+1% Crit Arrow Myth

OMGLAZERZ - Heavens Tear

Marutah - Harshlands wrote: »

Frequency of crits over a long period of time, and Chance to crit (which is what the crit% is) are 2 diffrent things..

i think were both just taking the crit% as diffrent things..

i highly doubt the game devs made it so that 6221 hits was the equalizer in crit precentage calculations.. if its a static unchangeable number like you say, the precent should be present no matter the sample size.. be it 1million hits or 1000 hits..

and crit% being the chance per hit makes total sense, as you would be able to see the results as more random (ie: spiking up and down, just as your graph does)

if it was your way, the graph should have flattend out way earlier and never have gone above the listed %... make up all the fancy graphs and **** you want..

just because it looks pretty dosent make it correct.

One can lead a horse to water, but they cannot make them drink.

Granrey - Sanctuary

OMGLAZERZ - Heavens Tear wrote: »

I want you to see something that will change you mind. This is taken from the data as it came in.


http://img300.imageshack.us/img300/4259/critrate5.jpg


I can throw numbers at you all day long, but this is the way it is: over the long term you will see the crit rate show up. There is your proof.


387 crits over 7539 attacks


Perfectly logged with no mistakes.


Doubt me now?

I still not see any info showing is not a chance. However, finding out is pointless since we can see that after 7000 hits the crit% becomes stable (chance or not). 7000 hits is good mark for you to check the arro myth.

All you have to do is 7000 hits with arrow% and 7000 without one.

Edit:

Since you did your test to a number close to 7000 with arrow and without one. I can say your results are correct IMO. (based on your first post)

Marutah - Harshlands

a) im not a horse
b) your not some all knowing guru.. everything you have posted is all based off one loosly defined 'logic' by you..

i still find it hard to belive that you dont accept the '% to crit' is just that.. % to crit per hit..

but instead you think its % of crits over the course of 6,221 hits.. it boggles the mind..


ps: im done with the thread, ill quit buggin ya and let you do your calcs.. all in all its pretty obvious that weither the % is in rerguards to 'per hit' or 'over time' like were debating the crit arrow is Not adding anything but a visual glitch..

ps: omg, you said your a heavy/arcane hybrid foxy? hows it working for ya? i have a heavy/arcane in her 30's, just been lazy leveling her.. its kinda nice so far though

Shiga - Sanctuary

Probably won't convince anyone who is completely mired in their ways, but for those wondering how it actually works, you can brush up on your understanding of the law of large numbers

OMGLAZERZ - Heavens Tear

Marutah - Harshlands wrote: »

a) im not a horse
b) your not some all knowing guru.. everything you have posted is all based off one loosly defined 'logic' by you..

i still find it hard to belive that you dont accept the '% to crit' is just that.. % to crit per hit..

but instead you think its % of crits over the course of 6,221 hits.. it boggles the mind..

And I didn't call you a horse. It's a saying...?


You said the other was a pretty picture, but it didn't drop fast enough, but it was only the latest trial I've run.

Here's one from one I did with my mage veno who has a stated 3% crit rate.

http://img5.imageshack.us/img5/1728/critrate3.jpg

263 crits over 8831 Venomous scarabs cast.

2.98% crit rate showing

This one seems to rise quickly to 3% and stay there, and spells don't have a chance to miss.

On the Heavy/arcane/hybrid veno...it's hard, and I have to restat away about 15 dex to fix her since I'm struggling to get her Magic damage up, and her into level 5 heavy armor.

Semyezza - Heavens Tear

-sigh- Crit arrows only work when you SHOOT them. Otherwise, why would they be on arrows? An Archer wouldn't be able to use it, since they'd use them up. -headdesk-

It's a cute myth, though. <_<"

OMGLAZERZ - Heavens Tear

Semyezza - Heavens Tear wrote: »

-sigh- Crit arrows only work when you SHOOT them. Otherwise, why would they be on arrows? An Archer wouldn't be able to use it, since they'd use them up. -headdesk-

It's a cute myth, though. <_<"

It would be even cuter if people weren't selling them for 5k each which means people see it, and go "ooo...more crit!" which is why I started this to begin with meaning a lot of people probably don't realize it's just a myth since I doubt archers are spending tons to buy them.

Rule - Heavens Tear

The primary issue is that the character sheets display an additional 1% when the crit arrow is in your ammo slot, even though a bow is not wielded.

Either way this represents a glitch, either that it gives one, or that it doesn't but displays it.

There is no conclusive evidence of either a period of hits during which the % of crits is achieved before reset or whether it is on a per hit basis. Either way in a large enough sampling it does not matter. In a 7k sample with zero chance to miss it seems LIKELY that the numbers represented would cover both possiblities. I'll leave the proof of that to the more mathmatically inclined.

The results of this ARE important. If as a wiz I get an additional 1hko over a period of time, then I save mana and grind time. Though HOW important is questionable, but I'll take every single percent I can get. At 10% now.

