Welcome
Welcome to darkrisings

You are currently viewing our boards as a guest, which gives you limited access to view most discussions and access our other features. By joining our free community, you will have access to post topics, communicate privately with other members (PM), respond to polls, upload content, and access many other special features. In addition, registered members also see less advertisements. Registration is fast, simple, and absolutely free, so please, join our community today!

Trip / Entangle Testing and Results

Testing data for the game

Posts: 320

Joined: Mon Jun 09, 2014 3:47 pm

Post Sun Aug 31, 2014 12:31 am

Trip / Entangle Testing and Results

I have two different days worth of ability testing written up, with the first day being far more bare bones than the second (I THOUGHT of the script I used to scrape the damage data while watching the first days trials run). I'll say this both up here and at the bottom - please feel free to challenge my conclusions, these are but my musings scribbled down generally far too late at night for my own good.

Additionally, I am taking requests for future sets of combat data to capture and analyze. I'll have some baseline testing I'll want to do, but figure there are plenty of things I can compare to a relevant control sample.

I apologize ahead of time for the massive wall of text.


----------------------------------------------------------
DAY ONE TESTING

Trip Testing 08-25-14


Background:
Two test characters were used, both were Kine Rangers (class has meaning in second round of testing. Per Kiirion, I was asked to test four conditions for trip testing: Character A and B both grounded, Character A Flown with Character B Grounded, Character A Grounded with Character B Flown, and with both A and B both flown.

Attacker’s trip, weapon skill, dodge, parry and shield block were at 100. Defenders trip skill was at 0.

Results:

Control Group - No Fly
Trip landed 118 times out of 118 attempts (100%) [0 misses]

Test Group 1: Attacker Flown, Defender Grounded
Trips landed 150 times out of 152 attempts (100%) [0 misses]

Test Group 2: Attacker Grounded, Defender Flown
Trips Landed 527 out of 639 attempts (82.47%) [112 misses]

Test Group 3: Attacker Flown, Defender Flown
Trips landed 416 out of 504 attempts (82.54%) [88 misses]

Data interpretation:

The attacker being flown is irrelevant to the equation based on current data. I stopped collecting data after the first 100-150 attempts as there wasn’t a single missed attempt, so there wasn't as much reason to keep going.

Fly gives you roughly an 18-20% defensive rate against hostile trips, based on current data. More iterations could help determine if the ranger is closer to 18 or 20, but currently it seems unnecessary.

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

Entangle Testing 08-25-14

Background

This had similar testing parameters as trip did. Character A and B both grounded, Character A Flown with Character B Grounded, Character A Grounded with Character B Flown, and with both A and B both flown.


Attacker’s entangle, weapon, dodge parry and SB were 100, and the defender had entangle set at 100 as well.

Results

Control Test - All Grounded
Entangle landed 222 out of 305 attempts (72.79%) [83 misses]

Test Group 1: Attacker Flown, Defender Grounded
Entangle landed 214 out of 316 attempts (67.72%) [102 misses]

Test Group 2: Attacker Grounded, Defender Flown
Entangle landed 256 out of 323 attempts (79.26%) [67 misses]

Test Group 3: Attacker Flown, Defender Flown
Entangle landed 225 out of 384 attempts (58.59%) [159 misses]

Data Interpretation

The various configurations of flying DID change the land percentage a bit, with a grounded attacker against a flown defender having the most success, while all parties involved being flown really caused a significant downward deviation. With the potential for a MUCH higher damage knocked associated with entangle (I’m already thinking of how to capture and compile that data), a lower to hit percentage is expected, but is the tradeoff between trip worth the lower chance to actually knock your opponent down AND is the damage worth it (again, see above, this data will be compiled in the future).


Ranger vs Ranger does appear to change how entangle works (similar to how monk vs monk changes how armthrow works). A missed entangle will cause the message ‘You trip on one of your own vines!’ to happen, although the attacking ranger will NOT be tripped and is only subject to the normal lag associated with the use of entangle. Is this a bug (if not, the output message should be changed to reflect the fact that the attacker is NOT being tripped, just being hurt).

