The cost of starting the shoe at a high bet
By Canth · July 20, 2026
A common piece of cover is to keep a bet out on the first hand of a new shoe instead of dropping to the table minimum at the shuffle. Spreading up late in a shoe and then betting the minimum on the very next hand is a tell, so some counters leave a bet out for one hand to smooth the change. That first hand is played into a neutral count, where the house has the edge, so the cover costs expected value. This is a measurement of how much, run through the simulator across a few common games.
How we measured it
Two players count Hi-Lo with the standard index deviations and play identically. One drops to the table minimum on the first hand of every new shoe, the way the count says to. The other holds whatever bet it had out when the shoe ended, capped at a set cover amount. End a shoe at a high bet and you open the next one at the cap; leave it at a middling bet and you open at that middling bet; leave it already at the minimum and you open at the minimum, because there is nothing to hide. After that first hand, both players are back on the same bet spread. They are dealt the same shoes, so the gap between them is the cover and not luck.
The bet never drops fully across the shuffle. Wherever you were when the cut card came out, you hold that bet (up to the cap) for one hand, then spread normally. It is a tax paid on the shoes where you had money out, carried one hand into the next shoe where the count is neutral. The bigger the cap, the higher the ev cost. Only the bet on that first hand changes; the cards dealt do not, so the two strategies stay matched shoe for shoe.
Example · $25 table, cover set to $200
| Last shoe ended on | First hand of the next shoe | Why |
|---|---|---|
| $25 | $25 | already at the minimum, nothing to hide |
| $50 | $50 | carry it, you were still spread up |
| $200 | $200 | carry it, right at the cover |
| $400 | $200 | carry, but capped at the cover |
Only that first hand is set this way. From the second hand on you are back on your normal ramp, betting the count.
The player plays heads-up, Hi-Lo with the standard H17 index deviations (the Illustrious 18, insurance at true count +3), DAS, no resplit aces, no late surrender, H17. The table pace is set per game: 130 rounds an hour in single deck, 170 in double deck, and 250 in the shoes. Each game runs its own ramp, a realistic green-chip spread from a $25 minimum up to $400, built by the bet-spread optimizer for that game’s edge. Risk of ruin is quoted at a $25,000 bankroll, small enough that this aggressive spread carries real risk, so the cover’s effect on it is visible rather than pinned at zero. The true count is estimated to the deck fraction an AP would use for each game: a quarter deck in single deck, a half deck in double deck, a whole deck in the shoes. We tried three cover levels: holding $50, $100, or $200on the first hand. Each number is the mean of five runs of 600 million rounds, three billion rounds a figure. Cost is reported per hour and per round; the per-round figure does not depend on the game’s pace, so it compares cleanly across games.
What it costs, by game
“Win rate” is the edge when you drop straight to the minimum, shown per hour and per round. The cover columns show how much of that you give back by holding $50, $100, or $200 on the first hand instead, per hour and per round.
| Game | Win rate | $50 cover | $100 cover | $200 cover |
|---|---|---|---|---|
| Single deck0.25-deck cut · H17130 rds/hr | $404/hr$3.11/rdN₀ 23h | −$0.10/hr−0.07¢/rd0.0% of edge | −$0.29/hr−0.23¢/rd0.1% of edge | −$0.58/hr−0.45¢/rd0.1% of edge |
| Double deck0.5-deck cut · H17170 rds/hr | $240/hr$1.41/rdN₀ 49h | −$0.37/hr−0.22¢/rd0.2% of edge | −$1.12/hr−0.66¢/rd0.5% of edge | −$2.14/hr−1.26¢/rd0.9% of edge |
| Six deck1-deck cut · H17250 rds/hr | $96/hr$0.39/rdN₀ 199h | −$0.21/hr−0.08¢/rd0.2% of edge | −$0.56/hr−0.22¢/rd0.6% of edge | −$0.95/hr−0.38¢/rd1.0% of edge |
| Eight deck1-deck cut · H17250 rds/hr | $75/hr$0.30/rdN₀ 279h | −$0.16/hr−0.07¢/rd0.2% of edge | −$0.43/hr−0.17¢/rd0.6% of edge | −$0.77/hr−0.31¢/rd1.0% of edge |
These are simulated figures. The per-round cost does not depend on the table pace, so it is the fair way to compare games run at different speeds. Average bet size across a whole session barely moves: the cover changes one hand per shoe, not the ramp.
Where the cover costs the most
The tax is paid once per shoe, so the cost turns on two things: how big a bet you carry, and how often it occurs. In raw dollars the cover costs the most in double deck, about two dollars an hour at a $200 cover, because that game runs a large spread where the shuffles come fast and the count regularly climbs enough that you end shoes near your top bet.
