Ensign
Join Date: Jan 2014
Posts: 15
Requirements Edit
1. Multiple submissions from one player would be accepted. However, the 5 CombatLogs must be 5 new logs each time.
2. Deadline has been removed. However, only CombatLogs parsed after 3/11/2014 would be accepted.

Introduction
I'm not sure if many players understand, but currently there is no perfectly accurate method to measure DPS in STO. STO does not have statistical results compared to other games such as League of Legends, World of Tanks, and etc. However, one may wonder, "Doesn't STO already have DPS League?" While DPS League does exist, there's no set variables to measure indivisual progress. Therefore, I want to introduce a DPS benchmark that emphasizes on one's ship build rather than teamwork.

What is the purpose of this benchmark?
To help all players understand what is realistic damage and DPS. The benchmark will also benefit players who are interested in personal progress.

Tools
Nothing is needed. Only 5 CombatLogs are required from each participant.

Requirements
1. Only Infected: The Conduit (Elite) CombatLogs will be accepted
2. Only Solo Queue CombatLogs would be accepted
3. Inviting friends or "random" teammates from 10k DPS, 20k DPS channels, and etc before a round would not be acceptable.
4. Character and Ship Build has to be consistent throughout the 5 CombatLogs.
5. Any faction and ship is welcome to participate.

Instructions
Quote:
Originally Posted by STO CombatLog Reader README
1. To create a CombatLog file, use the "/combatLog 1" command (without "" !) in the chat window of STO before a fight.
(Note: You have to use the command every time you log in the game)
2. After fighting you can use the "/CombatLog 0" command (without ""!) in the chat window to stop STO writing into the CombatLog.
(Note: You don't need to stop the CombatLog, but it may make the results better. Otherwise things like using a Tribble or Hypo in social zones (by any player) will be documented, too! That may distort the statistics.)
3. Find the path to your CombagLog.Log file. (the end will be "...\Star Trek Online\Star Trek Online\Live\logs\GameClient\CombatLog.Log" but you will need the whole path)
Please Note: Rename each CombatLog.Log file into the following format: Log##_PlayerName_InstanceName_MM_DD_YYYY. (Ex. 01_Lily_infected_03_11_2014)

4. Upload the 5 logs online to Mediafire, Google, Dropbox, etc. and make a forum post in the following format:

Quote:
Originally Posted by arpeggione
Player Name: Lily
Ship: Fleet Avenger
1. 01_Lily_infected_03_11_2014: http://www.mediafire.com/view/pirns1...03_11_2014.Log
2. 02_Lily_infected_03_11_2014: http://www.mediafire.com/view/cfl17i...03_11_2014.Log
3. 03_Lily_infected_03_11_2014: http://www.mediafire.com/view/qk6ucc...03_11_2014.Log
4. 04_Lily_infected_03_11_2014: http://www.mediafire.com/view/l006gw...03_11_2014.Log
5. 05_Lily_infected_03_11_2014: http://www.mediafire.com/view/ruvb6d...03_11_2014.Log

or

Player Name: Lily
Ship: Fleet Avenger
Folder: https://www.mediafire.com/folder/9uy...ar_Trek_Online
The following is a small math lecture that explains how to use this benchmark.
Quote:

What is behind this benchmark?

In order to understand this benchmark, one would need to know some Statistics. There's a rule in Statistics called the Empiral Rule.
The Empirical Rule:
Approximately 68% of the observations will fall in 1 standard deviation.
Approximately 95% of the observations will fall in 2 standard deviation.
Approximately 99.7% of the observations will fall in 3 standard deviation.
A normal distribution (a Bell curve) would be calculated from everyone's data. We can then interpret the Bell curve with the Empirical Rule.
Refer to this wiki for more information: http://en.wikipedia.org/wiki/68%E2%8...80%9399.7_rule

What makes this benchmark effective?
Besides having set variables such as random teammates, constant character and ship builds, there are many ways we can check the integrity of people's numbers.
There's a formula called z-score, which is [(Observed DPS) - (Average DPS)]/Standard Deviation.
For example, the average DPS is 7,000 with an standard deviation of 3,000. Someone claims they have 40,000 DPS. So, we plug these values into the calculator: (40,000 - 7,000)/3,000 = 11. What do we with the number 11? z-score tells us that any value that equals 2 or higher is somewhat unlikely. Any value higher than 3 is very unlikely, and considered an outlier. Since 11 is much, much higher than 3, we can assume that their 40,000 DPS did not follow our benchmark requirements because it is very unlikely to happen. We can do the same test for player's damage, time, and more.

