Old RA AM; An exploration of pixel math.

dblaha86
Myrmidon
 
Posts: 169
Joined: Jun 19, 2013 11:05

Postby dblaha86 » Jan 04, 2017 02:45

I thought this pixel math stuff made perfect sense at first. Looking at the video of AM that was analyzed in this thread here seemed to be perfectly accurate to determine this bards total hp. Then using that same method we could figure out how much AM ticks were healing for.

This picture decoding the AM ticks using that method:
Image

Then I thought I could be even more accurate and count pixels on a higher resolution of that same video at the very same moments. This would allow to zoom in even further on those health bars and get a more precise hp/pixel measurement. This is where I started to notice a problem. The healthbars are faded due to the poor video quality and it is anybodys guess where to actually start counting from. Do you start counting from where the fade completely ends or where the solid healtbar begins to fade? Just determining where the health bar actually ends will give you different numbers when you calculate it.
Image Image

Because of the fading health bar due to poor quality, any guesses where to start counting from would have a (+/-)3 pixel deviation in calculating how many pixels the 297 hp heal actually did. This severely affects the calculation of the bards total HP. If the heal is 25 pixels we calculate 1722 HP, if the heal is 22 pixels we calculate 1957 hp. Furthermore, the blur on the health bar is not constant so its always a different guess for each tick of AM making it even more difficult to determine tick values.

Most of the AM ticks in the pictures above look like 10 pixels which would mean 118hp AM ticks if he has 1722 hp, or 135hp ticks if he has 1957 HP. One of the examples(top right) looks like 8 which would mean as little as 95hp/tick @ 1722 hp, or 108hp/tick @ 1957 hp. The numbers are all over the place and it really seems this method is not reliable.

I decided to try and get a perfect resolution screenshot from Uthgard and see if using the pixel calculations could determine my known total HP using these same methods. I Dueled a player and then counted the healthbar. Screenshot. Even with high image quality the calculations are still off by nearly 100 HP.
Image

Not calculating the correct HP even when there is no quality issues with healthbar fades leads me to the conclusion that either Uthgards healthbars are innaccurate, or that this method as a whole is inaccurate and delivers a calculation of a lower total hp pool than the actual. This means that the AM tick calculations using this method are also lower than actual values. The values calculated for AM using this method have all been at least 100hp/tick and as high as 135hp/tick for me. This means to me that is not unreasonable to find that old AM actually ticks for 150, and if it is in fact only set to 75hp/tick currently that it is incorrect.

dblaha86
Myrmidon
 
Posts: 169
Joined: Jun 19, 2013 11:05

Postby dblaha86 » Jan 04, 2017 03:46

Also, note that in that analyzed video. The bard has Purple spec buffs and it is in toa. With extra bonuses to con and Purple buffbot buffs, his HP is likely greater than 2000, we are able to hit 1700 hp on classic uthgard settings with only base con buff, aug con 4, and toughness 2.

The aug con 4 is +24 con, which nearly negates the +26 possible con from TOA, yet he still has spec str/con buff which is going to increase by ~90 more con. He could have well over 2k hp if he has any +hit cap(up to 200 more?) toughness RA or Aug con on his character.

If his HP is over 2000, then AM is VERY likely ticking for ~150...

Additional HP test on uthgard:
On uthgard with purple druid buffs +151 total con, without any aug con or toughness. I am at 1842 HP. A normal TOA template would have 400/400 hit cap adding another flat 200, putting hp at 2042 without any increase to the con cap (a possible 101/101 with +26) or using any other RA's that would help such as aug con or toughness. 2k hp is very easily achievable in toa especially on a try-hard battlebard using a buffbot.

Heres another pixelmath undercalculation of total hp on uthgard with 1842 hp:
Image

After spending several hours today looking at this junk, I am thoroughly convinced that the video "prooving" the amount that AM should tick for made incorrect assumptions on the bards total hp and undercalculated the tick values. With this evidence presented here, you can accurately assume the bard in the video had 2000-2100 hp, thus putting those AM ticks at 150 hp which alligns with patch notes, comparisons to new RA AM, and player testimonials from 2003 about the RA presented in Issue 2447

Is there anything that prooves the ticks are only 100hp instead of 150 besides the flawed pixel math stuff?
Last edited by dblaha86 on Jan 06, 2017 07:05, edited 2 times in total.