AndromedaB - Sanctuary

how did u record so many hits?
and make that graph thing
i wana try it on my mage

AndromedaB - Sanctuary

how did u graph so many hits???????????????
i wana try it on my mage

OMGLAZERZ - Heavens Tear

AndromedaB - Sanctuary wrote: »

how did u record so many hits?
and make that graph thing
i wana try it on my mage

Pen and paper to jot the numbers down, entered them into a spreadsheet, created a table based off the data, then a graph using the table. Took a screen shot, and edited it in GIMP 2.4.

Asterelle - Sanctuary_1381265973

OMGLAZERZ - Heavens Tear wrote: »

I want you to see something that will change you mind. This is taken from the data as it came in.


http://img300.imageshack.us/img300/4259/critrate5.jpg


I can throw numbers at you all day long, but this is the way it is: over the long term you will see the crit rate show up. There is your proof.


387 crits over 7539 attacks


Perfectly logged with no mistakes.


Doubt me now?

This graph is very telling. If you were to ask me after the first 2000 hits what the crit rate really was I would guess it was aroung 6%. It's not until about 6,000 points that it becomes clear the real crit rate is 5%.

I thought I would look at some statistics for this...

For a binomial distribution of n points with probability p the standard deviation is
s = sqrt(p(1-p))/sqrt(n) so
n = p(1-p)/s^2

If we set s ~ 0.5% and p~3%
n = 1164

What this says is that after 1164 points we have a 68% confidence that the measured probability is within 0.5% of the actual value. Here are other confidence intervals.

1164 points = 68% confidence of being within 0.5%
4656 points = 95% confidence of being within 0.5% (also 68% confidence of being within 0.25%)
10560 points = 99% confidence of being within 0.5% (also 95% confidence of being within 0.33%)


I was wondering if you could please summarize your data for:

Displayed crit rate with no crit arrow:
Measured crit rate with no crit arrow:
Number of datapoints with no crit arrow:

Displayed crit rate with crit arrow:
Measured crit rate with crit arrow:
Number of datapoints with crit arrow:

For both I would say any measurement with > 6,000 data points has a high confidence of being right.

-Ast

AndromedaB - Sanctuary

Asterelle - Sanctuary wrote: »

This graph is very telling. If you were to ask me after the first 2000 hits what the crit rate really was I would guess it was aroung 6%. It's not until about 6,000 points that it becomes clear the real crit rate is 5%.

I thought I would look at some statistics for this...

For a binomial distribution of n points with probability p the standard deviation is
s = sqrt(p(1-p))/sqrt(n) so
n = p(1-p)/s^2

If we set s ~ 0.5% and p~3%
n = 1164

What this says is that after 1164 points we have a 68% confidence that the measured probability is within 0.5% of the actual value. Here are other confidence intervals.

1164 points = 68% confidence of being within 0.5%
4656 points = 95% confidence of being within 0.5% (also 68% confidence of being within 0.25%)
10560 points = 99% confidence of being within 0.5% (also 95% confidence of being within 0.33%)


I was wondering if you could please summarize your data for:

Displayed crit rate with no crit arrow:
Measured crit rate with no crit arrow:
Number of datapoints with no crit arrow:

Displayed crit rate with crit arrow:
Measured crit rate with crit arrow:
Number of datapoints with crit arrow:

For both I would say any measurement with > 6,000 data points has a high confidence of being right.

-Ast

am i the only one in highschool -_-

Nayiro - Sanctuary

AndromedaB - Sanctuary wrote: »

am i the only one in highschool -_-

I learned this last semester, and I'm in highschool xD

Llama - Lost City

You have a limited understanding of odds and percentages. Even if you have 99% crit rate you can still not crit more than twice out of 2 attacks. Its a 3% chance per attack which means every attack you have a very small chance of critting.

OMGLAZERZ - Heavens Tear

Llama - Lost City wrote: »

You have a limited understanding of odds and percentages. Even if you have 99% crit rate you can still not crit more than twice out of 2 attacks. Its a 3% chance per attack which means every attack you have a very small chance of critting.

I'm sorry, but who is this directed at?

WarrenWolfy - Sanctuary

Marutah - Harshlands wrote: »

b) your not some all knowing guru.. everything you have posted is all based off one loosly defined 'logic' by you..

He may not be a guru, but he's absolutely right.

OMGLAZERZ - Heavens Tear

In the top set, the % of crits rises quickly to ~6%, and then starts dropping to ~5% from about 1300 attacks in rising and falling but staying within a few tenths of a percentage of 5%. The top set was also done with both a Christmas Blessing and the Elite Wolf Fang arrows equipped for what would have been a 6% crit% which, as we can see, it clearly was not.


In the bottom set, the % of crits rises quickly to ~6% as well, and stays there with a initial drop at ~1800 attacks in to 5.4% before rising back to ~6% until 2400 attacks in at which point it starts dropping to 5% where it stayed within a few tenths of a percentage of 5%. The bottom set was done with just the Christmas Blessing for a 5% crit% which, as we can see, it clearly was.

