TabletMark is useful in that you can install it launch the program, click the battery test and it will (should) run until the battery dies. Then when you relaunch it after a charge it will detect it was in a battery run and report the results. That is also useful for making a comparison later to your prior run or to a like system to determine if a system has a different than baseline battery life. Example if user A and B report 4 hours and 8 hours battery life but TabletMark reports 10 hours for both the difference is usage profile with "all other things" being equal. I think TabletMark is a bit on the weak side as far as usage goes so times will be a bit longer than maybe the average user would experience. Its useful with the right perspective but it may also confuse and frustrate people unfamiliar with these tools.
As the great Greek Testing and Benchmarking philosopher Perfomaximus once said "Testing proves that testing works".
As you are finding out sometimes we have processes running amuck and those will impact battery life significantly. These rogue processes will also skew your battery reports same as if your baseball player was in a batting slump. From the example above, "ALL other things", may not be equal. In the real world, with users being different, this has always been the case. However, when we have Rogue Processes exacerbating the situation it makes questioning everything even more critical and comparisons that much more difficult. I had not experienced these rogue processes until the last few days, hopefully they are not common but they must be considered when users report incidents.
Lastly your usage profile, sans the rogue processes, is the ultimate determinate factor of your battery life. If you run x days with out rogue processes, look at the run time for those days as an indicator of where your usage profile falls. over time your true average will emerge. powercfg /energy will give you an rundown of what used "energy" as this report can be run even if you were plugged in. Then you have to decide is there anything I can do to affect these results?
One user here was quite vocal and adamant about NOT doing anything that would help his battery life; if anything he did just the opposite. Data tells us things, what we chose to do with that information is a completely different matter. There are manipulators all around us in this world with all sorts of agendas you have to deal with. Can you believe what you see, read, hear? is it real? what is real? What you see is real for you but may not be the true picture. Keep digging, the scientific process does not know politics or brands and if it does its corrupt.
Long winded way of saying your mileage may vary.