This is not new -- Pavel Tsatsouline described it in detail in an article at Dragon Door which is where I learned of it -- but it's interesting and hopefully my description is briefer and more easily understood.
Volkov has won IPF Masters, Moscow, and Russian championships in the bench press, with a raw bench of over 550 lb and 639 equipped. This apparently is the only way that he trains, so it cannot be too shabby a training protocol, as unusual as it is.
I find differing interests in it of being an interesting change-up workout, and also being a safe albeit unusual way to get an idea of 1RM in an exercise while also getting a good workout and not generally interfering with regular training. Very often a lifter puts off or just flat doesn't get a 1RM or a close estimate of one because of not wanting to "waste" a workout or because it might interfere with other training days, or being concerned for straining or even perhaps injuring himself. Not the case here in any of these regards, I think.
Or as for pure workout value, it allows fairly high number of sets while not being overtaxing on the CNS, yet still with, if desired, very-near-maximal effort at the end.
The method is simple.
Let's say that you don't have an exact target top work weight in mind, for example because you've never done this before.
If you don't, start with about 55% of what you guess for your 1RM. The exact value is not important. Either for your estimate being exact, or the figure having to be exactly 55%. But shoot for about that.
Decide how many reps you want to be doing per set. I recommend 2 or 3, though Volkov did anywhere from 1 to 5, though most commonly 2 or 3.
All sets are the same number of reps, unless unintentionally unable to do that many in the last set.
Then add some modest amount of weight per set, the same amount each set. The interval Volkov uses is about 4% his 1RM but again this need not be exactly followed. For example, if that would calculate to say 16 lb, don't worry, 15 lb would be fine.
End the sets when judging it unlikely that an additional increment can be handled.
Volkov does not do it, but if I think I probably can't get another set with the full increment but could do so with a little less, I'll make the last increment smaller.
So, example workout for estimated 500 lb 1RM. I pick this value for the sake of round figures. Let's say you plan to do doubles.
0) Very light warmup if desired.
1) 275 for 2
2) 295 for 2
3) 315 for 2
4) 335 for 2
5) 355 for 2
6) 375 for 2
7) 395 for 2
8) 415 for 2
9) 435 for 2
10) perhaps 455 for 2.
The article says nothing of Volkov's rest times and I have no other information on it. I take about an hour for this, allowing 1 minute rest at the very beginning, then fairly soon 2 minutes rest, and pretty soon getting into 3 and 4 minutes rest. By the last couple of sets, 6 minutes rest.
If doing it for the sake of an estimated 1RM for planning purposes, I'd say that a 1RM probably ought to be at least 5% higher than the highest double, more likely 10% or could even be a touch more; or if inadvertently dropping to a single on the last set, at least 5% more than that. And if you are quite sure that you could have gotten the inadvertent single (you'd planned a double) with say another 10 lb, for the sake of a good estimated you could add that 10 lb on, then add another 5% I believe. (We are not determining a factual 1RM, but a good estimate suitable for planning purposes or having a general idea.)
It's a nice workout and quite different from the usual.