A concrete floor covered with vinyl is not accurate for flashlight testing. Need a normal concrete, stone, metal or tile floor. I have a pair of zebras - for head and fot hand. I always worry about hand zebra glass as there is too little material to cover it. I don't think the flashlight will break if it gets a few good hits on the side of the bezel, but how will its water resistance change? Therefore, HDS is always paired with a Zebra, and with SF 6PX and a couple of smaller lanterns. The HDS bezel is not useless, it is durable and protects the glass from damage and at the same time maintains water resistance. I've dropped my Clicky many times on concrete and steel, there are quite a few dents, but it still works and remains waterproof (for my use). Rotary - I still do not really trust her in terms of water resistance of the swivel tail.
Predicting the future is a very difficult thing, yet few people succeed, even meteorologists with all modern instruments and satellites. I think it's the same with batteries. No person will be able to accurately determine the remaining flashlight time if they use several levels alternately, no matter - fixed or smooth. Only the processor in the flashlight can make it more accurately, and it will also not be accurate and does not predict the future, but only shows the current state except Nitecore. It can give a forecast in terms of running time, but it gives it on the basis of assumptions about the external environment embedded in it or the assumptions of the developers when they wrote this program. If the ambient temperature drops or rises too much, the operating time may change dramatically. Especially if the flashlight has a temperature-controlled brightness control.
The Nitecore in the TM series has a screen that can more or less accurately show the time, but with the proviso that it is configured for batteries with a certain capacity. H17Fx quite accurately reads the voltage on the battery. In Lupine, the processor constantly counts how much mAh was spent, but sometimes it makes a mistake if the flashlight was disconnected from charging before it was fully charged. But none still gives good accuracy. Alternatively, it would be possible to create a driver that could set the lamp's operating time, for example, 1 hour at one level, 10 hours at the second level, 1000 hours at the third level. Next, the timer starts and the flashlight will do everything to keep this time, that is, if the environmental conditions or the state of the battery change, it will regulate the light flux itself so that the battery is empty and turns off exactly by specified time. The flashlight also needs to detect the state of the battery that is placed in it, so as not just to follow the timer, but to change the timer 1 time based on the state of the battery just inserted. Accordingly, it must be able to work as a high-level battery analyzer. For example, if a battery with 50% charge is put in it, it will change the timer from 1000 hours to 500 hours and start counting from that moment. And he must have a screen that will show the information, or a radio link with a computer to show the data. In my opinion, this is redundant for flashlights, it is enough to carry spare batteries with you or a spare flashlight with spare batteries and always use the light sparingly and economically. And the systems for accurately determining the remaining operating time should be left to the mars/moon-rovers and satellites.
If Zebra made a version with a normally protected glass - a powerful steel bezel on a thread, so that it protrudes above the glass by at least 5 mm and 2 mm outside the diameter, and the button ring is also threaded, and powerful springs to fit batteries of different lengthsthen, that would be great! As well as thicker walls everywhere. But I'm already describing the Convoy))) The only thing that the Convoys are inferior to the zebras is the driver and the heat sink. In a zebra, it is better due to the one-piece body. But Convoy has a better selection of LEDs and maintainability.
As for control - control of 1 button will always have compromises and you will have to choose either one type or some of the others. 2 control buttons already gives more space. Better yet, have 5 buttons and a swivel magnetic ring - you can set up a smooth change in brightness with a button, instant access to the desired level from the off or running level, access to fixed levels by turning the ring, etc. But this obviously will not physically fit on a small lamp + it will be really expensive, respectively, there will be little sales and losses for the company. Maybe multi-button control can be organized in a smartphone, some kind of application for controlling an LED in a phone so that you can have both smooth and discrete levels, etc., but I think such an application would not be bought due to the fact that it no one needs a weak LED with complex management in the phone.
I don't like Anduril because there is no blocking the entrance to programming, as on HDS, HIVE on McGizmo, H17FX and some others. If you could make the necessary changes and then block so that the uninitiated (or I forgot the instructions) could not change anything, that would be good. Zebra also doesn't have a programming lock now, I don't like it either, but it's tolerable, since I can quickly change everything back if someone changes the settings, if I remember how I set it up