A KO counter works from one number: start the six deck running count at -20, add one as each 2 through 7 falls, subtract one for every ten and ace, and never divide. That is the system's whole pitch, laid out in our KO guide: one running total stands in for the conversion that balanced counts must perform at the table.
That raises a question worth testing: the count is built to drift upward as decks go by, so what is the number doing at half a deck, at three decks, at the cut card? We ran the experiment: two million shuffled six deck shoes, read one card at a time, with the KO running count and the Hi-Lo true count computed side by side at every half deck.
The drift is real and exact on average, and subtler than the framing suggests. Exactly one KO value meant nearly the same thing at every depth we measured. Every other value drifted.
What the simulation measured
No hands were dealt. The order of play does not change which cards come out of the shoe, and the count only cares about the fixed sequence the shuffle produced.
On the KO side, 2 through 7 count +1, the 8 and 9 count 0, and tens and aces count -1: 24 plus tags against 20 minus tags in every deck, so a full deck sums to +4. The count starts at the standard initial running count our KO article gives, 4 minus 4 times the number of decks, which is -20 for six decks. On the Hi-Lo side, 2 through 6 count +1, 7 through 9 count 0, tens and aces count -1, and the running count is divided by the exact decks remaining, the conversion our running count versus true count piece explains.
Checkpoints fell every half deck from 0.5 to 5.0 decks dealt. A cut card at 75 percent of six decks lands at 4.5 decks; our penetration and true count frequencies piece covers why that depth matters.
The climb is exact, the spread is not
The first result is almost boring. The average KO running count climbs by exactly 4 per deck dealt, which is what the tags predict: the mean at any depth is -20 plus 4 times the decks dealt, and the measured means land on that line at every checkpoint.
| Decks dealt | Mean KO | Std deviation | 10th pct | Median | 90th pct | At or above 0 | At or above +2 | At or above +4 |
|---|---|---|---|---|---|---|---|---|
| 0.5 | -18.00 | 4.48 | -24 | -18 | -12 | 0.0% | 0.0% | 0.0% |
| 1.0 | -16.00 | 6.05 | -24 | -16 | -8 | 0.5% | 0.2% | 0.0% |
| 1.5 | -14.01 | 7.02 | -23 | -14 | -5 | 2.7% | 1.3% | 0.6% |
| 2.0 | -12.01 | 7.64 | -22 | -12 | -2 | 6.6% | 3.9% | 2.1% |
| 2.5 | -10.00 | 8.00 | -20 | -10 | 0 | 11.7% | 7.5% | 4.6% |
| 3.0 | -8.00 | 8.10 | -18 | -8 | +2 | 17.7% | 12.1% | 7.8% |
| 3.5 | -6.00 | 7.99 | -16 | -6 | +4 | 24.5% | 17.4% | 11.7% |
| 4.0 | -4.00 | 7.64 | -14 | -4 | +6 | 32.3% | 23.6% | 16.3% |
| 4.5 | -2.00 | 7.02 | -11 | -2 | +7 | 41.6% | 30.9% | 21.7% |
| 5.0 | +0.01 | 6.04 | -8 | 0 | +8 | 53.3% | 40.2% | 28.1% |
The spread tells a more interesting story. The standard deviation grows from 4.48 at half a deck to 8.10 at three decks dealt, then shrinks back to 6.04 at five, a shape consistent with drawing without replacement, where the uncertainty of a running total is largest halfway through the pack.
The median stays below zero until five decks are dealt, so most of the shoe is played at negative values, as our KO guide tells new counters to expect. Even at the cut card, only 41.6 percent of shoes show 0 or better, and 21.7 percent show +4 or more.
21 Trainer ProKO alongside Hi-Lo, Omega II and Zen, with a true count drill and index deviation training. Virtual chips only.
Get the app
The one value that holds its meaning
The deeper question is what a KO number is worth. The design claim behind unbalanced counts, which our balanced versus unbalanced comparison lays out, is that, given the right starting value, the drift approximates the true count division, so one fixed running value can stand in for a converted threshold. We tested how well that holds by asking what the Hi-Lo true count averaged, at each depth, among shoes showing a given KO value.
| KO value | 1.0 | 1.5 | 2.0 | 2.5 | 3.0 | 3.5 | 4.0 | 4.5 | 5.0 |
|---|---|---|---|---|---|---|---|---|---|
| -4 | +2.20 | +2.03 | +1.83 | +1.57 | +1.22 | +0.73 | 0.00 | -1.22 | -3.66 |
| 0 | +2.93 | +2.85 | +2.75 | +2.62 | +2.44 | +2.20 | +1.83 | +1.22 | 0.00 |
| +2 | +3.30 | +3.25 | +3.20 | +3.14 | +3.05 | +2.93 | +2.74 | +2.44 | +1.83 |
| +4 | +3.63 | +3.66 | +3.66 | +3.66 | +3.66 | +3.66 | +3.66 | +3.66 | +3.66 |
| +6 | +4.05 | +4.12 | +4.19 | +4.27 | +4.40 | +4.57 | +4.87 | +5.49 | |
| +8 | +4.48 | +4.57 | +4.71 | +4.87 | +5.13 | +5.49 | +6.09 | +7.32 |
The table shows average Hi-Lo true counts by decks dealt; blank cells had fewer than 200 shoes at that value and depth. One value stands out: with the -20 start, a KO of +4 corresponds to an average true count of +3.66 at every depth from 1.5 to 5 decks dealt, and +3.63 at one deck, a cell with only 483 shoes. Below it, the same KO number stands for less and less as the shoe goes on: a KO of 0 averages a true +2.93 after one deck but +1.22 after four and a half. Above it, the same number stands for more and more: a KO of +8 averages a true +4.57 after two decks but +6.09 after four and a half.
