Proving the ATA/EVP Formula

[Starrz]

(Same as from official boards)

Something I've been interested in doing for a while now. Take note that this is not my formula, and I do not take credit for it.

I was presented with this simple formula indicating the relationship between ATA and EVP in PSU:

Miss Rate = EVP / (ATA*10)

Now, to show proof of it's accuracy:


-----------


Video #1

http://youtube.com/watch?v=EoiebnvHinw

In this example, we have the following:

Base ATA: 660
Python: 146 Acc
Burning Hit lv 30: 75% Acc.
Megistaride: 115% Acc.
Vanda Merha: 627 EVP

If we use the formula, we get:

627 / ((660+146) * 1.15 * .75 * 10) = Miss Rate

9.019% = Miss Rate

In the example, after 453 shots. We see that there have been 29 missed.
Resulting in a 6.401% Miss Rate.

Difference between the formula and the example: 2.618%

Video #2

http://youtube.com/watch?v=iH9sLDre7bw

In this example, we have the following:

Base ATA: 436
Falgahoh 154 Acc.
no PA
Zodiaride 115% Acc.

lv 105 Armed Servant (Kanohne). Approx. 792 EVP.

Using the formula we get:

792 / ((436 + 154) * 1.15 * 10) = Miss Rate

11.672% Miss Rate

In the example, after 115 shots. We see that there have been 9 missed.
Resulting in a 7.826% Miss Rate.

Difference between the formula and the example: 3.846%

------------

Video #3

http://youtube.com/watch?v=vSChFqXLK_0&feature=user

In this example, we have the following:

Base ATA: 98
Dalteri: 79 Acc.
No Buff
Raisei-sou lv 1: 53% Acc.

lv 25 Vahra: 240 EVP

If we use the formula we get:

240 / ((98 + 79) * .53 * 10) = Miss Rate

25.58% = Miss Rate

In the example, after 98 shots. We see that there have been 30 missed.
Resulting in a 30.6% Miss Rate.

Difference between the formula and the example: 5.03%

--------------

Video #4

http://youtube.com/watch?v=pddatggZ1m8

This one is a bit different as I needed someone else to do the shooting. So I'll be just observing someone else shooting. Thanks to CellarDoor for helping me out.

In this example, we have the following:

915 Base Acc.
244 Acc. Blackbull
No PA
No Buffs
lv 106 Svaltus Approx. 1975 EVP

For proof purposes:
http://i78.photobucket.com/albums/j101/ailleurs/ArmorEquip.jpg
http://i78.photobucket.com/albums/j101/ailleurs/BlackbullEquip.jpg
If we use the formula, we get:

1975 / ((915 + 244) * 10) = Miss Rate

17.04% = Miss Rate

In the example, after 242 shots. We see that there have been 44 missed.
Resulting in a 18.18% Miss Rate.

Difference between the formula and the example: 1.14%

Video #5

People said the samples weren't big enough, and they were right. Here's more samples in a video.

http://video.google.com/videoplay?docid=-5193128434752863795

Base ATA: 673
Python 146 Acc.
Burning Hit lv30 75% accuracy
Megistaride 115% accuracy

lv 85 Vanda Merha 627 EVP

Formula suggests a hit rate of 8.876%

1178 shots fired. 101 shots missed. Resulting in a 8.573% miss rate.


Difference between formula and example: .303%
--------------

So there you guys have it. Enjoy calculating Miss and Hit rates!



<font size=-1>[ This Message was edited by: Starrz on 2007-12-15 07:00 ]</font>


<font size=-1>[ This Message was edited by: Starrz on 2007-12-15 11:14 ]</font>
[u]

<font size=-1>[ This Message was edited by: Starrz on 2008-04-13 19:31 ]</font>

[DesertGunner]

The only thing I'm interested in is calculating rare drops

[AC9breaker]

Hey man that's some awesome stuff.
You gave me a half chubb with all that intel. (*'-')b

[ThEoRy]

Nice post. An interesting read and very informative.

