Discount Gambling

Spanish 21


Spanish 21 is a variation of blackjack, played with a “spanish” deck that contains no Tens, but offers many liberal rules and options that generally yields better odds than blackjack.  Basic strategy is a little more complicated, but almost all casinos allow the use of strategy cards, so you’re a lot better off playing Spanish 21 if your house offers it.  Some of the liberal rules of the game are:

  • double on any number of cards
  • re-doubling (sometimes up to 3 doubles)
  • double after split
  • surrender after double
  • all player 21’s pay immediately, regardless of dealer hand
  • bonus payouts for 6-7-8, 5 card 21’s, etc.

Basic Strategy

The basic strategy for a 6-deck shoe game, where the dealer hits on soft-17, and up to 3 doubles are allowed is presented in the strategy card below.  The card is color-coded, where red stands for hits, white stands for surrender, yellow means stand, and blue means double.  The exceptions to the color-coded rules are indicated by special notations in the box.  There are no exceptions to hit or surrender, or any decisions after doubling.  There are a few exceptions before doubling, where yellow stand boxes indicate when you should hit instead.  Also, there are blue double boxes that indicate when you should just hit the hand.

Basic strategy for Spanish21.

Basic strategy for Spanish21.

Stand Exceptions

The yellow boxes with a number or * ‘ ” symbol in them mean you should hit the hand when you’re drawing for a bonus payout.  For example, say my hand is 7-8 offsuit, and the dealer is showing a 2. I have a hard total of 15. The strategy table shows a yellow box with a 4* in it.  The * tells me I should take a hit, because I’m drawing to a 6-7-8 offsuit bonus hand (pays 3:2).

As another example, say I have 2 3 2 3 2 2, a six-card hard 14, and the dealer is showing a 6. I’m drawing to a 7-card 21, which pays a 3:1 bonus.  The strategy table has a yellow box with a 6″ in it, meaning I should take a hit and draw to either the 7-card 21 bonus, or the 6-7-8 suited bonus.  I have the former, so I take a hit.

Double Exceptions

Sometimes, it’s better to just hit a hand than double it, depending on how many cards you have.  This is because you have a chance of drawing to a multicard 21 bonus (remember, there’s no bonus after double, and it gets very expensive to see another card after doubling).  For example, if I have a 3-card 11 total (e.g., 4-5-2) and the dealer is showing a face card, the table shows a blue double box with a 3 in it, meaning I should just hit my hand, because I’m holding at least 3 cards.  This makes sense, because if my first draw is a 3, then my total is 14, and I’d have to surrender if I doubled.  But if I just hit to the 4-card 14, then I have an easy hit to the 5-card 21 bonus (I need a 7).  Even if I don’t hit my bonus, I have a “free” draw to improve my hand, instead of surrendering the hand.

Some Observations

Double and Re-Doubling

All double-down opportunities are on-average winning hands (+EV).  That is, a decision to double-down is never the “least of two evils” (i.e., lose less by doubling).  You double because it’s the way to maximize the win from the hand.  However, some re-doubles are losers, and you’re redoubling-down to lose less than surrendering the hand 😦

For example, one of the weakest doubles are player hard-9 against a dealer 3.  Doubling still yields a winning net of approx 7.8% of the initial bet, whereas the hitting yields only 5.5%.  So, passing up on this opportunity to double will cost you about 2.1% of your initial bet.  That’s a huge edge you can’t afford pass up.  As another example, consider soft-18 vs. a dealer 4.  Doubling in this case yields a 14.3% return on your initial bet, while hitting returns only 8.4%.  That’s a whopping 5.9% advantage you cannot forfeit.

So while I often see people passing up on their double-down opportunities, I also see people doubling-down when they’re not supposed to.  A lot of people want to double soft totals against a dealer 2 or 3.  Of course, you only do so against a dealer 4/5/6, and doing otherwise will cost you money.  For example, doubling down on an A-3 (soft-14) vs. a dealer 2 will turn a winning hand into a losing hand.  Just hitting in this case returns an average 5.4% profit on the initial bet.  Doubling down erroneously here will result in an average loss of -0.2% of the initial bet.  Yet again, a doubling mistake costs 5.6% of your initial bet.  This is throwing money away.

Some of the re-doubling opportunities are the “least of two evils”, where you’re just trying to lose less on average.  For example, say you doubled-down on a 4-6 (10 total) against a dealer 8, and you drew a 2.  Now, you have 12, and are faced with either surrendering (-100% of initial bet), redoubling (-98.5%), or standing (-101.5%).  The basic strategy says to re-double, because you’ll get a little return on your redouble (1.5% of your initial bet).  The worst case is to stand, where you’ll lose a little more than if you surrender.  But practically speaking, since you’ve already doubled your bet, this +/- 1.5% initial bet difference between your options is reduced to a +/- 0.75% difference relative to your doubled bet.  So it’s not a huge mistake to go any way you want to here.