So the approximation holds, on average, at exactly one value and loosens with distance from it. A fixed running value keeps its average meaning only at +4; below it, the same number marks a weaker shoe the deeper the deal goes, and above it, a richer one.
How often a KO value means a rich shoe
Averages can hide spread, so the last cut is blunter: among shoes showing a given KO value, what share had a Hi-Lo true count of at least +2?
| KO value | 1.0 | 1.5 | 2.0 | 2.5 | 3.0 | 3.5 | 4.0 | 4.5 | 5.0 |
|---|---|---|---|---|---|---|---|---|---|
| 0 | 100% | 98% | 95% | 88% | 79% | 67% | 52% | 35% | 18% |
| +2 | 100% | 100% | 99% | 98% | 95% | 90% | 83% | 72% | 56% |
| +4 | 100% | 100% | 100% | 100% | 99% | 99% | 97% | 95% | 92% |
The shares tell the same story. A KO of 0 early almost always coincides with a true count of +2 or better, because zero sits far above the -16 average at that depth; at the cut card the same value coincides only 35 percent of the time. A KO of +4 stays at 92 percent or higher at every depth measured, though even its share slips late in the shoe while its average holds.
What the simulation leaves out
This is six decks only, and other deck counts start from different initial running counts. The Hi-Lo side was divided by exact decks remaining, which no player has: a live counter estimates it from the discard tray, a skill our deck estimation practice piece covers, and estimation error would blur these tables. And the simulation follows the count, not the game: nothing here says what anyone should wager at any count value.
The Hi-Lo true count is a yardstick, not the truth: it measures how concentrated the high cards are per deck remaining, a fair reference for what a KO number is worth. Counting cards with your head is legal in the United States, casinos are private businesses that may refuse service to players they suspect of counting, and no technique removes the risk from the game. Our piece on whether card counting is illegal has the details.
Practicing with the drift in view
Three takeaways carry into practice. Expect negative numbers most of the shoe; the median does not cross zero until five decks are dealt. Know where you are, because the average count at any depth is -20 plus 4 times the decks dealt, and distance from that line says whether your shoe is running rich or poor. Treat fixed running values with depth in mind: their meaning slides except at one steady value.
None of that requires money or a casino. The 21 Trainer Pro upgrade teaches four counting systems, Hi-Lo, KO, Omega II and Zen, with a true count drill and index deviation training, all on virtual chips. Counting shoe after shoe from the correct start makes the drift something you feel rather than a number you were told.
21 TrainerFree on iPhone and Android. Instant right or wrong feedback with a clear explanation on every hand.
Get the app
Frequently asked questions
What number does the KO count start at for six decks?
Minus 20. The standard initial running count from the Knock-Out book is 4 minus 4 times the number of decks, so 0 for single deck, -4 for double deck, -20 for six decks and -28 for eight. The count then climbs about 4 per deck dealt on average, averaging zero once five decks of a six deck shoe are dealt.
Does the same KO running count mean the same thing all shoe long?
Mostly no. One value, +4, corresponded to an average Hi-Lo true count of +3.63 to +3.66 at every depth we measured. Below it, values meant less and less as the shoe went on: a KO of 0 averaged a true +2.93 after one deck but +1.22 after four and a half. Above it, values meant more and more, so a fixed threshold holds steady only at that value.
Why is the KO count negative for most of a shoe?
Because it starts at -20 in a six deck game and climbs only about 4 per deck dealt. The median stayed below zero until five of six decks were dealt in our simulation, so most of the shoe is played at negative values. That is built into the design, not a sign of counting badly.
Is counting cards with the KO system illegal?
Counting cards with your head is legal in the United States. Casinos are private businesses and may refuse service to players they suspect of counting, with jurisdictional differences in how far that goes. No technique, KO included, removes the risk from the game; counting at best yields a small long-run edge with large short-term swings.
How do you practice the KO count without a casino?
A shoe of cards at home is enough, and a trainer makes the loop faster. The 21 Trainer Pro upgrade teaches KO alongside Hi-Lo, Omega II and Zen, with a true count drill and index deviation training, all on virtual chips. Practice never needs real money, and accuracy is the only thing at stake.