Oddly enough, there were only TWO instances of entangle flat out missing a target against another ranger, and it only occurred when both the attacker and defender were flying (these were only noted as they were unforeseen outcomes that broke my script’s tracking on both occurrences). [2 out of 384 attempts, 0.52%]


----------------------------------------------------------
Last edited by Some Guy on Mon Sep 01, 2014 11:52 am, edited 1 time in total.
Yeah? Well, you know, that's just like, your OPINION, man.

Posts: 320

Joined: Mon Jun 09, 2014 3:47 pm

Post Mon Sep 01, 2014 12:10 pm

Re: Trip / Entangle Testing and Results

DAY TWO TESTING

Background

This is a similar set of parameters to the first part, except for two exceptions to the primary test:

#1) It’s Ranger vs Non ranger (mage) - So there is no interaction of an enemy entangle chance to miss.

#2) The attacking ranger is now hasted.

#3) All tests, including the prior days tests, are all same race vs same race to keep stats / size etc constant. Both the attacker and a victim have a multifaceted gem on to max stats.

In addition, I’ve finished writing out a new data gathering script and can now track the damage output of each entangle attempt to see what can be learned. This series of trials was also run for considerably longer, so instead of the 1328 attempts in my first set, I now have over 3000 attempts in my second.

I will also break results down per trial, as adding damage in makes a control group (henceforth defined as a non flying attacker vs a non flying victim).

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

Control Trial (No Fly), Attacker Hasted

The control data set has 1089 captured instances of entangle.

Out of 1089 attempts, the attack landed 100% of the time (1089/1089).

Damage wise, there is the following breakdown:

Image

There were no hits greater than === OBLITERATES === recorded.You also have a 35.35% chance to do *** DEMOLISHES *** or above damage (and a 70.71% chance to do UPPERCASE damage+), greatly exceeding similar a similar lag amount spell like bash for potential. All damage values are before sanctuary / magic ward. While the attacker is not against a ranger, and is hasted, the miss rate became 0.

Trial #1 - Attack Fly, Victim grounded, Attacker Hasted

This data set has 470 total attempts in it.

Out of 470 attacks, 462 landed (leaving the miss rate at roughly 1.70%).

Damage wise there was the following breakdown:

Image

There were no hits greater than *** DEVASTATES *** collected in this sample. You have a 29.78% chance to do *** DEMOLISHES *** or higher damage, which is a 5.57% decrease overall compared to the control set. You also have a 70.42% chance to do UPPERCASE damage or better, which is only 0.29% lower than the control set, showing that for the most part better than normal damage potential is for the most part preserved. All damage values are before sanctuary / magic ward.

Trial 2 - Attacker and Victim flying, Attacker Hasted

Trial 2’s data set has 635 instances of entangle logged.

Out of 635 attempts, the attack landed 629 times (leaving a miss rate of 0.94%).

Damage wise, there is the following breakdown:

Image

There were no hits greater than *** DEVASTATES *** collected in this sample. You have a 30.55% chance to do *** DEMOLISHES *** or higher damage, which is a 4.8% decrease overall compared to the control set (note per chart below, there is a closer to 7% drop in the *** DEVASTATE *** rate in this trial the largest individual drop in any category, partially offset by a 3ish% gain in the *** DEMOLISH *** rate, the highest individual raise in any category). You also have a 67.70% chance to do UPPERCASE damage or better, which is 3.01% lower than the control set, indicating that a flying victim helps to lower some of the total damage potential, given the uptick in low quality damage done. All damage values are before sanctuary / magic ward.

Trial 3 - Attacker Grounded, Defender Flying, Attacker Hasted

The third and final trial had 843 captured instances of entangle.

Out of 843 attempts, the attack landed 100% of the time.

Damage wise, there is the following breakdown:

Image

There were no hits greater than === OBLITERATES === recorded, supporting the belief that if the attacker is not flown there is a higher high quality damage potential. You have a 36.53% chance to do *** DEMOLISHES *** or above damage (1.18% higher than the control group). You also have a 70.22% chance to do UPPERCASE or better damage, staying roughly in line with the control test(roughly 0.49% under). All damage values are before sanctuary / magic ward.

Trial Comparison

When you compare all trials to the control, the following chart can be built to illustrate the differences of each damage time.

