Latest in Gaming

Image credit:

Encrypted Text: Rogues in theory vs practice


Every week, WoW Insider brings you Encrypted Text for assassination, combat and subtlety rogues. Chase Christian will be your guide to the world of shadows every Wednesday. Feel free to email me with any questions or article suggestions you'd like to see covered here.

The goal of every rogue is to deal the maximum possible damage. In order to accomplish that goal, we have to figure out the best stat and rotation configuration. Discovering the best stat and rotation configuration would take weeks of work if done empirically. We can't simply apply brute force to make the issue go away. That strategy only works against mages and hunters.

In order to quickly determine the optimal configuration, many intelligent rogues have worked on modelling our damage and rotations mathematically. As WoW's mechanics and systems have evolved, so have these models. While they're getting more and more accurate at predicting rogue performance, they're still not 100% in sync with what actually happens in-game. Rogues are a class played with a keyboard, not a pen and paper.

Parses are the golden standard

The goal of every rogue is to deal the maximum possible damage. I don't care about what a spreadsheet or simulation says my DPS should be. I only care about the actual results. The DPS meter at the end of an encounter is the one true measure of our performance. World of Logs is where thousands of rogues aggregate their performance to allow us to share combat logs and to compete with others. DPS Bot is a tool that compiles WoL's data into an easy-to-read chart to track global trends and performance.

Blizzard's balancing job isn't to make sure all of our simulations are accurate. Blizzard's balancing job is to make sure that rogues are performing well on DPS meters and combat logs. Our simulations and models are designed to give us the information we need to reach the top of the meters, but it's up to us to execute on that information. Our real world performance is the only metric that actually matters.

We need data, not anecdotes

Blizzard has equipped the rogues of Mists with all of the tools they need to top the meters against Blade Lord Ta'yak in Heart of Fear. Overall, assassination rogues are doing significantly more damage to Ta'yak than any other DPS class. If a retribution paladin does more damage than you on this encounter, that's not Blizzard's fault. It could be the result of a large gear differential, poor play on your part, extremely bad or good luck, or any other number of factors. A parse of this performance is not individually valuable, and proves very little.

In order to be able to find trends or accurately discuss rogue performance, we need a lot of parses. The more rogues that submit parses to WoL, the more accurate our view of rogue performance will be. Without the thousands of parses at WoL, we wouldn't be able to figure out if rogues were doing great or doing awful as a class. The catch is that since we have all of this parse data, we also can't argue against it. You can't claim that ret paladins are better than rogues if there are a thousand examples that prove you wrong.
Parses are the most important balance tool

Ghostcrawler has mentioned on Twitter that DPS Bot's information can be valuable when discussing class balance. In addition, he's also commented that our community simulations are not necessarily accurate and that more emphasis is placed on parses. World of Logs and DPS Bot only contain data taken straight from our combat logs, which is unimpeachable. You can't argue with the actual results.

According to DPS Bot, rogues are the top melee class overall and certainly competitive on every encounter. You are now equipped with the knowledge that we are doing just fine in PvE. If you read a post on the forums about how awful our damage is in PvE, you know better. Many rogues will try to use anecdotal data about how some shadow priest in their raid beat them to prove that the class as a whole is performing poorly. When discussing balance, we have to look at the big picture.

Simulations and models have a purpose

The purpose of tools like ShadowCraft is not to accurately predict your DPS. There are far too many factors that weigh in on any given individual performance. What ShadowCraft is designed to do is give you the best possible information on stats and rotations, thus enabling you to perform at your best. ShadowCraft isn't about the final result, but rather about the best way to get there.

ShadowCraft and other tools are going to help you properly reach your hit and expertise caps, they're going to help you pick the right gems and reforge the right stats, and they're going to help you make the right choices. They have nothing to do with balance and the DPS numbers that they give you are only estimates and not to be taken seriously.

Contributing to the discussion

If you want to help the rogue community, you can contribute in one of two ways. The first, and easiest way to help is to simply submit as many combat parses as possible. Use the /combatlog command to capture your performance in a raid environment and submit it to World of Logs. It's free and you'll be giving the rogue community more data that we can use to analyze balance and to find trends. You'll also be able to compare your logs with other rogues to find new strategies and tricks.

The second way you can help is to contribute to the models of rogue damage that are being worked on. Follow the discussions at Elitist Jerks and assist with any research that needs to be done. The rogues working on the DPS models often have questions about specific interactions and mechanics which can be solved with some quality time attacking a training dummy. Rogues were recently able to diagnose the issue with Mutilate not applying poisons by testing the ability on a dummy.
Sneak in every Wednesday for our Mogu'shan Vaults guide, a deep-dive into the world of rogue rotations -- and of course, all the basics in our guide to a raid-ready rogue.

From around the web

ear iconeye icontext filevr