To avoid confusion on the graphs as to why there are two lines (red and black).

Red line: % of successful attacks that were crits.

Black line: % of ALL attacks that were crits.


The red line is only important if the melee attack critical is only determined upon a successful attack which is, at this time, unknown. This is why I actually would have preferred doing this by casting spells, but auto-attacks render more data faster.

This also means that the bottom graph where it looks like it's sitting at 6% excessively long before dropping: it actually isn't if you follow the black line which is at ~5.7% and doesn't actually go back up to 6% after falling below that at ~1000 attacks in.



Mob: Rhinodrake Megalith
Number killed: ~625

Total attacks : 6058
Crits : 281
Misses : 293
Non-crits : 5484

Crit% of all attacks: 4.64%


Expected crit% : 6% (base 4%+Christmas Blessing+Elite Wolf Fang Arrow)

Difference: 1.36%



Mob: Frail Blob
Number killed: ~825
Crits: 387
Misses: 378
Non-Crits: 6774
Total 7539

Act Crit % 5.13

Expected Crit%: 5% (base 4% + Christmas Blessing)

MissEbil - Heavens Tear

But the point stands:
1% isn't that much of a chance of a crit anyway.
PLUS: 1% Arrow = 1k, Blessing = around 2mil

Arrow > Blessing

OMGLAZERZ - Heavens Tear

MissEbil - Heavens Tear wrote: »

But the point stands:
1% isn't that much of a chance of a crit anyway.
PLUS: 1% Arrow = 1k, Blessing = around 2mil

Arrow > Blessing

Except the Arrow doesn't do anything without a bow equipped.


Don't you read the charts?


The top one was with BOTH the Elite Wolf Fang arrow equiped AND the Christmas Blessing Equipped.


The bottom one was with JUST the Christmas Blessing Equipped.


It follows logically that the arrow did not actually increase the chance to crit by 1% like the Christmas Bless did.

FYI, the base crit chance for my venomancer is 4%.

Christmas Blessing > Elite Wolf Fang arrow.


I'm sure I have data on that spread sheet detailing just the Elite Wolf Fang arrow equipped. Let me check really quick.


Here's the end numbers I had on the Elite Wolf Fang arrows, to clear this up.

I realize the total set sizes are not large, but it's what i have.

Crits : 118
Non-crit hits: 3352
Misses: 194
Total Attacks: 3664
Crit% 0.0322


So, with just the Elite Wolf Fang arrow equipped, she hit a 3.22% crit rate which is below her base crit rate of 4% and 1.78% below the 5% crit% showing on her character sheet.



So, without the Christmas Blessing on, and just the Elite Wolf Fang arrow, you can clearly see that it's not increasing the crit %.

In fact, I'd say it was driving it down to 3%.

Chart coming soon to a post near you. b:pleased

MissEbil - Heavens Tear

The charts made my head hurt.

What I was trying to say, is I'm willing to waste 1-2k on an arrow on the CHANCE it might work. I can't afford a Blessing because people are... people, and greedy.

OMGLAZERZ - Heavens Tear

MissEbil - Heavens Tear wrote: »

The charts made my head hurt.

What I was trying to say, is I'm willing to waste 1-2k on an arrow on the CHANCE it might work. I can't afford a Blessing because people are... people, and greedy.

A chance it might work?


There is no chance it might work unless you have a bow equipped.


It's like carrying around a rabbit's foot because it's "good luck". It doesn't really bring good luck, just warm fuzzy feelings to make you feel better. Nothing more, nothing less.

FateBlade - Lost City

Really cool, when get +1% critical for free (from arrow). b:victory

Smexxyfox - Heavens Tear

lmao somone didnt read the thread at all..

Its intresting tho how it seems like you actually get a lower crit rate with the arrow and without it you managed slightl above the expected amnt.

OMGLAZERZ - Heavens Tear

FateBlade - Lost City wrote: »

Really cool, when get +1% critical for free (from arrow). b:victory

There is no +1%. It's a visual glitch.

Smexxyfox - Heavens Tear wrote: »

lmao somone didnt read the thread at all..

Its intresting tho how it seems like you actually get a lower crit rate with the arrow and without it you managed slightl above the expected amnt.

The red line, the one you seem to be referring to, isn't of all the attacks, that's the black line, the red line shows the % of just successful attacks. I'd have to run the numbers out even further to see how they act at 10k, but I don't have the time, well, I just don't feel like it right now.

FateBlade - Lost City

So, better dont use +1% critical arrow? b:shocked

OMGLAZERZ - Heavens Tear

FateBlade - Lost City wrote: »

So, better dont use +1% critical arrow? b:shocked

That's my conclusion.

FateBlade - Lost City

You sure? How you know that?

Granrey - Sanctuary

FateBlade - Lost City wrote: »

You sure? How you know that?

He already proved it, see the graph.

FateBlade - Lost City

But Blessing +1% crit. is good, right?