[Bitey]

Was gonna post this before work, but did not have the time, so now I will do so.

Anyway, I do not see proof for a formula. All I see is some math done around a random formula with links to a few videos that prove your formula wrong with real miss rates that are far different than your calculations. If your formula was within 1% with almost every calculation (giving decimal numbers and what not) then I might be convinced you have something here, but instead your formula if off by double that usually and even in one case by 5%. 5% is a huge difference when actually calculating possible misses for DPS because that 5% is an extra 5 hits or more missed a minute on some guns...

On a side note though, your videos did help prove something useful. No matter how much ATA you pack on a character with natural stats and equipment, the biggest deciding factor in your hit rate is whether you are buffed with Zodial or not. Both of your videos buffed with Zodial had a much greater chance to hit the foe than the not buffed videos. Especially Video 1 compared to video 4. Video 4 had A LOT more accuracy with equipment and base stats than video 1 and video 1 even used a PA which lowered it's ATA even more while video 4 used no PA at all, yet video 4 ended up with a miss rate that was 3 times that of video 1.

[Starrz]

On 2007-12-15 21:52, Bitey wrote:
Was gonna post this before work, but did not have the time, so now I will do so.

Anyway, I do not see proof for a formula. All I see is some math done around a random formula with links to a few videos that prove your formula wrong with real miss rates that are far different than your calculations. If your formula was within 1% with almost every calculation (giving decimal numbers and what not) then I might be convinced you have something here, but instead your formula if off by double that usually and even in one case by 5%. 5% is a huge difference when actually calculating possible misses for DPS because that 5% is an extra 5 hits or more missed a minute on some guns...

On a side note though, your videos did help prove something useful. No matter how much ATA you pack on a character with natural stats and equipment, the biggest deciding factor in your hit rate is whether you are buffed with Zodial or not. Both of your videos buffed with Zodial had a much greater chance to hit the foe than the not buffed videos. Especially Video 1 compared to video 4. Video 4 had A LOT more accuracy with equipment and base stats than video 1 and video 1 even used a PA which lowered it's ATA even more while video 4 used no PA at all, yet video 4 ended up with a miss rate that was 3 times that of video 1.

Actually, the formula is very accurate. The reason for the difference gaps is that I can't possibly calculate the thousands of samples it would take to get close to the formula's expected hit rate. That, and the formula provides an average hit rate, while in the examples, they are exactly that, examples. Hitting and missing is still random, and you're obviously not going to get exactly average every time. :]

And regarding using zodial and not using zodial, you're not looking at the examples correctly. The reason why video 1's hit rate was so high compared to video 4 was that lv 85 Vanda Merha have roughly 1/3 the EVP of lv 106 Svaltus. For you see, changes in Acc. do not effect the miss rate as much as EVP does.

Adding on to that note. It is very important that people not compare the videos to one another. As I purposely made them very different. Using different classes, different Acc. values, different monsters, different weapons, different PAs, and different buffs. You just can't compare them. :]

[Pillan]

If you want some more accurate data verifying the formula, you can check it out here (http://suimasen.sakura.ne.jp/aaaphp/index.php?%A5%CD%A5%C3%A5%C8%A5%EF%A1%BC%A5%AF%A5% E2%A1%BC%A5%C9%2F%B8%A1%BE%DA#c6779cf6).

[Starrz]

I didn't think people would be this skeptical of facts. Anyways, here's another test using over 1000 samples.

http://video.google.com/videoplay?docid=-5193128434752863795


Base ATA: 673
Python 146 Acc.
Burning Hit lv30 75% accuracy
Megistaride 115% accuracy

lv 85 Vanda Merha 627 EVP

Formula suggests a hit rate of 8.876%

1178 shots fired. 101 shots missed. Resulting in a 8.573% miss rate.


Difference between formula and example: .303%