Dealer 4/5/6 Upcard

I get  happy when I see a dealer 4, 5, or 6 upcard.  I really get happy when I have a low soft total, because this is a chance to double and redouble your bet a few times.  In all cases, opportunities to double increase your average win from the hand.  In general, with a dealer 4/5/6 upcard, take all the opportunities to double and split, as they all return more expected win (i.e., they’re not “least of two evils” moves).

For example, you split 3-3 against a dealer 3, because you lose less (-7.5% of the initial bet) than if you took a hit (-11.1%).  But if you have 3-3 against a dealer 4, taking a hit is still a loser (-7.2%), while splitting becomes a +1.5% average winner.  Of course, you have to bet more (and may have to double it), but you can’t afford in the long run to cost yourself this 8.7% mistake.

All the doubles and redoubles with soft totals, even against a dealer 4 are winners.  For example, if you have an A-3 (soft 14) against a dealer 4, doubling will return a 15.2% profit on your initial bet, while hitting will yield just 9.1%.  This 6.1% difference includes the possibility of redoubles down the line.  Even A-7 (soft 18) against a dealer 4 yields more for doubling (14.3%) vs. just hitting (8.4%).  As I said before, all redoubles against a dealer 4/5/6 increase your winnings vs standing.  For example, if you (re)double to a soft 18 against a dealer 4, you’ll eek out a little more profit (28.7% of initial bet) by redoubling than by standing (23.7%).  Of course you’re putting a lot more at risk by doing this, especially if this is your third double.  I tend to start with a small initial bet, so I don’t care if I have to 8x it.  But, risking an additional 4 initial bets to extract out a gain of 5% of the initial bet may not be worth it to you.  You’re really only getting 5%/4 = 1.25% return on your 4x redouble bet.

The examples given against a dealer 4 upcard are even more pronounced against a 5 or 6 upcard.  The bottom line is: a dealer 4/5/6 upcard is good stuff.  I especially like the 6 upcard, and doubling down with low soft totals.

Breaking The Rules

Almost all of the people I play with don’t believe in math, and they display nothing but contempt for it.  I really learned to keep my mouth shut on the matter.  You can get into a fight over this, believe me. At best, I’ll say indifferently, “well, the book says …”.  They don’t believe in math, but they believe in the existence of the book.   The only math the players practice at the table is “would-of” calculations to see who to blame  for taking/not-taking a dealer bust card, or low card that gave the dealer a good total.  It’s really complicated stuff.  They see a lot of patterns in the cards, and hold a lot of superstitions.  Like if the dealer has a weak upcard, they still believe someone at the table has to hit (I guess take a low card), otherwise the dealer won’t bust.  People go ballistic over this stuff, especially at Spanish 21, where you’re supposed to hit hands like 14 against a dealer 3, and 12 against a dealer 6.  All of this just results in more blame opportunities than regular blackjack.

Common Mistakes

Some of the plays in the basic strategy table are overlooked by a lot of players, and these mistakes are significant, however infrequent.

Hand Correct Mistake Cost (of initial bet)
9 vs. dealer 2 hit double 1.8%
17 vs. dealer A hit stand 1.8%
8-8 vs. dealer face card hit split 4.1%

Gambling More

Sometimes, you just want to gamble more, despite the odds.  The best thing to do in this case is to choose the better gambling opportunities.  Of course these are often double-down or split opportunities, where you can get more money into play.  Some of these optional plays are listed in the table below, along with their cost, in terms of the initial bet.  The higher the cost, the worse the mistake is.  Note that there’s a wide range in the cost of mistakes that you’ll see at a table.

Hand Correct Mistake Cost (of initial bet)
A-3 vs. dealer 3 hit double 0.04%
3-3 vs. dealer 6 hit split 1.1%
3-card 11 vs. dealer face card hit double 2.3%
A-8 vs. dealer 6 stand double 7.8%
doubled 12 vs. dealer face card surrender stand 7.9%
12 vs. dealer 6 hit double 13.2%
doubled 12 vs. dealer 6 stand re-double 13.8%
doubled 17 vs. dealer A surrender stand 13.9%
doubled 12 vs. dealer 9 surrender re-double 18.3%
A-9 vs. dealer 6 stand double 24.4%
doubled 12 vs. dealer face card surrender re-double 25.2%
doubled 13 vs. dealer face card surrender re-double 42.7%

For example, you can decide to gamble and double an A-3 vs. a dealer 3.  The end result of the hand, assuming you finish the hand correctly, is only worst by 0.04% of the initial bet, had you just hit, then finished out the hand correctly.  That’s a very small cost for gambling it up, and getting some more money out there.  It hurts a little in the long run (which is too long for any of us to actually see), but in the short run, it’s just increasing your bet.

Often I see people re-double 12 against a dealer 6.  I guess they figure if you’re supposed to hit all 12’s, might as well double on them too.  In reality, the hand is a loser, and you’ll only get 62.6% of your amount bet back by standing.  Re-doubling will lower the return to 48.4% (of the amount bet before the last redouble).   It’s a bad bet, but some people like to gamble.  They can get lucky, by not busting.  At that point, they’ve doubled their expected return on the hand.