Do you have an real example?

Spreadsheet 1: https://docs.google.com/spreadsheets...Tux5xE/pubhtml
Graph 1: http://i.imgur.com/teGBqKv.png
Blue is Time (s) taken. Red is Damage Dealt. Orange is DPS.
This spreadsheet is my personal logs and I have found these values. On the DPS Summary page, one can see that the mean is 6,795.422. The Standard Deviation is 3,704.279.
Someone claims they did 22,000 DPS. We check by [(Observed DPS) - (Average DPS)]/Standard Deviation: (22,000 - 6,795.422)/3,704.279 = 4.104. This means that we can consider their DPS an outlier or they are very impressive.
If one is lazy and does not want to calculate these numbers, you can see I calculated the Normal Distribution at the bottom of each page. Then just follow the wiki section on Expected/Approx Frequencies.

How do I use the table of Expected/Approx Frequencies?

Follow this chart: http://en.wikipedia.org/wiki/68%E2%8...her_deviations
Then lets refer to the earlier example: Someone claims they can do 22,000 DPS with this benchmark. 22,000 DPS is over 4 Standard Deviations. This means that it is around 1 in 15,787 chances or about "Every 43 years (twice in a lifetime)."
That means they are very good or they could be lying.

Are there any other ways to interpret the data?
Refer back to the Empirical Rule, lets say in Spreadsheet 1. A player has gotten 16,000 DPS in an ISE run. Since 16,000 DPS is 2.5 standard deviations away from 0 (or the median). He or she achieved an record better than 95% of the player population (according to this benchmark).

Q&A

I don't see what taking the average of random ppl with random fits and flying skills/styles in random pugs would tell you.

As shown by Spreadsheet 1, everyone's values add up and creates a distribution which can be interpreted by the Empircal Rule or Expected/Approx Frequencies (Refer to the small math lecture).

How come your max damage recorded is so low?

I seem to have bad luck in finding high DPS pugs. I think it is because everyone is in DPS channels. Therefore, I need your help in getting a more complete picture!
I also estimate that for this benchmark, players will get anywhere from 1,000 to maybe 30,000 DPS. Any higher would probably not be possible, but someone can always prove me wrong.

If you want to measure 1 persons ability/contribution, why not use Solo ISE?

Solo ISE actually dilutes the damage. One would end up having approximately 12,000,000 total damage which makes your DPS higher than it should be.

This exercise is frustrating, why do you bother?
As a player who can carry pugs efficiently (refer to Spreadsheet 1), the only answer I can think of is perhaps this benchmark is suggesting that your ship build is not all rounded. Your build is perhaps lacking a good balance between offense and defense.

The variance of players is very high, how would you handle this difference?
Refer again to Spreadsheet 1, I don't have any problems with variance since I am collecting the results of every player that appears in a CombatLog.

If you want to measure team performance, how come you don't use DPS League?
Because I'm not interested in DPS league. What I am interested about is about a player's personal progress in a practical setting: Solo Quene.

I like DPS League better, Why should I use this benchmark?
If one likes DPS League better, then by all means you should continue using that. However, this benchmark can also be used with any other types of DPS measures. One can have a full ISE build for DPS league runs but can also create a another setup that is suitable for all instances.

I don't want to use your benchmark
Okay then, take care and have a nice day. Thanks for not bothering me.
Personal ISE Logs for reference. What the Solo Quene spreadsheet would be modeled after.
Spreadsheet 1: https://docs.google.com/spreadsheets...Tux5xE/pubhtml
Graph 1: http://i.imgur.com/teGBqKv.png
Blue is Time (s) taken. Red is Damage Dealt. Orange is DPS.
I will began to calculate the mean, median, standard deviation, normal distribution, once I get enough players to provide me their CombatLogs.

