What do you expect to happen? What do you expect to happen? This is a question that you're probably asking yourself on a daily basis. Usually you don't realize you're asking yourself this question. But you do this explicitly when looking at work that you need to do. This is a continuation here on Developer Tea of our series about backlog psychology. And in today's episode, I want to talk about this idea of expectation. We all have expectations. These are beliefs about what we think will occur. That's all an expectation. An expectation really is beliefs about what we think will occur. But usually, usually our expectations are in some way different from reality. And this difference may be subtle. And usually it probably is relatively subtle. Or it may be drastic. We talked in the last episode about the importance of practice. The idea that your processes should create some kind of stability. Some homeostasis. Repeatability. So that you can get into the habit of practice. So there's not so much variation that your iteration from one to the next week or one to the next month looks wildly different. This allows your team. It allows your engineers. It allows you to practice. But what does practice provide? Practice on its own provides us experiences. Experiences. If you've ever filled out a resume. If you fill out an application for a job, you're asked about your experience. Your years of working are an analog to the number of experiences you've had. And the number of experiences you've had provides you usually with better expectation machines. But even for people who have a ton of experiences. Whether that's a long time in the industry or a lot of compressed experience. There is still a hurdle. In our expectation building. And it's related in part to something called a binary bias. And there are other biases or cognitive dissonance that may contribute. But the basic idea of a binary bias is that we take continuous information. And we try to create categories out of it. In particular for expectations. We have a binary bias. To believe that either something will happen or it won't. That there is a 100% chance or a 0% chance that our expectations are correct. But the truth is that we will almost never be 100% correct about our expectations. This is also why our estimates for how long something should take. Are almost always going to be wrong in one direction or the other. Especially for things that are larger in scope or in complexity. And so the bug in the system that we're finding here. The backlog psychology kind of error that we might make. Is putting all of our eggs in one basket. Imagining that our one expectation is the only one we should prepare for. The truth is that we cannot predict the future. Both in terms of simple things like estimating. But also in terms of our expectations. In terms of what actually happens. What happens when we build this feature? What happens in a couple of months after that feature is built? Humans are not very good at imagining exactly what will happen. But we are good at something. We are good at something that can help us in this particular arena. And that's the adjustment that I want you to make. This is the takeaway for today's episode. Your adjustment should be that you expect more than one thing. To happen. On a probabilistic basis. Expect more than one thing on a probabilistic basis. What does this mean? It means that instead of having a singular expectation for a particular item of work. Whether that's your estimation expectation. Or how you think it's going to be delivered. What exactly the solution is. Instead of developing a singular expectation. Develop a handful. Usually for something as simple as a task. Two or three is enough. I'd recommend going to three. So you avoid other biases related to either or. That tend to show up when you only have two options. But having more than one expectation. And assigning some kind of probability ranking to them. This is not saying that you're assigning an explicit percentage number. Because that again requires precision. And humans are not very good at precision. What we are good at. Is comparison. If you were given three outcomes. Three outcomes from a particular work item in your backlog. And you were told to rank them in order of most likely to occur to least likely to occur. Humans are much better at this task. Than they are with assigning an absolute probability to each of those items. The truth is the thing we expect to occur. May only have a 40% chance likelihood. Of occurring. But it's greater than any other item in the list. Now it's very possible in that same scenario. That another item has a 39% chance of occurring. And we should be able to identify this. By expanding our expectations. Instead of only having one expectation. You're expecting a variety. You're expecting a range of possible outcomes. It would be foolish to say that you're expecting. A singular outcome. And exactly in line with this. I also recommend that if you're doing. Some kind of estimation. That you lean towards a range. A similar strategy. If your organization does require atomic estimations. Where a range is not accessible. Or possible for your tooling maybe. Lean towards the high end. Establish a range. And then take the high end for your estimate. This tends to be a safe route to go. And the reason. For this is very simple. There is a lot of complexity. That we don't know about. It's very possible. That your work scope. Goes towards that high end. In fact. Most research indicates. That estimation tends to be correct. If we use the high end. As our primary estimate. That tends to be actually kind of the center. Of the distribution of estimates. Versus actual time taken. For given tasks. But the loss of overestimating. Is almost always. More acceptable than the loss. Of underestimating. In any case. When you can. When you can. Consider multiple outcomes. That is what you should align. Your expectations to. Thanks so much for listening. To today's episode of Developer Tea. If you enjoyed this discussion. Then please do two things. One. Subscribe in whatever podcasting app. You're currently using. If you've already subscribed. Send this to someone else. You think will enjoy it. That is a great way. To help Developer Tea. Grow. To new audiences. To new ears. So I definitely appreciate. When you all share this. With new people. Who have never heard of the show before. And then the second thing is. Come and join our Discord community. Head over to developertea.com. Slash Discord. To join totally free today. Thanks so much for listening. And until next time. Enjoy your tea.