English


uiWebPrevious1uiWebNext

#1 Report | Quote[en] 

A few weeks back, a conversation occurred where a member said that they travel around without being hit. They have a parry modifier at 300. Over the course of a few days, others claim to have the same experience. They stroll in PR and no problemo. I called BS.

I looked at my dodge which, at the time, was around 220. I wasn't a master of ele yet. I immediately started on a mission to obtain that 300. So I mastered ele to bring me up to 250. I crafted a set of +2 dodge light armor, a +20 buckler, and +20 dagger. Voila, 300.

Immediately went out to test. And surprisingly to me, it appeared that I was dodging less at 300 than I experienced with merely 260 (innate + armor).

I brought this up in uni and a private conversation followed how it would be cool to test this. After many hours of xmas box hunting, I decided to get started.

Test conditions are:
Obtain ~1000 attacks
Repeat under different dodge mod conditions:
naked (250 dodge)
+armor (260 dodge)
+buckler (280 dodge)
+dagger (300 dodge)

Tertiary conditions:
vs 1 mob
*facing me immobile
*behind me immobile
*while mobile

vs 6 mobs
*facing me immobile
*behind me immobile
*while mobile

I went to silan and picked the suckling yubo. I will move up to higher level mobs and aggressive as well. My alt will be nearby to rez so I can continue the test. I will likely need to craft some non-dodge and non-parry bonus armor for the higher levels so I may take a few extra hits before dying. We'll see how this goes. Or, I may limit it to 100 samples at high levels in the interest of time. The larger the sample space, the better picture obtained. If I were to base conclusions off the first ten attacks, I would think the lvl1 yubo has a 50% accuracy against me.

I enabled logging and imported the data into a spreadsheet. I removed all chat and unrelated messages. The raw data is pasted here: https://paste.ubuntu.com/p/BVcQxBW58C/

I didn't manage to get a thousand attacks, but it's close enough in my opinion.

With 250 dodge and versus a lvl1 non-aggro mob, I initiated the fight with a taunt. It hit me immediately. These mobs must have a massive to-hit bonus. The yubo conducted critical hits which gave it three attacks.

Test 0 (facing me immobile vs 1):

883 offensive attacks
315 hits (35.8%)
495 dodge (56%)
032 misses [my defense doesn't equate but] (3.6%)
030 critical hits (3.4%) of which:
**7 slow movement
**8 stamina reductions
**6 -20 dodge de-buff (avg 55% attack accuracy)
**9 -20 parry de-buff (avg 55% attack accuracy)

I have data to sort for the other tertiary tests. I will follow this with the remainder after I wake up. With two-hundred-fifty innate dodge, I was surprised to see six mere points above majority in success. I predicted eighty percent dodge. I now expect a negative logarithmic trend after all tests are complete.

Last edited by Thgilsrooc (3 years ago)

---

#2 Report | Quote[en] 

That is very interesting research you do.

#3 Report | Quote[en] 

I got a PM in RC saying that dodge/parry is capped at 60% if the modifier is 50 levels above the mob.With this information, a dodge of 300 would succeed 60% of the time with a q250 mob.From my initial numbers, at dodge 250, I was succeeding 56% with a lvl1 mob. I was 249 points above the level of the mob, so the numbers don't seem to add up. I will run this test again with another yubo (I won't die this way). Standing there naked and let it attack me for a few hours.I'll return from afk and let the larger sample space give me the stats. I should have had enough to return a 60% dodge with 800 attacks, so we shall see.

---

#4 Report | Quote[en] 

The conclusion: it is logarithmic. I was away for hours and gathered well over 4000 attacks.

Here is the raw data: https://paste.ubuntu.com/p/9gcjD23Hyw/

From the numbers included, my character DID dodge 60.2%. With some guidance from others, the dodge information lies here in the wiki: https://en.wiki.ryzom.com/wiki/Melee

It's capped at 60% after all. And when I 'noticed' a decrease in dodge chance with 300 modifier, it was only the perception of lower dodges. I'm not a programmer, but I wonder if the chance to hit is calculated on a dice roll.

Anyway, you'd have to notice 60% throughout a looong history of fighting. Take a gander at the data.

I'd prefer to have high resists than a slightly higher chance to dodge over the long term. A high resist works regardless. I have some nice mats collected for dodge modifiers, but I will now focus on new recipes for resists down the board.

Have fun bark,

Rooc

Edited 2 times | Last edited by Thgilsrooc (3 years ago)

---

#5 Report | Quote[en] 

I think this may have been a difference in definition, and not *necessarily* BS. How does one define "Not being hit". Does that mean:

1. I walked thru PR and chose my path carefully enough and was patient enough to sneak by all the mobs w/o being hit. One of our HOs back in the day used to take a Level 20 alt toon on a world tour about once every 3 months

2. If I can remember correctly, parry w/o boosts or special items

Base Toon: 250
Staff & Shield: +20 +30
NPC Armor: +24

With that level gear you can stand in front of a mob and take hits but many of them will be no damage. You can actually do more damage to yourself with HP credits by attacking the mob, then just standing there.

3. Another way of looking at it I will use an example for. Back many years ago, we decided we wanted to see just how successful we could be digging sups, figuring out the seasons, and refill times, etc. We got very good at it.... tho it kinda made it boring.

As you know many sup spots have aggro on or close to the spots. And wanting to avoid killing the last one, when sup time was coming. So in our team of three, two were in dig gear and I would taunt the last mob(s) around in circles while they dug out the sups .... Yes, I'd get hit but, between running and the high parry, the HP regen rate was higher than the damage rate. When they were done harvesting the sups, I'd still be at or v. worst case, very near the 6700 or whatever HP I started with.

As for the data .... nice work / interesting results

---

uiWebPrevious1uiWebNext
 
Last visit Thursday, 28 March 18:44:18 UTC
P_:

powered by ryzom-api