Exactly this! When doing R&D for new systems or potential improvements, costs and commissions should be the last thing on your mind. We're not even talking working trading software yet at such a stage, unless for those with very active profitable systems already, which are probably 1% of the 1% or even less portion of all traders. In practice, commissions is a big first hurdle. For puny accounts (<$30000), it may turn trading into losing proposition. Ditto for daytraders or very active trading. That and spread do indeed turn most theoretical systems into the category of unworkable. Thus there's frustration about this topic. Here's at least 5 stages of trading systems development: 1. R&D (reading, bright ideas, brainstorming, checking out "stuff") 2. Rules (creating trading plan, coding) 3. Testing (backtesting, simulations) 4. Execution (testing waters, forward testing, scaling up) 5. Death (End-Of-Life, scaling down, manual trailing) At any point in time, people are usually stuck in one of these 5 modes for longer periods of time, depending on where they are in the system lifecycle-wheel. All of these stages require quite different mindsets and trains of thoughts, so very rarely someone is working on more than 2 of these stages simultaneously. It's actually very unnatural for people to multitask, and each of these stages require full and clear mental capacity. This also colours how traders think about what is "true" and not, at any point in time.