Gambling Less

Sometimes, you don’t care what the odds say, you just don’t have the appetite to re-double or split a hand that you don’t really like.  Your fears are generally well-founded.  In these cases, you have to put more money down, sometimes just as the “least of two evils”, to improve the return of a losing hand.  Below are some of the plays that fall into this category, along with the cost of  the mistake, in terms of the initial bet.  The realistic cost of the mistake probably needs to be divided by 2 in the case of a split, and by the amount of the (re)double bet in case of a double.

Hand Correct Mistake Cost (of initial bet)
doubled 12 vs. dealer 8 re-double rescue 1.7%
doubled soft 18 vs. dealer 4 re-double stand 5.0%

More Articles

posts categorized under Spanish 21:

78 Responses

Subscribe to comments with RSS.

  1. Ted said, on September 14, 2009 at 7:31 pm

    Hey, just curious, where did you get your figures for the % return for individual plays? I’ve been looking for this information extensively across the internet and wonder if you have a full table of “indexes.”

    • stephenhow said, on September 14, 2009 at 7:36 pm

      I wrote my own program to calculate each option EVs for a given hand. I was really into Spanish 21 for a while, until I barred myself from the only game in town 😦

      • Ted said, on September 15, 2009 at 4:55 pm

        Can you e-mail me the program, or can I pay you for it?

      • stephenhow said, on September 15, 2009 at 5:15 pm

        It’s C++. Do you know how to compile it?

      • Sharon said, on May 5, 2010 at 12:26 am

        You barred yourself or were you barred by the casino? If you barred yourself, then you can simply unbar yourself. If the casino barred you, then you have a problem. I am going to presume that the casino barred you. That leads me to my question:

        What did you do to get yourself barred from the only game in town?

        and my follow up question:

        Now that you, Katrina Walker and others have revealed the vulnerability of Spanish 21 to attack by advantage players such as yourself, has casino management begun employing any countermeasures?

        Specifically I am wondering about the bet spread. If an advantage player such as yourself can not make a spread of at least 1 to 30 for minimum and maximum bets because doing so sets off alarm bells in the pit, then such an advantage player is going to find it more difficult to attain the kind of edge that can be attained in many of the traditional blackjack games.

        Of course, if spanish 21 is the only viable target in town, then wouldn’t really matter to an advantage player such as yourself.

      • stephenhow said, on May 5, 2010 at 7:09 am

        I barred myself because of one particular (ex-) floorperson at Sycuan named Trish. I was just betting $5 every hand, not varying my bets. This woman was just humorless, on some minor power trip. Instead of customer service, she gave me a hassle over buying people’s surrenders at the table. Anyway, months after my public self-barring event, I heard that she got demoted from floor supervisor to dealer.

        Anyways, there are other +EV games in town (see Mississippi Stud). You’ll get an average +1.5% edge on every hand, without having to count any shoe. BTW, I’m not an advantage player. I expect to lose in the long run. I’m just trying to waste the most amount of time for the least amount of money 🙂

    • Wilson Fang said, on April 21, 2013 at 11:13 pm

      I spent 1 year plus to compute the house edge and test my idea in Pontoon, 3 times double-down, late surrender allowed, double-down at any number, double down allow after split, 3 splits allowed for all cards except A, 1 split for A… can anyone tell me when is the optimal time to double?? what is the house edge??

  2. Ted said, on September 15, 2009 at 9:09 pm

    I can download a compiler.

    • stephenhow said, on September 15, 2009 at 10:32 pm

      Ok, will upload the source code after work. It runs as a simple command-line tool, so you can do >spanish21 -h to get the usage.

  3. stephenhow said, on September 15, 2009 at 11:52 pm

    Ok, the source code is uploaded. There’s a link for it under “More Articles” above. When it’s compiled, try:

    >./spanish21 -u 6 6 7
    dealerHand: 6
    playerHand: 6 7
    EV -0.166567

    stand : EV -0.198398
    hit : EV -0.166567
    double : EV -0.417523
    split : EV -1.000000
    surrender: EV -0.500000
    rescue : EV -1.000000

    Note that EVs for N/A options may be -1; ignore them.

    >./spanish21 -h
    usage: spanish_21 [-da] -u upcard p0 p1 [p2 ...]
    -d: doubled hand
    -a: use dealer approximation (practical)
    -u: dealer upcard (required)
    p0 p1 [p2 ...]: player hand (must specify at least 2 cards)

  4. ZenMaster_Flash said, on October 31, 2009 at 5:01 pm

    The E.O.R. for the 8 deck Stand 17 variant of Spanish 21 that is dealt in New Jersey, Ontario and a few midwestern stores is quite different from the published figures for the more commonplace 6 deck H17 version of the game, to wit:

    > A: -0.16%
    > T: -0.06%
    > 9: -0.02%
    > 8: -0.03%
    > 7: 0.01%
    > 6: 0.07%
    > 5: 0.12%
    > 4: 0.1%
    > 3: 0.07%
    > 2: 0.05%

    Note the accelerated value of the ACE and adjust your Card Counting efforts to reflect that.

