It's easier when graphs from different batteries are on the same graphs, then it's obvious that the one with the taller and longer curve is the better battery. HKJ's website allows two batteries to be compared up to 5A.
A taller curve basically means it offers the potential for a brighter light, but that doesn't mean much unless you have a flashlight that pulls high current. Some great batteries for this are the LG HE2, Sony VTC4 and VTC5, and the Samsung -20R and -25R. These typically don't have very high capacity.
A longer curve means it has great capacity (mAh) and will stay on longer. The Panasonic NCR18650B is what most people will think of. While it has high capacity, it's not good at delivering high current.
I believe all of us tests pairs of batteries, which helps average out our own human error, but can also show if batteries don't provide consistent results. For example, Ultrafire batteries frequently perform inconsistently, which can make them dangerous in multi cell lights. Many batteries from Japanese manufacturers are often very closely matched, and are much safer to use in multi cell lights. I'd call consistency one of the strongest indicators of quality, if not the greatest.
When shopping for a battery, you should balance your need for current and capacity, along with price, plus other things like if it's a button or flat top, weight, thickness, length, quality and safety (protection circuit or IMR chemistry).
HKJ would be the best person to do a longer explanation. I may give it a shot if he doesn't, but it'll be after I move.
Does this mock up of HKJ’s Comparator look about right, correct me if i’m wrong on anything thanks.
.
.
As i understand it in simple terms, the more that the discharge curve forms a right angle towards the top right hand corner of the graph, the better the cells are for handling that load.
The higher the load on the cell, the more the voltage sags and results in reduced capacity.
Also when a pair of cells are tested, ideally the lines would be on top of each other so you can only make out one discharge curve for the two tested cells.
Now, if only we could get our hands on some of those CRX IO cells…
I do not give high rating to unknown brands, but I may say they have a fantastic performance.
Anyway if I got that kind of curve I would start to look for faults in my equipment (The different current traces will never be on top of each other, especially not when you get up to a few amps).