I think everyone is different and theres is no 1 correct answer for all your questions, just keep trying what makes sense and what works for you and you will be able to make your own way.
For me at least i open a chart and start looking for patterns or ideas, can be on any chart on any timeframe. If i see a pattern or get an idea that i would like to test out i start testing it on different charts, different timeframes to see if the idea makes sense in another chart/timeframe. If the idea looks decent enough and robust enough i will continue working on it. Does it need more trades or fewer trades? More constraints or less? Does it look like it needs some sort of trailing stop loss or should it let all trades run to target/stop etc.
I will just keep working on that 1 idea until it either looks like shit or it looks decent enough to start putting into demo/live.
If i come up with an idea and im not in front of my computer ill write the idea down on my phone and test it later.
I dont have an exact formula for what i test and how i test it when i make an algo, i sort of try to figure out what my algo is missing or what it needs and then i try to implement that without curvefitting or destroying my algo while doing it.
I would say for some people having a written down rulebook for everything that needs to be checked and done is probably a slower way to go, but in the end it might give you better results! I dont know, i think this varies from person to person.
1 user thanked author for this post.