Sunday, January 14, 2018

Continuous Delivery of Games

The traditional overhead of "getting a game ready to release" can take quite a bit of time and reduce how quickly you can release new features to your players and respond to what the market is telling you.

Take, for example, the steps a mobile game development team I recently visited goes through for a release every few months:

  1. Merge the development branches
  2. Fix the merge-caused bugs
  3. Optimize, polish and debug for release quality
  4. Send to QA
  5. QA does their regression tests
  6. If bugs are found, return to step 3
  7. If quality is good, submit
  8. If the submission is rejected, return to step 3 to fix it
This took a lot of time! As a result, the team could only release major features several times a year, often behind those of their competitors who captured many of their impatient players.

Continuous delivery is a series of practices that ensures your code and assets can be rapidly and safely released to players by delivering every change to a production-like environment and ensuring that the game functions as expected through painstaking automated testing.

As more games move to more continuous delivery, tools and practices for supporting the ability to deploy with less hassle become more valuable.  Examples are:

  • Feature toggles
  • Continuous integration
  • Unit testing
  • Automated gameplay testing
  • Blessed build indicators
  • Form a stability team
  • Integrate QA
  • Etc
Most of theses practices can be found in my latest book Gear Up!

Thursday, December 14, 2017

The Challenges of Using Story Points for Games

Story Points - The Challenges
Many game teams have used story points to measure items in the their product backlog.  There has been a lot of confusion, misuse and overuse of them.

Let’s look at some of these challenges

Using points for stories that multiple disciplines work on
Story points are meant to drive cross-functional behavior and to focus the cross-functional team on delivering increasing levels of value (velocity).  This is done by encouraging discussions about improving collaboration.

For example, if QA is a bottleneck, then the discussion turns to questions such as:
  • “How can we create functionality that doesn’t require so much QA effort?”
  • “What can the rest of the team do to alleviate this bottleneck?”
  • Etc.
Cross-functional story points are meant for complex work that is a bit unpredictable, such as developing a new mechanic or character behavior.  Teams swarm on complex work, iterating and fixing things daily.  It’s should be a true cross-discipline effort

If the work is more predictable, such as content production, then we can use functional-oriented points, or directly using time estimates with Kanban tools.  With such work, there is a flow of handoffs and we can level that flow by measuring each step (however we’re still trying to optimize the delivery of value for the entire flow, not individual steps).

Thinking story points are precise
Story points are meant to be a fast and efficient way to continually estimate and reestimate the cost of features (stories).  This helps us prioritize them and to foster conversations that don’t take up enormous amounts of time.

There is a prevalent myth that the bigger our plan, the more accurate it is.  In reality, it’s the opposite.  The better plan is the one that is frequently revisited as we learn more.  A hundred one-hour planning sessions held over two years produces better results than 100 hours of planning done up front.  Story points allow such frequent planning by being brief and focused.  An hour or two of backlog refinement every sprint is usually enough.

Translating story points to time estimates
Story points are often directly converted to time estimates.  There are a few problems with doing this:
  • Time estimates depend on who is doing which part of the story.  Any two developers can have a magnitude of difference in the amount of time it takes to complete something based on skill, experience and other factors.  Cross-functional stories don’t provide this information.
  • As rough, relative measurements, story points are not precise enough to convert to time.  This is a reason that we plan using time for a limited horizon, such as a sprint’s duration.  A story that takes 5 minutes to assign points to might take an hour to task out in sprint planning.
So what good are story points?
Story points exchange the illusion of detailed up-front planning (big documents, detailed schedules) with more frequent and brief planning sessions based on the emergent knowledge we gain over the course of developing a game.  We learn the true cost of making something fun based on execution, not documentation.

By continually updating a forecast based not just on promises made at the start of the game, but based on what reality is telling us during development allows us to better manage the outcomes.  Agile planning is better at hitting fixed dates with quality than any other approach.

The big challenge is adopting the mindset to use them correctly.  Too often our stakeholders are not aware of the power of such tools and want specific answers up front.  That’s the biggest challenge.

Wednesday, October 18, 2017

Agile game development training course is now free

The "Agile Game Gems" video training course is now available for free.

This 48-minute course is a useful resource to introduce and refresh game developers on the purpose and vocabulary of agile game development.

I will be using this course to prepare students for my training cso we can better hit the ground running on adopting and improving agile. 

Links for the course's reference material are linked to in the video's description.

Tuesday, August 15, 2017

Mobile Game Lesson - Rapidly Responding to Change

I’m slightly embarrassed to admit that the mobile game I play the most is Solitaire.  Solitaire is a great game to play when you are waiting to board an airplane or when trying to fall asleep in a hotel.  I do both of those things quite often.