User avatar
Cirdan
Eagle Knight
 
Posts: 527
Joined: May 19, 2016 22:48

Postby Cirdan » Jan 04, 2017 04:18

Instead of worrying about when the health bars begin and end, why not pixel math the difference in pixels between his heal of himself for 297 and the regen tick from AM? You can also compare it to the few times he takes damage before he heals himself.

User avatar
Abydos
Game Master
Game Master
 
Posts: 6636
Joined: Jan 22, 2011 21:14

Postby Abydos » Jan 04, 2017 04:42

FWIW be cautious of pixel math. Encoding/compression schemes can cause precision problems as regions of screens are ignored. Something to keep in mind :-p

dblaha86
Myrmidon
 
Posts: 169
Joined: Jun 19, 2013 11:05

Postby dblaha86 » Jan 04, 2017 05:17

Abydos wrote:FWIW be cautious of pixel math. Encoding/compression schemes can cause precision problems as regions of screens are ignored. Something to keep in mind :-p


Oh i absolutely agree, my whole claim here is that the pixel math is incredibly inaccurate and is consistently undervaluing total HP every time i check it with GOOD quality screenshots where i actually know the characters total HP.

dblaha86
Myrmidon
 
Posts: 169
Joined: Jun 19, 2013 11:05

Postby dblaha86 » Jan 04, 2017 05:36

Cirdan wrote:Instead of worrying about when the health bars begin and end, why not pixel math the difference in pixels between his heal of himself for 297 and the regen tick from AM? You can also compare it to the few times he takes damage before he heals himself.


--Oh you probably mean comparing the 297 heal to the AM ticks irregardless of total HP. I'll try doing that but I still dont see how that solves any issues with the accuracy, it will still have faded bar and rounding issues I believe.
Last edited by dblaha86 on Jan 04, 2017 06:23, edited 1 time in total.

User avatar
Cirdan
Eagle Knight
 
Posts: 527
Joined: May 19, 2016 22:48

Postby Cirdan » Jan 04, 2017 06:05

dblaha86 wrote:--Oh you probably mean comparing the 297 heal to the AM ticks irregardless of total HP. I'll try doing that but I still dont see how that solves any issues with the accuracy, it will still have faded bar and rounding issues I believe.


Yes, that's what I meant. If the faded bits at either end are part of the problem, this should get around that. Of course it's not perfect and you have to be careful and sometimes move frame by frame but it does help you get a general range.

dblaha86
Myrmidon
 
Posts: 169
Joined: Jun 19, 2013 11:05

Postby dblaha86 » Jan 04, 2017 07:45

Cirdan wrote:
dblaha86 wrote:--Oh you probably mean comparing the 297 heal to the AM ticks irregardless of total HP. I'll try doing that but I still dont see how that solves any issues with the accuracy, it will still have faded bar and rounding issues I believe.


Yes, that's what I meant. If the faded bits at either end are part of the problem, this should get around that. Of course it's not perfect and you have to be careful and sometimes move frame by frame but it does help you get a general range.


The image below has the first 2 healthbars showing the 297 hp heal. If you open the picture in MS Paint and zoom in you can see the fade I was talking about. Its not the 0% and 100% mark of the hp bar I was referring to, its the fade where the hp bar is at. You can see anywhere from 1-4 pixel fade to where the current hp bar value 'ends'. It is hard to tell if its 23 pix or even up to 25, and this happens for each of the AM ticks as well. Then the AM ticks its hard to determine if they are anywhere from 7-10 pix.