  5. ZenMaster_Flash said, on October 31, 2009 at 5:03 pm

    No I cannot compile, debug, or otherwise crunch code; although I suppose that it could be run as a UNIX executable via Terminal

  6. EmeraldCityBJ said, on November 9, 2009 at 10:55 pm

    When reviewing your basic strategy table, I observed that your chart indicates that you should HIT rather than DOUBLE a soft-14 (Ace-Three) with a dealer 3 showing. This strategy is inconsistent with what Michael Shackleford has published at, and with the strategy published “The Pro’s Guide to Spanish 21” by Katarina Walker.

    I have also written a program to calculate EVs for Spanish 21 decisions, and my numbers are consistent with your chart. I agree that hitting A3v3 is preferable to doubling. My thought is that both Shackleford and Walker used an infinite deck for their calculations and that when you take into account the removal of the ace and both threes from the 6-deck shoe, it’s barely enough to make a difference in the strategy decision. I asked both of them a question about this, and both of them responded that their strategy is correct. I’d be interested in hearing your thoughts on this.

    • stephenhow said, on November 9, 2009 at 11:45 pm

      Yeah, I think we’re probably right here. But, think there’s always some small (and different) assumption we all use (e.g., I don’t carry suit information in my recursive EV calculator, and I make some approximation for the 678s bonus). So, it’s not surprising we have one small difference on the margin. This decision point is probably very close between doubling and hitting, and the difference is probably due to differences in our small assumptions. Of course, if you made no assumptions, then I trust your results 🙂

  7. Ted said, on November 16, 2009 at 9:43 am

    Hi Stephen–

    Could you please tell me the expected value for hitting versus standing on hard 17 versus an ace up in Spanish 21. I know the correct play is hit, but I’m wondering how much I would lose by standing and avoiding having all the players (and the dealer!) jump down my throat at each opportunity.


    • Ted said, on November 16, 2009 at 9:44 am

      By the way, I play a S17 game, don’t know if that makes a difference.

      • stephenhow said, on November 16, 2009 at 12:08 pm

        In a S17 game (dealer stands on soft 17), if you have a 3-card hard-17 total (e.g., 6,7,4), you should stand. Your stand EV is -0.522146 and your hit EV is -0.525567. If you have a 4-card hard-17 total (e.g., 6,7,2,2) you should hit. HitEV = -0.483529, StandEV = -0.523382.

  8. ZenMaster_Flash said, on November 16, 2009 at 4:32 pm

    TED plays a S17 game.

    You are making errors in Basic Strategy.

    You surrender a (2 card) 17 vs. a dealer’s Ace, and hit a 3 (or more) card 17 vs. the Ace.

    Do you need indices for this play ?

    • Ted said, on November 16, 2009 at 4:36 pm


      I don’t count.

      I know you surrender a two-card 17 against the ace and do that with no problem. My question (I guess I should clarify) was about three- and four-card 17s.

      Thank you Stephen for the information. Sounds like a wash on three-card 17s, but a definite hit on four-card 17s versus the dealer ace.


  9. ZenMaster_Flash said, on November 16, 2009 at 4:50 pm

    StephenHow’s equity figures are probably correct, but I count cards.
    I do not use the unbalanced Hi-Lo count offered by K. Walker.
    But what that count offers for 17 vs. Ace is as follows.
    Remember that Kat’s count is unbalanced.

    Late Surrender your 2 card 17 vs. Ace = −6 (slightly negative)

    Hit 3 card 17s vs. Ace = −3 (barely negative)

    Rescue (forfeit) your [doubled] 17 vs. Ace = +2 (very positive)

    p.s. Basic Strategy for [a 6 card] 17 vs. a dealer’s 8,9,10 is to HIT

    • EmeraldCityBJ said, on November 17, 2009 at 9:07 am


      I would not characterize a true count of -3 using Kat’s system as “barely negative”. It would be more accurate to state that it’s barely above the neutral true count of -4. Basic strategy is to hit, but if only a few small cards have been removed from the pack, standing becomes the better play.

      • Ted said, on November 17, 2009 at 2:36 pm

        As a side note: Does anyone actually count (and win) at this game? (Obviously you guys use some form of counting or another). I bought Walker’s book but the system seemed amazingly hard. I’m just trying to learn basic strategy at this point.

  10. ZenMaster_Flash said, on November 17, 2009 at 5:08 pm


    Yes, and other than myself I know several people making “good money” at this game.
    As a Professional Player I have depended almost entirely on this one game for my living
    this year.

    The S17 game is VERY much better than the H17 game, with an acceptable Hous Advantage of .33%.

    The Basic Strategy will, naturally, NOT give you an advantage, any more than it can in BJ.

