Way back in the olden days, at the dawn of this millennium, I had a very odd career. I ran factories. More specifically, I took over bad factories, and made them into good factories. The hours sucked, but the job was a lot of fun. When I started playing WoW, my attitude towards raiding evolved fairly directly from my experience running industrial facilities. When I got afflicted with raid leadership, I applied a lot of the lessons I learned about improving factories to improving a raid. I’m not sure if any of what follows is useful advice for anyone else – but it’s how I moved from manufacturing car parts to manufacturing boss kills.
Root Cause Analysis
There’s a limited amount of time in any given raid, and the less of it you dedicate to talking about stuff, the more of it you can spend killing stuff. Root Cause Analysis is all about solving causes, not solving symptoms. It may sound like it’s going to take a lot of time out of your day, but it’s the sort of thing that only takes seconds in reality, and can save hours of wiping.
The most basic technique in root cause analysis is “Five Whys”. Set out your problem, and some of the stuff you know already (Who, What, When, Where, How, How Many). And then ask Why? And whatever the answer is, ask Why? again. And again. Most of these steps you can probably do in your head, you already know the answer to them, if you’re paying attention. And when you get down the fourth why, and you don’t know the answer, you may find yourself asking the raid a question that seems completely out of left field – because they’re still thinking about solutions to symptoms. That’s okay, they’ll get used to it.
If you solve a symptom, the problem will occur again. You have to solve a lot of symptoms before you luck into something that really solves the cause. It’s what’s referred to as a “band-aid” fix, and sometimes it’s even the right thing to do. But you can save a whole lot of time in raid by at least knowing the difference, and knowing when you need to dig down and fix a cause.
There are any number of techniques for RCA, and I’m not going to go into them here. Most systems work – so long as you bear in mind the most basic principle: Fix the root cause, don’t waste your time on symptoms.
Process Optimization (with a nod to Lean Manufacturing)
There’s a few basic common-sense guidelines for process optimization on a plant floor. Here’s a few bullet points I’m going to discuss:
- Reduce steps
- Optimize the slowest stage in a process first
- Put the right people in the right jobs
Those are the things that I pay the most attention to when progressing on a boss. There are always other things to bear in mind. I’ve said it before, though – specific mechanics aren’t important in any strategy building process. Any mechanic can basically be abstracted as a cost to raid performance. Your goal is to find the lowest-cost and most repeatable procedure for succeeding at the mechanic. That’s process optimization.
Reduce steps is clear – movement is bad, it has a cost associated with it. One thing to bear in mind with movement when building strategies is that “when” and “how often” are as important as “how far”. If you can set up a series of short moves buried by instant casts, that may hide the movement penalty. If you can plan your movements for a phase with a low DPS requirement, that may effectively hide the movement penalty. But overall, move less and move better is always going to be a big gain to raid performance.
Optimize the slowest stage first – what does that even mean? In the context of any given bossfight, it means set up for the hard part. Use heroism and potions and other cooldowns in order to get through the hard part. Stack the raid for the hard part. If you’re going to throw manpower and time and energy at something, it’s not going to be P3 Al’akir, it’s going to be P1 or P2 – and that might differ by raid. Due to the way practice and muscle memory works – the hard part may well change. A long time ago, we optimized a lot of Ragnaros around P2. These days, we’re entirely set up for P3 and P4 DPS – practice has made P2 handle itself. Optimize the slowest stage first. Spend your effort where it generates the most rewards.
Simplify. Boss fights have a lot of mechanics. In a lot of the most interesting boss fights, one mechanic will encourage you to do one thing, and another will encourage the opposite. You always have the option of doing incredibly complex timing-dependent things and setting up a beautifully choreographed performance that looks great on kill videos. You also have the option of stacking up a couple of raid cooldowns and just ignoring some mechanics. Brute force isn’t always pretty, but it’s a valid strategy – and sometimes it’s the best strategy. Brute force isn’t the only option for simplifying, though. When given two competing strategies, one of which consists of “Move from A to B”, and the other consists of “imagine the encounter area as a 2-D mapping of a 4-D manifold described by something only Theck would understand”, I don’t really care how efficient the second strategy is. If I can’t explain it to an average five year old, it’s probably not workable.
Put the right people in the right jobs. Different people have different strengths, and different classes have different strengths. Utilize the strengths of your people. Maybe you have a tank who does amazing threat (DPS) and never ever dies, but he can’t kite to save his life. Every time he moves, healers get cleaved. Don’t put him in the kiting job. Always be willing to honestly evaluate your people, and use them effectively. It doesn’t help anyone to put the wrong people in the wrong jobs – it just makes everyone frustrated.
I ended up raid leading by accident. I didn’t seek out the job, but I’m glad I got the chance to do it. It’s been difficult and rewarding – generally at the same time. It certainly helped to have some prior experience that I was able to apply. In some ways, managing a guild of thirty people is harder than managing a company many times that size – but keeping in mind some of the things above was how I muddled through.