How to select power direction over USB-C

Method 1: Reverse/inverse the sides/plugs of the USB-C to USB-C cable. Might works or not.

Method 2.1: Use a USB-C to micro-USB(male) cable and a micro-USB(female) to USB-C(male) adapter.

Method 2.2: Use a USB-C to USB-A(female) adapter and a USB-A to USB-C cable.

Method 2 should always work. However you cannot use PD charge then.

Explain for method 1: I've heard C-C cables have sides to determine host / device. But I can't find the source. From my personal experiment, it depends on cables.

In every case, the Mac tries to make power flow from the battery to the computer.

How do I reverse this? How do I make power flow from the Mac to the battery?

I believe @amdyes gave a workable answer with a mention of using two USB-C to USB-A adapters/cables back to back. The device with the cable having a male USB-A female receptacle will only source power. The device with the USB-A male plug will only sink power. Flipping the cable around will reverse the flow of power. This forces a flip in the roles that each device will go into regardless of how they negotiate power transfer.

Apple uses USB-BC, USB-PD, and it's own BrickId for communication between USB power sources and sinks. USB-BC allows for power transfer up to 7.5 watts, and always in one direction. USB-BC and BrickId were developed to allow for simple power bricks to charge devices. There's no defined method for USB-BC and BrickId to reverse power flow when both devices are able to source and sink power.

USB-PD allows for power transfer up to 100 watts, and switch direction of that power transfer. While USB-PD can manage power transfer between 7.5 watts and 18 watts it is expensive to do so. Expensive not just in dollars but also in the volume it would take for the electronics inside small devices, and expensive in the power consumed by the electronics needed to communicate by USB-PD.

Apple filled this gap between USB-BC and USB-PD with BrickId and uses it on their 12 watt power supplies. Older Apple USB-A power bricks rated for 10 watts (and likely any other power brick rated between 5 and 18 watts that may exist) used what is now a deprecated version of USB-BC and/or BrickId.

I'll see people complain that Apple is "doing their own thing" with BrickId, and therefore violating some USB specification. Since Apple is a member of the group that writes the USB spec, and enforces compliance to the spec, makes it unlikely that they are violating any USB spec. There's plenty of other electronics manufacturers "doing their own thing", which may or may not violate the USB specs.

Qualcomm saw the same gap between USB-BC and USB-PD that Apple did, but they filled this with their Quick Charge protocol. Quick Charge 4 is something of an umbrella specification which combines parts of USB-BC, USB-PD, and prior versions of Quick Charge along with a few new features. This is Qualcomm and friends "doing their own thing" but at least attempting to not step into where it might violate the USB specs.

There's other means of getting power into and out of a USB battery pack but what Apple and Qualcomm are doing have been the most popular. There's a lot of overlap in what Apple, Qualcomm, and others will do to maintain compatibility with as many USB devices as possible and therefore keep customers happy. What still happens though is that manufacturers will have to choose one out of the two or more options on how to solve a problem, because implementing all of them will be expensive or simply impossible.

Quick Charge and BrickId both use the data pins on a USB-A port to communicate the capabilities of a power source to a power sink. I suspect that your power pack is using one means to communicate, the MacBook is using another, and so nothing is communicated.

I went into some detail here to hopefully help people know what to look for in finding the problem, as well as describe why the problem even exists. This might not bring a solution but should give guidance on repeating the problem. The term "BrickId" is not likely to be found in any documentation as this is a term not meant for public consumption. Quick Charge and many other trademarks used to describe battery charging methods over USB are meant for public consumption. Searching the documentation for these terms should help in seeing where the problem might be.

I will admit my desire to go into some of this detail is a bit of a rant against complaints of some manufacturer "doing their own thing". That's what happens when two companies find the same problem, at the same time, and don't know the other is working on a solution. This sucks for us consumers to have to look for conflicts in standards but I can imagine far worse things to deal with.