    Level Two and Level 3 counts make the game very profitable; but all Card Counting, even with Katarina’s simple unbalanced Hi-Lo Count, come with a wrinkle. That is that the Play of your hands becomes VERY complex when compared to BJ. e.g. You have a 17 and the dealer has an Ace. You have an index to direct you when to Late Surrender. You have a different index for Hitting a 3 card 17. You have a 3rd index for “rescuing” (surrendering) a multi-card 17 after doubling. To make matters more complex the most common hands have indices that are subject to further adjustment, depending on how many cards you hold. e.g. You have 11 vs. the dealer’s 10. You have learned that Basic Strategy is to double BUT it is a completely borderline play. With a 3 card 11 facing any dealer high card the correct Basic Strategy play is to HIT instead of splitting – BUT – if the True Count is sufficiently high it overrides the decision to HIT. A borderline True Count dictates the basic Strategy play. Most of these exceptions are for hands of 4 or more cards. The important ones require just 3 card hands. The extremely rare play of hitting a HARD 18, 19 or 20 vs. a dealer Face card !

    These examples are very incomplete – but are hints for just how devoted to study and practice one need be to gain an advantage at this game. Many months of heavy daily study and practice are required. It will vary by the individual, but most people would do well to devote 4 hrs. per day for three to five months. BUT, very few people are willing and able to devote that many hours to intensive study / practice. Also, note that bankroll requirements are extremely high. If your local casino offers $15 or $10 as minimums, you’ll need a wide bet spread; indicating that to keep your “Risk of Ruin” to a reasonably acceptable level will require a cash bankroll of perhaps $20,000+ at minimum. Unless you are willing to place tens of thousands of dollars at risk, this is NOT FOR YOU.

    • EmeraldCityBJ said, on November 17, 2009 at 11:28 pm

      I would argue that there are not “several” people making “good money” playing Spanish 21. Sure, there probably a handful of folks who’ve bought Kat’s book and tried out the system, but I seriously doubt there are more than 10-20 people world wide who can demonstrate a significant long term profit from the game. When considering the millions of casino patrons, or even the hundreds of people trying to count cards at blackjack, the number of Spanish 21 APs is exceptionally small.

      I would also argue that the game is not “very profitable” just because someone decides to use a Level 2 or Level 3 system. An advanced system is only marginally more profitable than a simple system. I personally use a Level 2 system because I learned how to play the game on my own before Kat’s book was out. There’s no reason for me to switch to Kat’s system since I’ve already made the effort to learn something which does perform a little better. However, if I had it to do over again, I would probably choose to just learn Kat’s system.

      I do agree that Spanish is much more complicated and much higher variance compared to blackjack. The strategy deviations for the bonus payouts, plus the added indices you need to learn does take more time to memorize. Also, learning more indices in Spanish increases your win rate more than learning a comparable number of indices in blackjack. To mitigate risk, my recommendation is to size Spanish 21 bets to be roughly 2/3 to 3/4 your blackjack bet with the same edge. For example, if you’re betting $100 with a 1% edge on blackjack, I’d recommend betting around $75 with a 1% edge on Spanish. Unless your bankroll is such that you are frequently betting up to the table max, you’re not going to make as much on Spanish with the same bankroll and same RoR.

  11. ZenMaster_Flash said, on November 17, 2009 at 5:14 pm

    In my post of a moment ago, I found something that needs clarification.

    The final sentence of the penultimate paragraph should read as follows:

    The extremely rare play of hitting a SIX card HARD 18, 19 or 20 vs. a dealer Face card
    Is actually a Basic Strategy play.

    I do not have an index for this rare play, but it is not a play that can be altered via the True Count.

  12. ZenMaster_Flash said, on November 18, 2009 at 5:53 am

    I respect EmeraldCity Blackjack, someone that I have corresponded with in the past and know from online BJ forums.

    However I am not going to split hairs over the word “several”. Unless I have actually seen the person in action I am too cynical to accept their statements at face value. I have seen only a “few” people that I know extracting regular profits from this game.

    I do take exception with Emerald City BJ’s post re: Level 2 and 3 counts.
    The boost in advantage is very significant when moving up to a level 2 or 3 count.

    Spanish 21 differs from BJ in several ways, but none are more crucial than the power of Basic Strategy Deviations. Knowing a solid range of indices is analogous to the strategic playing of single deck BJ. Spanish21 is dealt as a 6 or 8 deck game. In a shoe game perfect bet-sizing trumps play deviations. In a hand-held BJ game, bet sizing alone will only work with a very good rule set. Play deviations are crucial. This is why such “pitch” BJ games ought to be attacked with a strong level 2 count, such as Hi-Opt Ii, AO II or ZEN; as these counts have the highest Playing Efficiency, though the first 2 listed require side-counted Aces.

    • EmeraldCityBJ said, on November 18, 2009 at 10:38 am

      My sims suggest that my Level-2 system only increases my win rate about 5% over Kat’s system. In other words, if Kat’s system gets you $30 per hour, I could win $32 per hour using my system with a comparable spread. If this extra $2 is important to you, then go for it. I would argue that many folks will lose the $2 difference and then some by making more errors with a more complicated system.

  13. ZenMaster_Flash said, on November 18, 2009 at 10:51 am

    What are the “tags” for your count ?

    • EmeraldCityBJ said, on November 18, 2009 at 11:16 am

      Aces = -2
      J,Q,K = -1
      2,3,4,5,6 = +1

      Starting count is 0, and the neutral true count is 0.