Solo Queue
Spreadsheet: https://docs.google.com/spreadsheets...S5Egdo/pubhtml

Last edited by arpeggione; 03-15-2014 at 12:44 PM.
Captain
Join Date: Jul 2013
Posts: 1,472
# 2
03-11-2014, 07:09 PM
I don't see what taking the average of random ppl with random fits and flying skills/styles in random pugs would tell you. Anyway, I suspect resulting average will be low, but you already know that.
Critical Hit Calculator
Captain
Join Date: Apr 2013
Posts: 3,315
# 3
03-11-2014, 08:22 PM
Over 9000.
Vice Admiral Rylana - U.S.S. DNT Omega X || Vice Admiral Lyzara - I.R.W. DNT Omicron X
Vice Admiral Kailiana - R.R.W. DNT Theta X || Vice Admiral Talina - I.R.W. DNT Tau X
Vice Admiral Victoria - U.S.S. DNT Upsilon X || Lt. General Dannika - I.K.S. DNT Sigma X
Vice Admiral Shondra - R.R.W. DNT Alpha X || Lt. General Rosanna - I.K.S. DNT Iota X
=== Vice Squad/House of Tlhap-Jen ===
Captain
Join Date: Jan 2013
Posts: 3,158
# 4
03-11-2014, 08:34 PM
Quote:
Originally Posted by noblet View Post
I don't see what taking the average of random ppl with random fits and flying skills/styles in random pugs would tell you. Anyway, I suspect resulting average will be low, but you already know that.
Well, it's no more unreasonable a benchmark than space bullet with private snowflake clubs flying individually dysfunctional ships. Arguably, this is a more practical and meaningful point of comparison using ships with practical rigging, instead of ships only capable of performing a singular rigid dance. When you pug it, you have to build a ship that works, not a ship that is entirely tailored to one specific piece of unchanging content and circumstances, which can essentially be flown by a timed script in a choreographed, prescripted dance.

It would also certainly be useful to include some other piece of non-ISE, even non-Borg, content, as part of the benchmark for the same ship. Otherwise you end up with a hyper-specialized ISE machine which translates poorly to anything else.

Last edited by doffingcomrade; 03-11-2014 at 08:38 PM.
Captain
Join Date: Jun 2012
Posts: 3,229
# 5
03-11-2014, 08:43 PM
Or you could use SB234 times instead, since its something you can run solo so no friends helping on the debuffs, not multiple targets to AOE, and shielded so no straight-to-hull advantages. Plus, while its not exactly going to one-shot you or anything, it has enough offense and durability that if you run in there with all offense NO defense, it'll still last long enough to hurt you. Still imperfect and the timer can be gamed, but a better measure of what ship by itself can do.
Ensign
Join Date: Jan 2014
Posts: 15
# 6
03-11-2014, 11:06 PM
Quote:
Originally Posted by noblet View Post
I don't see what taking the average of random ppl with random fits and flying skills/styles in random pugs would tell you. Anyway, I suspect resulting average will be low, but you already know that.:P
Try parsing yourself 5 times in random queues using my listed requirements. After, compare these five CombatLogs to your five best ISE results. What made those five best ISE records possible?

Quote:
Originally Posted by rylanadionysis View Post
Over 9000.
Care to help out by sharing some CombatLogs? :)

Quote:
Originally Posted by reginamala78 View Post
Or you could use SB234 times instead...
I appreciate the ideas of using other instances as benchmarks. But first, I will need help on completing this current test. If this benchmark goes well, I will be willing to create more types of comparisons.

Also, a spreadsheet and formula explanations are added to my first post.
https://docs.google.com/spreadsheets...5Egdo/pubhtml#

The success of this experiment rides on your help, post some CombatLogs today!
Captain
Join Date: Apr 2013
Posts: 3,315
# 7
03-11-2014, 11:34 PM
Since you use combatlogreader just look up my @handle (same as my forum name) in the DPS League

Shondra has hit over 29k, Rylana over 16, Victoria over 17 (as far as I remember, I stopped chasing DPS a while ago)

All of my record runs were done pugging. For some reason my DPS is always crap when I run with people that have higher deeps than me. C'est la vie.


Also, according to CLR DPS League the average DPS of all reported players is 7162.56 as of the time of this post.
Vice Admiral Rylana - U.S.S. DNT Omega X || Vice Admiral Lyzara - I.R.W. DNT Omicron X
Vice Admiral Kailiana - R.R.W. DNT Theta X || Vice Admiral Talina - I.R.W. DNT Tau X
Vice Admiral Victoria - U.S.S. DNT Upsilon X || Lt. General Dannika - I.K.S. DNT Sigma X
Vice Admiral Shondra - R.R.W. DNT Alpha X || Lt. General Rosanna - I.K.S. DNT Iota X
=== Vice Squad/House of Tlhap-Jen ===

Last edited by rylanadionysis; 03-11-2014 at 11:39 PM.
Lieutenant
Join Date: Dec 2012
Posts: 71
# 8
03-12-2014, 12:22 AM
Been hitting regular 10-12k on my beamscort in pugs but its squishy and can't handle return fire in a group of 4 other 3-4k Dps. I've topped 18k on it.