Image

The data table supporting the chart are below:

Image

Flying rangers are bad for damage, but only get a slight to hit % decrease. When these data sets are compared to the ranger vs ranger, un-hasted damage sets, you can see the extreme differences between to hit (between a 41.41% to 20.74% decrease). Both parties flying lowers the top end of the damage range considerably, with a large drop in the higher damage *** DEVASTATE *** offset by half the gain in the lower damage *** DEMOLISHES *** .

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

Hopefully I didn't end up butchering any of my conclusions by going cross-eyed and looking at the wrong trial (I would be lying if I said I didn't do that at least twice), if you see any grave errors please let me know and I'll get things corrected as soon as I can.

Like I said in the top of the first post, I'm willing to take requests (eventually - timing TBD) on what to scrape next data wise. I'm still working on some scripting for how to log things a little better (I also think I will be able to set up some variable #math to get some data on how much actual damage is done when you see *** damage text ***, but if Kiirion wants to release actual values so I don't have to run the trials I'd cool with that too :D).

Generally, until I get my control trials written, skills like enhanced damage and critical strike or weapon flags like sharp aren't testable (I need to be able to get something to compare data that includes them).

What do you think? I welcome all questions, comments, love and hate.
Yeah? Well, you know, that's just like, your OPINION, man.

Posts: 248

Joined: Wed Jun 11, 2014 1:54 am

Post Tue Sep 02, 2014 9:50 am

Re: Trip / Entangle Testing and Results

Quick question: since you didn't set a Dam Roll value for the ranger on the damage tests, does it mean that Entangle's (or even other damaging skills) damage isn't affected by the stat?

Entangle landing rate seems to reflect what everyone sort of already knew: trip is way more reliable.
IMO, it should land really really well against grounded targets because it's vines...from the ground. Or even deal more damage.

It's not really important but it would be nice having an idea of the % of those secondary effects on certain skills/spell, such as:
Water bolt's rust, winter blast's chill, buddha palm's and poison dagger's double cast and burn/poison, mind blast mana thing...

Posts: 320

Joined: Mon Jun 09, 2014 3:47 pm

Post Tue Sep 02, 2014 10:15 am

Re: Trip / Entangle Testing and Results

Envenom wrote:Quick question: since you didn't set a Dam Roll value for the ranger on the damage tests, does it mean that Entangle's (or even other damaging skills) damage isn't affected by the stat?


As far as I saw, dam roll does not have bearing on this skill. Kiirion can reconfirm that for me with a look at the code.

Envenom wrote:Entangle landing rate seems to reflect what everyone sort of already knew: trip is way more reliable.
IMO, it should land really really well against grounded targets because it's vines...from the ground. Or even deal more damage.


I believe it's meant to be about on par with bash in terms of reliability (ranger vs ranger excluded), which seems to be fairly accurate, depending on the circumstances. The damage seems to vary more.

Envenom wrote:It's not really important but it would be nice having an idea of the % of those secondary effects on certain skills/spell, such as:
Water bolt's rust, winter blast's chill, buddha palm's and poison dagger's double cast and burn/poison, mind blast mana thing...


These are certainly things I could work on trying to capture (I don't think I'd need a control group for it either, given magic, for some of them... hmmm) but those should be fairly easy to tease out if I tweak my tracking script.
Yeah? Well, you know, that's just like, your OPINION, man.

Posts: 4

Joined: Fri Jan 16, 2015 2:06 am

Post Fri Jan 16, 2015 6:02 am

Re: Trip / Entangle Testing and Results

I've been curious for a long time if you want to do the tests. What is the damage increase/hit increase for the attacker when the defender is grounded.

Posts: 320

Joined: Mon Jun 09, 2014 3:47 pm

Post Fri Jan 16, 2015 10:08 am

Re: Trip / Entangle Testing and Results

There is a damage increase per this thread only counts if the victim is out of combat (on the initial round of combat, by my interpretation) and sitting/resting/being sprawled but out of combat.
Yeah? Well, you know, that's just like, your OPINION, man.

Return to Testing

Who is online

Users browsing this forum: No registered users and 0 guests

cron
Donate Now
Donate Now


Designed by ST Software for PTF.
Hosted by © 2017 FreeForums.org | Create a free forum | Powered by phpBB
About FreeForums | Legal | Advertise Here | Investors | Contact FreeForums.org
Report Violation
suspicion-preferred