      • Wilson Fang said, on April 21, 2013 at 11:20 pm

        Yes, totally agree. this is the best count for Pontoon….
        Ace = -2,
        J,Q,K = -1,
        2,3,4,5,6 = +1

  14. Ted said, on April 10, 2010 at 1:31 pm

    Stephen: Here is a challenging question for you. I was playing Spanish 21 the other night and the woman next to me was down to her last $10. She was dealt a 10 against a 2 and she let me double on her hand. She got a 2 on the 10. At that point, should I have told her to rescue and taken $5 back? My advantage had been lost when the 2 was dealt, so I’m thinking we should have rescued if she had agreed to it.

    For what it’s worth, we lost the hand.

    • ZenMaster_Flash said, on April 10, 2010 at 5:57 pm

      !0 vs. 2 is a proper double.
      It is NOT a proper “rescue“.

      You simply never do better by rescuing against a dealer 7 or less.

      You will lose less than 50% of your money bet by standing.
      Look at it this way.
      If you are in a situation where you can win (here by the dealer’s bust)
      just 25% of the time then you do NOT surrender.
      25% means 1 time in 4 you win, and 3 times in 4 you lose.
      So … In 4 attempts you win 1 and lose 3 times for a net loss of 2 units.
      If you surrendered all 4 times you would be out 4 times ½ for a net loss of 2 units.

      • Ted said, on April 11, 2010 at 7:25 am

        ZM_F — Thanks for the reply. I know that you never rescue against a 2. However, I believe my situation was different. We all know that 12 against a 2 is a net loser by about 75%. The reason you stand on that double is because the probabilities are you have a 25% chance of winning twice as much money, a 100% chance of losing one of those bet, or a 75% chance of losing both bets, as you said.

        In this case, I had only half of the action on the hand. I am no longer getting the return for the money that I expected when I offered to double on her hand. My bet (and hers as well, I think) is a net loser. Shouldn’t I surrender, since I can’t hit?

        Stephen, do you have E.V. values for this play?

  15. ZenMaster_Flash said, on April 10, 2010 at 6:05 pm

    I am confused here by the Basic Strategy chart in the INTRODUCTION above [for the respilt H17 game.]

    Since when is splitting 3’s against a 6 a “mistake”
    Also doubling a soft hand against a dealer Face looks really weird to me.
    Can that too be correct?

  16. Zen_Master_Flash said, on May 5, 2010 at 7:09 am

    Sharon, et al,

    I have been backed-off 3 times at Spanish21, (and narrowly escaped on a 4th recent occasion.

    Influential casino consultant, Bill Zender, has succeeded in teaching the casino industry that the game is subject to the predations of rabid Card Counters.

    If you are unfamiliar with “Z Man”, see:

    Katarina Walker’s fine book is outdated ONLY in that respect.

  17. John C said, on September 11, 2010 at 12:13 pm

    Hi Stephen, the link for your Spanish 21 source code doesn’t work. Can you fix that? I think it’s amazing that you’ve done all this analysis just for fun. You must be a helluva programmer.

    • stephenhow said, on September 12, 2010 at 7:01 am

      You can get my Spanish21 source here. One day I need to clean up all my blackjack code, re-write it in Java, and make it open-source at Sourceforge.

      • Richard said, on March 15, 2015 at 3:47 pm

        stephenhow said, on October 25, 2012 at 12:00 pm

        Yeah, that’s a dead link, plus it was old, sloppy code. Unfortunately, I don’t give out my new, good code anymore. It can pretty much analyze any blackjack variant quickly and accurately, so it’s too good to give away now 🙂

        Based on your last post on this thread, SO MUCH FOR OPEN SOURCE and making it available at Sourceforge!

  18. Chris Burke said, on August 31, 2011 at 8:48 am

    As a recreational player, could an A-5 (modified, I would assume) or simpler counting method be used to even out the odds, if not swing them in favor of the player?

    It appears most commentors on this site are pretty sophisticated in using counting methods, so I apologize for the rookie question.


  19. zenmasterflashcatdata1 said, on August 31, 2011 at 11:09 am

    No. Something as simple as an Ace-Five count wouldn’t be capable of altering the House Advantage sufficiently to make the game beatable.

    I user a complex level 2 count with > 100 indices.

  20. Chris Burke said, on September 2, 2011 at 7:36 am

    Well, as a recreational player, I am not necessarily interested in “beatable” so much as “break-even”, or closer to break even.

    I know the more complicated counts are required to adjust for the complex nature of some of the payouts and the lack of 10s — I was just curious if an A-5 would be of any benefit (or some modified version of that)?

    I go with a couple thousand and play $15 hands for hours and hours, and generally break even or lose maybe $100 or $200, or make about the same, over a weekend. Over 10 weekends this year, I am prob up $600 cash, plus my room is comp’d at a nice hotel in AC. I pay for meals, but I have a pretty successful day job so I am not sweating it either way.

    My good friend and I just sit, hang out, drink, talk, and play, so that is why I go, not necessarily to win, so I was thinking something more simple might help my odds someone, or reduce their odds.