The first mobile Solitaire game I played had great graphics and some good features.  It also did a few things that annoyed me such as having slow animations when “undoing” moves and not having a feature that quickly completed the hand once all cards were exposed (a Solitaire hand is "solved" once all cards are face up),

Then one day, after Apple upgraded iOS, the game just kept crashing on startup.  This kept up for a few days.  Because I was traveling that week, I was having some serious withdrawals.  So I ended up trying a few competitive Solitaire games and finding one that not only worked with the latest iOS, but also had fast undo animations and the autocomplete feature.  I ended up buying that game and never going back.

By chance, I ended later ended up working with the studio that made the original game and was able to meet the small team that was supporting it.

It turns out that the team were unaware of how the market perceived this very profitable game of theirs or what the competition was doing.  Part of the problem was having a very deep feature development pipeline that didn't respond to change very well.

It was an eye-opening experience that such a simple game could suffer from the problems that larger games could as well.  The mobile game market can be very fickle.  Crashes, exploits or the emergence of a similar game doing something better can change you market overnight.  You need to be able to have some bandwidth set aside to deal with them.  Plan on change!

Thursday, July 13, 2017

Why a "Scrum for Video Game Development" Course?

What areas does a Scrum for Video Game Development course cover that others don't?

Plenty. Video game development has unique challenges that other industries do not.  For example, death-march-crunches are harder to avoid.  Mainstream app development can manage crunch by managing technical debt and scope.  Video game development has more varieties of debt (art, tech, design, content production) that have to be managed differently. We cannot just cut content production for schedule reasons and ship four hours of single-player gameplay to meet a schedule. This requires a different approach.

We discuss this and many other game-development-unique approaches in my interactive workshops.  I invite you to bring your questions or challenges to discuss. We'll share what many studios have discovered through experimentation and practices that were captured in my more recent book: Gear up! Advanced Game Development Practices.

Typical areas of discussion include:
Come join the next Scrum for Video Game Development course and bring your questions and experience.

Wednesday, June 21, 2017

Sprints or a Kanban?

There seems to be some confusion that using Sprints or a Kanban is a competition of sorts with "one being better than the other".  Might as well argue whether a hammer is better than a saw.  Anytime someone gives a blanket statement that "one is better than the other" it means they misunderstands both.

Sprints are more suitable to complex problems that a cross-discipline team will swarm on to solve.  Complex work has "unknown unknowns" that require experimentation and defy planning and estimation.  The time-box is a limit of time that is established to touch base with the business side and to replan our next move on a complex mechanic.

A Kanban, a way to visualize a lean flow, is used for complicated work.  Complicated work has "known unknowns", like creating levels and characters for a game with established mechanics.  The variations are manageable. It is more predictable and uses hand-offs of work through a flow.

Using Sprints to manage complicated work results in batched work and an artificial division through sprint planning and review. It hides discipline inefficiencies and leads to split stores which create no value individually. Imagine the cost of buying a house if every one was a custom concept home built by a guild of craftspeople.

Using a Kanban to manage complex work results in turbulent flow that either creates inefficiencies and a lack of transparency or artificial deadlines to call things "done" when they really aren't.  Kanban doesn't handle the back-and-forth of exploration very well. This creates debt and can limit the creative potential of a game. It's why we don't use an assembly line to design a new car.

I often see Kanban used for complex work because there is still an "upfront planning" mindset that thinks a new uncertain game mechanic can be broken down into bite-sized discipline centric steps and pushed though the teams. Sprints are abandoned because the developers cannot be trusted to take larger goals and break down the work as they see fit.

Choose the best tool for the work.  Wielding a hammer with great skill to cut long pieces of wood into smaller ones isn't as useful as using a saw.

Saturday, June 10, 2017

Department Silos Can Kill

In 2002, GM engineers discovered that the ignition switches on some low-cost cars had understrength  springs.  As a result, a heavy keychain or knee-bump could switch the ignition for those cars off.

It was considered a rare occurrence and not exceptionally dangerous.

Before the defected part was recalled in 2014, it was blamed for over 120 deaths.  Many of those who had died were young; parents had bought the low-cost cars for their children, considering them safe.

The systemic reason for GM ignoring the severity of the problem was that the engineers who designed the ignition switch were not familiar with how the ignition switch impacted other components of the vehicle. They weren't aware that switching off the ignition disabled the power steering and airbag deployment circuits.  Disabling power steering and airbag deployment was a deadly combination.

Unintended component interaction is why I discourage the creation of most component teams.  Graphics teams, physics teams, audio teams, etc. all sound efficient and they are efficient in creating graphics, physics and audio systems, but the cost of late integration and the emergent systemic problems is too great.

Players want games, not components, but at least our mistakes don't kill them.