So the running count is nine after the first round (assume a full table and a lot of low cards. Rare, but for this example...) in a six deck game. 23 cards were dealt to six hands and a dealer.
The true count is 1.619.
Assume further a strategy that says bet 1 unit at TC < +1, two units at TC +1, and 4 units at TC +2. Assume one unit = $10.
At a table that would allow you to bet with chips and spare change, I think the bet called for with a true count of 1.619 would be $32.38. That amount could be wrong, but it's good enough for this example.
Nobody has the agility to arrive at this precise number while sitting at the table. Nobody is going to bet even $32 at a $10 table.
If I were to see a running count of nine after the first round in a shoe (where I started with one unit - $10) I might bet $10 again, I might bet $20. I'd think: RC of nine divided by six is... something more than one, but it's the first round, and anything more than $20 is too radical a jump. 23 cards out of 512 is a very small sample, and uh-oh, the dealer wants to know whether I'm in this round, because I'm taking all this time to calculate.
At an early point in a shoe at a low TC, it might matter, but not as much as when the TC is high.
If you bet 7 units at TC +3 and 9 units at +4 however, an accurate TC of 8.2 makes a big difference. Throw in a split and a double, and we're talking real money - 6 or 8 units difference between TC +3 and TC +4.
I do the math in my head as fast as I can, but I get too excited at high counts. I think to myself that with this high a count, being accurate doesn't matter. I probably err on the side of putting out MORE than a precise calculation calls for, probably miss downward swings in the TC and estimate that more of the shoe has been dealt than is true.
I don't know how bad this tendency is. Being more aggressive at a high count - essentially increasing the slope of the ramp in mid-shoe - might be beneficial. But this is an endeavor that lives on a razor-thin margin, analyzed and strategized by folks with more than an "impression" or a "feeling". I am guessing that there are those who will say that more significant digits matter a lot. But are they in the real world betting $32 at a $10 table?
So how to arrive at a more precise wager? Is there a general approach? Truncate the TC so as not to be too exposed? Round up to gain an advantage? Bet 8 units when the TC is "not quite" +4? At a positive count, don't worry, be happy?
While he might believe that many over-estimate things, I've no doubt he could do it in his sleep, blindfolded. Why would he switch because of others?