    I play in AC – S17, 8 deck shoe — wizardofodds says perfect play puts me at -.40% — if I could get close to break even with a simple method, it would be more fun, bankroll last longer, whatever.

  21. zenmasterflashcatdata1 said, on September 2, 2011 at 8:34 am

    The Ace-Five count was never intended for a shoe game.

    However, that is not to say that it is worthless. It has some slight value, but nowhere remotely
    close enough to create a break-even game, even with an extremely aggressive bet-spread.

    The (unbalanced) Hi-Lo Count offered by Katarina Walker in her text is a simple Level One Count.

    Even If played with a very conservatively narrow spread (i.e. 8 to 1) and only the important indices

    the game is an undeniably profitable play.

  22. Chris Burke said, on September 2, 2011 at 8:53 am


    So you think that picking up her book and applying the Hi-Lo count she describes reasonably easily and make the game beatable?

    It might be worth it. Again, more of a mental exercise for me, but if you think it is worth it, I would give it a try.

  23. Steven George said, on October 16, 2011 at 11:16 pm

    The Spanish 21 table where I play offers an optional Match-the-Dealer side bet for both the up-card and the down-card. The minimum hand bet is $3 ($2 Tuesday). The maximum MD is $100 each. I use a simple level 1 balanced count but only a 2-to-1 hand bet spread. What I do instead is: try to hit the suited-MD that pays 9-to-1 for $900 for instance on the King of Hearts. Without a simulation program, I estimate that when my count is greater than +9, I have an MD advantage. That is when I place $100 on each of the MD circles. One time I hit a “double-suited, double-non” that paid $2,600. Have you ever tried to simulate such an idea?

  24. zenmasterflashcatdata1 said, on October 17, 2011 at 6:25 am

    The Match the dealer bet is NOT beatable. You were just lucky.
    The “count” its NOT (in any way) relevant.
    The house advantage remains 3.0% at all counts !

    • Steven George said, on October 17, 2011 at 7:39 am

      I am quite certain that the MD percent does change! For instance, If the first half of the shoe came out as ALL cards less than 7 (including the A), then the remaining half shoe would contain only 6 different denominations of 7 thru K. This would dramatically improve the odds of a MD as compared to a reshuffle with 12 different denominations. The world seems to be very close-minded about a MD player advantage. In straight BJ, the suited-MD pays 11-to-1 because with 13 denominations it is more difficult to get than in Spanish with only 12. I will continue to bring home the cash with my MD strategy! Thank You

      • EmeraldCityBJ said, on October 17, 2011 at 12:44 pm

        It is possible to win money counting for the match bet, but it’s not a very practical strategy. Since you don’t provide the tags to your system or your initial running count, it is impossible to know if you are playing a winning game. With the counting systems I have analyzed, I can say that the best win rate I could come up with is $30-40 per hour if you flat bet $100 only when you have an edge. The problem is the bet is pretty high variance and you’ll need a sizable bankroll to absorb the swings. I would argue that if you have a bankroll sufficient to play the match, there are far better ways to utilize it. Instead of the $30-40 per hour betting the match, you could play other games and make hundreds an hour with less risk.

  25. zenmasterflashcatdata1 said, on October 17, 2011 at 11:39 am

    Your comment about “the first half of the shoe” suggests that you have not worked this through mathematically.

    For such an extreme deviation from normal to occur would happen perhaps once in your entire lifetime – if you played nonstop from birth through death and expired at the age of 124.

    • Steven George said, on October 18, 2011 at 10:24 am

      Your comment about playing non-stop for 124 years is irrelevant to the problem solution. The extremely rare (non existent) chance of 144 consecutive small cards being dealt from the top of a reshuffle is only intended to illustrate the upper limit of a possible player advantage. It only demonstrates that EV does change on the MD side bet.
      The two other, more positive replies, also deserve my response.
      First, without having BOTH the up-card AND down-card match, it is true that the variance it too great to justify a single max $100 MD on the “up-card only” circle. The loss of playing basic on the hand offsets all MD efforts. However, with the additional opportunity to match the down card at the SAME time, the variance is reduced to an acceptable level. I have found that in general only 1 shoe in 6 will offer more than a single opportunity to bet the $100 MD on both circles. And that with the “up-card only” available the bankroll must be about $9,000. With both up-and-down matches present, $2,300 is nearly always enough to avoid busting out Note that where I play it is permissible to wager max on MD and min on hand. Some places don’t allow the MD to be greater than hand (this kills my strategy).
      Second, I said “balanced level 1”. This means start at zero and end at zero. It means +1 for 6 denominations and -1 for the other 6 denominations. It makes NO DIFFERENCE which 6 you choose for +1 or -1. You could randomly choose 2,3,7,9,J,K for +1 if you don’t want to match them. Simply pick the 6 that makes it easiest for you to not make counting mistakes. I use 2-thru-7 for +1 (paint,9,8,A are minus). For simplicity, I only go by running count (no true count math). This means that I always wait until the fifth round to decide if max MD is a go having at least a +10 running count. The casino where I play always deals more than 4 decks between shuffles, so I still get enough opportunities at greater than +9 running count. Also, if the running count drops from, for example, +15 to +9 late in the shoe, I will stay with BOTH $100 MD bets one more round. The opportunity exist to “jump-in” with two hands and a total $400 on MD. This causes extreme hostility at the table when you do get a suited match. Other players think that you have just “stolen” their good hand and that you have “screwed up” a good shoe. So, I don’t ever play more than just one hand.