Measured against the win rate, double deck in this situation costs the most. Single deck’s edge is so large that even a $200 cover is about a seventh of a percent. The deep shoes are the opposite: the cover costs under a dollar an hour, but they earn so little that it comes to about one percent of the win rate.
The heavier the cover, the more it costs
Obviously, how much you leave out matters too. Holding $200 on the first hand instead of $50 is four times the bet into the same neutral count, and it costs four to five times as much. There is a straight trade-off between how convincing the cover looks and what it takes off your win rate.
Double deck, H17. The cost climbs with the bet you hold: a higher cap both raises the first hand and lets more of your mid-sized finishes carry, all at a true count near zero where the house has the edge.
Variance moves a little
Betting one extra big hand per shoe does add variance. It is a small effect (one hand in a shoe of fifteen or more), and it is largest in single deck, where a big cover towers over a $25 minimum. The table puts the drop-to-minimum baseline next to the $100 cover. SD/hr rises about a percent in single deck and a few tenths in the shoes. At a $25,000 bankroll, risk of ruin runs from about half a percent in single deck to roughly nine percent in eight deck, and the $100 cover lifts it a few tenths of a percent in each game. The bite is largest in the deep shoes, where the base risk is already highest and the thin edge gives the added variance the most room to work, so it is worth weighing before you hold a large cover on a small bankroll.
| Game | SD / hr | N₀ | RoR @ $25,000 | |||
|---|---|---|---|---|---|---|
| Drop-to-min | $100 cover | Drop-to-min | $100 cover | Drop-to-min | $100 cover | |
| Single deck0.25-deck cut · H17 | $1,939 | $1,961 | 23h | 24h | 0.46% | 0.52% |
| Double deck0.5-deck cut · H17 | $1,685 | $1,699 | 49h | 51h | 1.47% | 1.61% |
| Six deck1-deck cut · H17 | $1,359 | $1,364 | 199h | 203h | 7.34% | 7.61% |
| Eight deck1-deck cut · H17 | $1,257 | $1,261 | 279h | 285h | 9.25% | 9.53% |
RoR is the lifetime figure at the $25,000 bankroll and scales with it: on a larger bankroll the standard deviations lift the win rate enough to push it toward zero. At this bankroll the $100 cover nudges it up a few tenths of a percent, most visibly in the shoes.
A worked example: a CAC2 double deck
Put real dollars on it. Here is a spread from a real situation: a $25 to $400 ramp (a 1-to-16 spread) on a double-deck H17 game with late surrender, DAS and 62% penetration, played heads-up at 161 rounds an hour off a $45,000 bankroll, counting CAC2 with its own H17 deviations. It is a strong local game, and the player wants to keep it, so they hold a small bet on the first hand of each shoe rather than snapping back to the $25 minimum.
| First hand of the shoe | Win rate / hr | Cost / hr | Cost / round | N₀ | RoR |
|---|---|---|---|---|---|
| Drop to the $25 minimumno cover | $217 | n/a | n/a | 64h | 0.16% |
| Hold $50neutral-count bet averages $29 | $217 | −$0.570.26% of edge | −0.35¢ | 65h | 0.16% |
| Hold $100neutral-count bet averages $35 | $216 | −$1.590.73% of edge | −0.99¢ | 66h | 0.19% |
| Hold $200neutral-count bet averages $46 | $214 | −$3.291.51% of edge | −2.05¢ | 71h | 0.26% |
Native engine, 5 × 40M rounds per arm, common random numbers; RoR at the $45,000 bankroll. The first hand carries the prior shoe’s ending bet capped at the cover, so a $50 cover lifts the average neutral-count bet only part of the way up from the $25 minimum, while a $200 cover lets more of the big finishes carry.
Is the cover worth it?
Cover is never free, but this one stays cheap. In dollars it is at most a couple of dollars an hour, in double deck, where you spread the biggest. As a share of your win rate it bites hardest in the deep shoes, about one percent at a $200 cover, because their edge is so thin; single deck barely notices it. Either way the numbers are small, and the need for cover shrinks with the decks: a bet that moves at the start of an eight-deck shoe draws little attention.
The practical read is to keep the cover light and only where you draw heat. On the CAC2 double deck above, doubling the minimum to a $50 first hand costs about a quarter of a percent of the win rate, while holding a full $200 runs about one and a half percent. A light cover buys most of the camouflage for a fraction of the cost of holding a big bet out. You can price your own game (decks, cut, spread and cover level) in the bet-spread simulator: set your ramp, then set the start-of-shoe bet and watch the edge move.
Numbers from the AdvantagePlay native Monte Carlo engine (5 × 600M rounds per arm, common random numbers). Cover and drop-to-minimum play share the same shoes, so the cost is a within-shoe difference rather than the difference of two noisy win rates. Your results will vary with rules, spread and how deep the game is dealt.