sk starting points

[Troxx]

Gamparse is 100% accurate in reporting dps under these circumstances as well as max hit, min hit, and average hit.

Gamparse is only buggy when you have a group going on when engage/starts for relative players may be different (ie the monk or rogue doesn't start attacking until 5-20 seconds into the fight.

[DeathsSilkyMist]

Originally Posted by Troxx [You must be logged in to view images. Log in or Register.]

Gamparse is 100% accurate in reporting dps under these circumstances as well as max hit, min hit, and average hit.

Gamparse is only buggy when you have a group going on when engage/starts for relative players may be different (ie the monk or rogue doesn't start attacking until 5-20 seconds into the fight.

I do not trust gameparse. I have never used it, and don't know how it parses the data. You are admitting it is buggy, which means I cannot know if your parse is also bugged.

I let you provide non-video evidence because you do not have a Youtube account.

Please humor me and do not use game parse. I want the raw logs, not a screenshot of an excel file.

[Lune]

He isn't admitting it is buggy, he is acknowledging that in one specific circumstance, comparing the DPS between individuals who started fights at different times, you aren't getting a like terms picture. It is even accurate in this situation, it's the interpretation of that specific situation that you have to be careful with.

For all other applications, it is just aggregating log data and is 100% accurate.

It's your data that's going to be shit because it's subject to human error.

[DeathsSilkyMist]

Originally Posted by Lune [You must be logged in to view images. Log in or Register.]

He isn't admitting it is buggy, he is acknowledging that in one specific circumstance, comparing the DPS between individuals who started fights at different times, you aren't getting a like terms picture. It is even accurate in this situation, it's the interpretation that you have to be careful with.

For all other applications, it is just aggregating log data and is 100% accurate.

It's your data that's going to be shit because it's subject to human error.

He is admitting it is buggy lol. If it can break in a certain situation, I am not sure why you are confident that it cannot break on others. Are you a developer for gameparse?

With the raw logs, other people can check them to see if human error occurred. Nobody can check if an error occurred in gameparse.

[Lune]

Originally Posted by DeathsSilkyMist [You must be logged in to view images. Log in or Register.]

He is admitting it is buggy lol. If it can break on a certain type of parse, I am not sure why you are confident that it cannot break on others. Are you a developer for gameparse?

You have no grasp of math or language, you lose hold of meaning so easily.

He used the word buggy, but trust me when I say, having used Gamparse, what he means is that the DPS readout for different players isn't accurately portrayed because it's based on different start times for a fight.

That has no bearing on what he's currently doing.

How do I know vBulletin isn't taking your words and changing them to promote your argument. Are you a developer for vBulletin? That's you right now. It's a fucking log compiler.

[DeathsSilkyMist]

Originally Posted by Lune [You must be logged in to view images. Log in or Register.]

You have no grasp of math or language, you lose hold of meaning so easily.

This is nonsense. You are making the claim gameparse cannot be buggy. You have no evidence for this, and Troxx has openly admitted gameparse has issues in certain scenarios. I am not sure why you are defending this when you have no ability to back up your statement.

There is no reason why Troxx cannot provide the raw logs.

[Lune]

Originally Posted by DeathsSilkyMist [You must be logged in to view images. Log in or Register.]

This is nonsense. You are making the claim gameparse cannot be buggy. You have no evidence for this, and Troxx has openly admitted gameparse has issues in certain scenarios. I am not sure why you are defending this when you have no ability to back up your statement.

There is no reason why Troxx cannot provide the raw logs.

I am not, I'm explaining to you what Troxx is saying. He is saying one specific feature has limited utility. That feature isn't even bugged, it just doesn't account for different times players start fights. It's completely independent of a single player parsing combat with mobs.

If someone is in combat for 6 seconds and they get a string of hits, it will show much higher DPS than a player who has been in combat for 2 minutes with normal distribution of hits. That's it. Unique to the DPS feature between players.

[Troxx]

Ok there were so many issues with trial 1 I'm going to put it in the scrap bin and we will try again. The results were significant but there were issues. Even though the parses support my hypothesis and understanding of this game, I cannot rightfully stand on them confidently and proclaim it is accurate. Regardless, I will post findings.

I had to move along because other players needed the camp more rightfully for quest related reasons.

Methods: Single pull static spawn undead. Lull neighbor. Pull. Fight down to 20% or just below. Root. Camp. Log back in. Heal, rinse repeat.