But heres a rough calculation anyways with this image...
Heal = 297/24pixels = 12.375 hp/pix
1st AM tick = 9 pix*12.375 = 111.375
2nd AM tick = 7 pix*12.375 = 86.625
3rd AM tick = 9 pix*12.375 = 111.375
4th AM tick = 9 pix*12.375 = 111.375
==AVG== 105.18 hp/tick

Comparison Directly to heal image:
Image

You are correct that we dont need a total hp value to try to estimate the tick values. The total hp calculation can just be considered an additional calculation which shows how the pixel counting is flawed. The numbers just don't add up and there is something else going on with the HP bars that cannot be counted by pixels. The ToA bard with red buffs is being calculated to have only 1726 hp when he more likely has over 2000 hp and my uthgard character is only calculating to 1675 hp when he really has 1842. It makes me question how accurate comparing anything with the pixel counting really is, whether its an AM tick to that 297 heal or an AM tick compared to estimated total health, it's flawed. Seeing as the calcs keep undervaluing total hp it makes me feel like it is transitive to calculating the am ticks low as well. This calculation happens on TOA bars as well when i can see each pixel perfectly and there is no fading.

If you consider that same image and assume 2050 total HP; (still uncertain of actual AM tick pixel count)
Heal = 297/21pixels = 14.143 hp/pix
1st AM tick = 9 pix*14.143 = 127.287
2nd AM tick = 7 pix*14.143 = 99.001
3rd AM tick = 9 pix*14.143 = 127.287
4th AM tick = 9 pix*14.143 = 127.287
==AVG== 120.22 hp/tick

Idk, all this pixel counting is just goofy ;0

User avatar
Thalien
Phoenix Knight
 
Posts: 1016
Joined: Apr 18, 2009 00:00
Location: Marburg

Postby Thalien » Jan 04, 2017 09:19

Think you can't rely on this. I mean find values by the bars.
I tried some research for power per acuity.

Pendragon Casting red shield for 31 power.

50 ava caby 135 int -> 95% power after cast -> 620 power? 31pt=5%?
50 sara caby 115 int -> 90% power after cast -> 310 power? 31pt=10%?

so 20 int means half the power. clearly not

:?: :?:
Fabienne - Blademaster
Junia - Bard
Audrina - Ench
-----
Zoe - Sorc
Stella - Pala
Devon - Scout

User avatar
GreenP
Phoenix Knight
 
Posts: 1303
Joined: Sep 07, 2008 00:00

Postby GreenP » Jan 04, 2017 12:36

Cirdan wrote:Instead of worrying about when the health bars begin and end, why not pixel math the difference in pixels between his heal of himself for 297 and the regen tick from AM? You can also compare it to the few times he takes damage before he heals himself.


Thats what i tried here.
viewtopic.php?p=382340#p382340
Formerly known as
Touareg- RR10 noobVW on Uth1.
Crossfire- RR7 noobAnimist on Uth1.
Ohnoes- RR9 noobAnimist on Lolgins.

Click this LINK if you want to see the truth about Hibernia

Shinyuka
Myrmidon
 
Posts: 152
Joined: Oct 19, 2013 13:15

Postby Shinyuka » Jan 05, 2017 01:45

Cirdan wrote:Instead of worrying about when the health bars begin and end, why not pixel math the difference in pixels between his heal of himself for 297 and the regen tick from AM? You can also compare it to the few times he takes damage before he heals himself.


Sorry i don't get that.. He spent alot of effort into explaing his point of view and why he thinks pixel math is NOT accurate... and your reply is something like: "yea yea whatever, why dont u just pixelmath.." WOW!

That appears very disrespectful to me, and somewhat medival-ish.


Return to Bard

Who is online

Users browsing this forum: No registered users and 1 guest

Friday, 24. January 2020

Artwork and screen shots Copyright © 2001-2004 Mythic Entertainment, Inc. All rights reserved. Used with permission of Mythic Entertainment. Mythic Entertainment, the Mythic Entertainment logo, "Dark Age of Camelot," "Shrouded Isles," "Foundations," "New Frontiers," "Trials of Atlantis," "Catacombs," "Darkness Rising," the Dark Age of Camelot and subsequent logos, and the stylized Celtic knot are trademarks of Mythic Entertainment, Inc.

Valid XHTML & CSS | Original Design by: LernVid.com | Modified by Uthgard Staff