Tonight in a pug about an hour ago I hit 17k in my vesta with a 3k and a 4k Dps in the group.

My fleet assault cruiser has 24k in a great group before but I've not been able to replicate that.

I think I'll take this challenge on in my vesta...
Ensign
Join Date: Jun 2013
Posts: 13
# 9
03-12-2014, 12:32 AM
Quote:
Originally Posted by arpeggione View Post
Introduction
With the growing amount of DPS threads, I've always wondered, What is an accurate distribution of individual DPS among players? Many of us already knew that some players have already achieved 30k, 60k, or even 80k DPS. However, these high values are results of team effort. Also, to include these results into a DPS distribution would be inaccurate because there are no set variables. Therefore, I would like to propose a small experiment to willing participants.

Tools
Nothing is needed. Only 5 CombatLogs are required from each participant.

Requirements
1. Only Infected: The Conduit (Elite) CombatLogs will be accepted
2. Only 5 CombatLogs will be accepted from each participant
3. Only Solo Queue CombatLogs would be accepted
4. Inviting friends or "random" teammates from 10k DPS, 20k DPS channels, and etc before a round would not be acceptable. Participants would be asked to parse again if the requirements are not met.
5. Character and Ship Build has to be consistent throughout the 5 CombatLogs.
6. Any faction and ship is welcome to participate.
7. Only data parsed and posted from 3/11/2014 to 3/16/2014 1200 hours forum time would be accepted. I will post again once the deadline is up.

Instructions (from STO CombatLog Reader README)
1. To create a CombatLog file, use the "/combatLog 1" command (without "" !) in the chat window of STO before a fight.
(Note: You have to use the command every time you log in the game)
2. After fighting you can use the "/CombatLog 0" command (without ""!) in the chat window to stop STO writing into the CombatLog.
(Note: You don't need to stop the CombatLog, but it may make the results better. Otherwise things like using a Tribble or Hypo in social zones (by any player) will be documented, too! That may distort the statistics.)
3. Find the path to your CombagLog.Log file. (the end will be "...\Star Trek Online\Star Trek Online\Live\logs\GameClient\CombatLog.Log" but you will need the whole path)

Please Note: Rename each CombatLog.Log file into the following format: Log##_PlayerName_InstanceName_MM_DD_YYYY. (Ex. 01_Lily_infected_03_11_2014)

4. Upload the 5 logs online to Mediafire, Google, Dropbox, etc. and make a forum post in the following format:


What will happen to everyone's CombatLogs?
https://docs.google.com/spreadsheets...S5Egdo/pubhtml
Everyone's data would be calculated using these example formulas:
DPS =E2/D2 (Damage/Time).
Average Damage =AVERAGE(Logs!E2:E6) (Average Damage from Logs page cells E2 to E6)
Average DPS =AVERAGE(Logs!F2:F6) (Average DPS from Logs page cells F2 to F6)
Average Time =AVERAGE(Logs!D26) (Average Time from Logs page cells D2 to D6)
Complete data would be published a day or two after 3/16/2014. (I am a college student, so expect to have delays).
I will also use summary statistics on everyone's damage and DPS.

Remember to be honest to yourself and your fellow players.
I like this. I've been doing the exact same thing for quite a long time now. I know exactly what your point is. I know exactly why you're curious about it. You will not receive favorable acknowledgement from a large portion of the dps channel crowd. If I find time I may upload some of my results. I just moved, things are not very organized here yet.

The point, for those that may have missed it while out chasing butterflies and dps, is to find out what YOU do for a group. What YOUR contribution to a random group is. What YOUR average dps is in an average group. It shows clearly how badly any ship can be boosted to ridiculous dps.
Republic Veteran
Join Date: Jun 2012
Posts: 148
# 10
03-12-2014, 01:56 AM
I'd suggest you only use Solo ISE runs for your calculations to avoid interference from teammates. Skills like tactical fleet or beta / FOMM / disruptor debuffs will always have an influence on the DPS values of every participant.

I had 26.6k+ when soloing ISE. It was a rather defensive setup, so I'd guess one might reach 35k or even more using broken beta spamming pets.

Visit the German STO forum: www.sto-center.de/forum
Youtube: https://www.youtube.com/user/STOHellspawny
Weapon Damage Calculator: Download
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off


All times are GMT -7. The time now is 07:03 AM.