# of fights 7 with each setup
Str variables: 140 vs 160

Limitations of fights:
-High green xp mob, not blue (this is a problem)
-Due to the above, fights were not as long as they should have been
-High variability from fight to fight. Ironically my highest dps was at 140 str but the averages still are relevant
-Very limited hits logged due to aforementioned low level of mobs and not enough time in camp to get a more appropriate # of hits

Strengths of this test/parse:
-Same mob 100% of the time
-Within xp range (albeit barely)

160 strength:
Av DPS: 29
Av Hit: 62
Max hit: 120

140 strength:
Av dps: 25
Av Hit: 58
Max hit: 116

Difference in DPS: 16%
Difference in av Hit: 6.8% (not consistent with findings in av dps - that's RNG luck one way or another)
Difference in max hit: 3.5%

Weapon used:

https://wiki.project1999.com/Darkmetal_Falchion

Goal was to prolong the fights as long as possible.

Buffs: Self only
-Strength difference accomplished by removing gear, none of which had + attack effects
-Yaulp was never used

*bonus fries: there was a significant parse-able difference in damage taken in removing pieces of gear


TLDR:

Parse, though it showed what i wanted it to show ... it has fatal flaws. I will try again and will brainstorm mobs that are dark blue exclusively moving forward.

TALLY HO!

What could make this easier?
-finding a a much higher level mob
-having a dedicated healer so individual fights could be longer
-having a corner to back into so repositioning is a non issue

[Troxx]

Originally Posted by DeathsSilkyMist [You must be logged in to view images. Log in or Register.]

Can you please provide the raw logs, and not the gameparse. You have admitted multiple times in the other thread you think gameparse is incorrect.

I am happy to once I have a sample that is worthy of scrutinizing.

First set supports all the things I want it to. I am mature enough, however, to acknowledge the many flaws and wait for a more reliable pool of data before proceeding.

I hope you can appreciate that.

First round of data is garbage despite the fact it showed what I wanted it to show.

I'm a physician. Was a biologist before. I place a lot of value on qualitative data rather than any random BS that supports my preferred outcome. I am basically allergic to confirmation bias ... to my very core.

[DeathsSilkyMist]

Originally Posted by Troxx [You must be logged in to view images. Log in or Register.]

Ok there were so many issues with trial 1 I'm going to put it in the scrap bin and we will try again. The results were significant but there were issues. Even though the parses support my hypothesis and understanding of this game, I cannot rightfully stand on them confidently and proclaim it is accurate. Regardless, I will post findings.

I had to move along because other players needed the camp more rightfully for quest related reasons.

Methods: Single pull static spawn undead. Lull neighbor. Pull. Fight down to 20% or just below. Root. Camp. Log back in. Heal, rinse repeat.

# of fights 7 with each setup
Str variables: 140 vs 160

Limitations of fights:
-High green xp mob, not blue (this is a problem)
-Due to the above, fights were not as long as they should have been
-High variability from fight to fight. Ironically my highest dps was at 140 str but the averages still are relevant
-Very limited hits logged due to aforementioned low level of mobs and not enough time in camp to get a more appropriate # of hits

Strengths of this test/parse:
-Same mob 100% of the time
-Within xp range (albeit barely)

160 strength:
Av DPS: 29
Av Hit: 62
Max hit: 120

140 strength:
Av dps: 25
Av Hit: 58
Max hit: 116

Difference in DPS: 16%
Difference in av Hit: 6.8% (not consistent with findings in av dps - that's RNG luck one way or another)
Difference in max hit: 3.5%

Weapon used:

https://wiki.project1999.com/Darkmetal_Falchion

Goal was to prolong the fights as long as possible.

Buffs: Self only
-Strength difference accomplished by removing gear, none of which had + attack effects
-Yaulp was never used

*bonus fries: there was a significant parse-able difference in damage taken in removing pieces of gear


TLDR:

Parse, though it showed what i wanted it to show ... it has fatal flaws. I will try again and will brainstorm mobs that are dark blue exclusively moving forward.

TALLY HO!

What could make this easier?
-finding a a much higher level mob
-having a dedicated healer so individual fights could be longer
-having a corner to back into so repositioning is a non issue

Multiple issues here:

1. I do not know when gameparse decides a fight has started. This can skew the average DPS if there is dead time between autoattack on and autoattack off.
2. You need to exclude ripostes. For example, I got 4 ripostes on the first turtle test, and 9 on the second turtle test. That is hundreds of damage that is random, and can skew either data set.
3. You used a 1h weapon, and I used a 2h weapon. There are differences in how those formulas work.
4. You need to tell us what mob you were fighting.
5. Your difference in DPS was 4, and mine was 3. The only reason why your percentage is higher is because your numbers are lower (because you used a 1H sword). I am not sure why you are celebrating when the numbers are similar, and they may be a bit inflated.