best test.
    5% Fast Cast on mule, both BLM/WHM casting Cure 4: 6 main, 4 mule

    So you'd either need a huge sample size to figure things out this way, or there actually isn't a difference between the two cases server-side.

    Spoiler: show
    Code:
    input /equip ring1 "Dark Ring"
    wait 1
    input /equip ring1 "Rajas Ring"
    send Sjugy "/ma cure4 byrth"
    cure4 byrth
    This is with default.xml loaded in spellcast

  13. #13
    Failed Sex Ed
    Join Date
    Aug 2007
    Posts
    2,391
    BG Level
    7

    That was not my experience with it. Testing a 5% difference showed 90% of the time the higher value landing the cure first. I believe I was testing 40% vs 45%. So there is some issue with that testing method or something else is going on. I will repeat the test I did years ago here in a few minutes.

    edit:Yeah it is still working perfect for me.

    Method:
    2x 75rdm(level synced)
    Jennica: AF1+1 hat, AF2 body, loq earring
    Shaddix: AF1+1 hat, AF2 body
    Alias m1g1 key on g15 keyboard to input /ma cure jennica on both instances
    unload spellcast
    Jennica equips and unequips ether ring.
    press m1g1 causing both characters to cast cure1 simultaneously

    Jennica cures for 30hp 9 times. Shaddix cures for 30hp 1 times.


    I also unequip loq on jennica just to remove any chance it was my machine somehow making jennica land them first. And got 5 landed 30hp from shaddix and 4 from jennica.

    And this once:
    http://dl.dropbox.com/u/2629723/fc.jpg

    no idea how that happens..

    either way though, my method is sound down to 2% fastcast difference.

  14. #14
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Well, that's pretty awesome. How do you have it set up? My way was way in the favor of my main (it's in the spoilers above).

  15. #15
    Failed Sex Ed
    Join Date
    Aug 2007
    Posts
    2,391
    BG Level
    7

    Forgot to say was in chocobo circuit. I use g15binder plugin for windower on both instances, it allows me to press one button to trigger the same action for both characters. Spellcast needs to be unloaded since it adds delay. Ideally you would make an alias for one of the gkeys on the keyboard that does /raw /ma cure character. But for some reason I couldn't get /raw to work with it so I just had to go with /ma cure character.

    edit: This method probably would work fine down to 1% fastcast if you were to use something like blm/sch light arts + quake.

  16. #16
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    I play on a laptop, so I'm not sure that route is open to me. Raksha did some testing with the more obscure pieces of SCH machinery: https://www.bg-wiki.com/bg/Talk:Light_Arts


    Though I'd love for someone to confirm the assumption, right now I'm assuming that SCH Job Ability recast is like Fast Cast with a 1:1 ratio of casting time to recast reduction.

  17. #17
    Groinlonger
    Join Date
    Oct 2006
    Posts
    2,964
    BG Level
    7
    FFXI Server
    Fenrir

    It looks that way. Might be interesting to throw some Fast Cast gear into the mix and see if they stack.

  18. #18
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Quote Originally Posted by Mojo View Post
    It looks that way. Might be interesting to throw some Fast Cast gear into the mix and see if they stack.
    I did more complicated cases when I was trying to figure out an order for the main Spell Recast page. It looks like everything is just multiplied together as separate terms.

  19. #19
    Sea Torques
    Join Date
    Jul 2009
    Posts
    679
    BG Level
    5
    FFXI Server
    Leviathan

    Quote Originally Posted by Byrthnoth View Post
    I play on a laptop, so I'm not sure that route is open to me. Raksha did some testing with the more obscure pieces of SCH machinery: https://www.bg-wiki.com/bg/Talk:Light_Arts


    Though I'd love for someone to confirm the assumption, right now I'm assuming that SCH Job Ability recast is like Fast Cast with a 1:1 ratio of casting time to recast reduction.
    Not sure if this should belong with the LA/DA information, but Accession/Manifestation nullifies the effects of LA/DA, including the bonuses from Argute Mortarboard and Scholar's Loafers.

  20. #20
    BG Content
    Join Date
    Jul 2007
    Posts
    21,105
    BG Level
    10
    FFXI Server
    Lakshmi
    Blog Entries
    1

    Thanks, I added sprinkled the information throughout the wiki. What happens when you Accession+Celerity something? "Normal" recast? With Argute Loafers/weather and Celerity/Accession, is it -10% or -20%? That should tell us if they're separate terms.

+ Reply to Thread
Page 1 of 12 1 2 3 11 ... LastLast