The Changing Nature of Estimates

April 26, 2019

When I was developing the holiday based Sleuthhounds games, The Valentine’s Vendetta, The Halloween Deception, and the The Yuletide Tail one thing that was tremendously helpful was that I knew I wanted to release the games for their respective holidays. That meant that I had hard stop dates for when the work had to be done, which in turn meant that decisions of what would and would not be included were, in a way, easy to make. If anything threatened the finish date then it had to be cut. It was that simple. With my upcoming Sleuthhounds: Cruise game I don’t have that fixed “do not pass go” date. Consequently, estimates and my ability to stick them are much more important and also much harder.

The tricky thing with estimates is that they’re approximations. Best guesses, ideally formed on at least some previous historical experience, but still guesses. After five other games in the Sleuthhounds series I’ve come to understand the types of techniques needed to realize most things in the newest game and roughly how long they should take to implement. But “most” and “roughly” are not absolutes.

Back at the end of January, when I completed implementing the critical path of Act 2 of the game, I made a very rough estimate for how long Act 3 would take to reach the same state. At that point, I only had very broad strokes determined for what would happen in Act 3 with a lot of question marks to resolve. My high level guess at that time was that Act 3 would be finished roughly in the middle of April.

Come the middle of March and I had filled in most of the question marks and had a much better understanding of the important story beats in Act 3. The high level guess changed into a more detailed estimate, which put the completion date on April 15th. This was based on having most of the story and puzzle design worked out although there were still a few areas of uncertainty.

Since that time, work has been proceeding at a steady pace on Act 3. Most of the implementation time has been very close to the estimates with only a few items coming in appreciably under or over their expected time. However, obviously April 15 has come and gone. The reason for the change in date, or more appropriately the change in the estimated date, stems not from the technical side of implementation but from the more creative side of story development. As I wrote earlier this month, when I actually got to the point of working on various scenes I found that it was necessary to expand them story wise.

Act 3, with its non-linear nature, has a lot of story bits that need to be lined up. Inevitably, when working on a story, holes are discovered. Places where things don’t quite make sense or where I realize that a needed scene is missing. It’s making sure the story is solid more than anything that has caused a change in the date. That’s because as I got more into the details of the story and learned more about it overall, I developed more knowledge on what needed to happen. In the case of Act 3 that meant the addition of several scenarios that I couldn’t have even conceived of back in January.

Such is the nature of estimates. They’re best guesses but guesses that need to be refined over time as more about a problem is learned. Sometimes estimates can shrink but mostly estimates will increase. That’s simply a byproduct of learning that there were aspects that hadn’t or couldn’t have been considered when the earlier estimates were made. Rarely do you learn more about a problem that results in there being less work.

All that said, my last reported estimate was that the critical path for Act 3 would be completed by the end of April 26th, which is today. However, I’m pleased to report that the critical path for Act 3 was actually completed yesterday. Not only that, but I incorporated two additional sequences that I hadn’t even considered from two weeks ago.

My original guess called for just over ten weeks to implement Act 3. In the end, it took about eleven and a half. Given the high level point at which I was starting I consider that a quite acceptable margin of error. So what’s next now that Act 3 is scripted? That’s easy. I’ll be moving onto Act 4, the final act of the game. How long will that take? I’ll let you know as soon as I’ve estimated it.