      • EmeraldCityBJ said, on October 18, 2011 at 12:46 pm

        The $30-40 per hour win rate I posted assumed that both the up and down match were available. The systems I simmed are similar, but probably not quite as good as the one you suggest. Regardless, I still think you’re making less than $50 per hour and that your risk of ruin with a $9k bankroll is through the roof. With a $100k bankroll, I’d be hesitant to bet much more than quarters on the match bet. I may be more risk-averse, and generally play with a RoR of 1% or less.

  26. zenmasterflash said, on October 18, 2011 at 5:48 am

    EmeraldCityBJ knows of what he speaks; however using some specialized count to track the likelihood of a profitable “match the dealer” bet would have the player utilizing just basic strategy in order to play his hands while flat-betting. Overall, the losses from that strategy will never be recouped by the rare opportunity to bet M.T.D. This ought to be plainly obvious, although a two-person mini-team approach would overcome that fault.

  27. zenmasterflash said, on October 18, 2011 at 10:48 am

    The statement:

    “I use 2-thru-7 for +1 (paint,9,8,A are minus)”

    suggests that the poster imagines that he can ignore the E.O.R. of the card ranks.

    • EmeraldCityBJ said, on October 18, 2011 at 12:49 pm

      EOR for this bet is pretty meaningless. Each single card, regardless of rank and suit will produce the same EOR. To see an EOR difference, you have to remove various combinations of two or more cards.

    • Steven George said, on October 18, 2011 at 5:15 pm

      EXACTLY correct! EOR applies only to sizing and playing the HAND bet strategy. It has NO impact on the relative frequency of one set of 6 denominations vs its opposite set of 6. And it is this measure that makes the MD advantage possible.

  28. zenmasterflash said, on October 18, 2011 at 1:30 pm

    Note: The official rules for Span21, as promulgated by Masque Publishing, prevent the player from betting more on the M.T.D bet than on the regular bet. Also forbidden is placing such a bet on another player’s hand or playing more than two hands yourself.

    Of course, where the game is offered by a casino that is a tribal enterprise rules may or may not be observed.

    • Steven George said, on October 18, 2011 at 5:24 pm

      Yes, where I play they do enforce the restriction of not playing the MD on other players’ spots. The play of multiple hands is allowed and ENCOURAGED! I have seen them actually open a new Spanish table solely for the purpose to allow a single whale to play 7 spots “heads-up”. I know of only two locations that allow the MD to be greater than the regular bet. One of these places has recently lowered their max MD to $50.

    • White_Gypsy said, on August 10, 2016 at 3:20 pm

      Ha! I play the non-tribals and play more than two hands most of the time.

  29. Jimmer Flimmer said, on May 16, 2012 at 5:56 am

    Match the Dealer is a terrible bet. I was just curious what the actual odds are because a lot of people I see plying take this bet just because they feel like a baller when the suited match comes out, and I like to make fun of them especially if they ask me why I’m not playing my MD. The way I figure it is you have any two cards (let’s assume they are always two different cards. If they are not it’s even worse for you). Then there are only 7 perfect match cards left in the entire shoe (assuming none already came out which I think again if it were any different, it would be worse for the player). Using 8 decks (which is what my local place does) 8*48 minus the 2 that you have is 382. So 382 to 7 odds of a perfect match or about 55 to 1. You have two cards however and so two chances to match, so your overall chance of winning is about 27 to 1. But I’m going to tell people 30 to 1 to make it seem like a big difference from the measly 12 to 1 theyre getting.

    I just want to point out that when I play Spanish 21 I get absolutely no heat at all. I am only 21 though and dont play all the time, but I feel that I would have to be winning for a looooong time before anyone became suspicious.

  30. ZenMaster_Flash said, on May 18, 2012 at 4:12 am

    The MD side bet has a large house edge of 2.99%
    learn to keep your trap shut and do NOT tell others how to play.
    just tell them that you are unlucky and that you obey your hunches.

  31. Kelvin Morin said, on October 18, 2012 at 9:45 pm

    Hi Mr. How. I was hoping to get the source code for Spanish 21 as suggested above, but it does not appear to be available anymore. Is there any way I can obtain this code from you?

    • stephenhow said, on October 25, 2012 at 12:00 pm

      Yeah, that’s a dead link, plus it was old, sloppy code. Unfortunately, I don’t give out my new, good code anymore. It can pretty much analyze any blackjack variant quickly and accurately, so it’s too good to give away now 🙂

  32. Marc Gray said, on March 16, 2015 at 8:43 am

    Some Span21 S/W is so badly needed !

    • Wight_Gypsy said, on May 18, 2015 at 8:47 am

      Very true.

  33. ZenMaster_Flash said, on May 20, 2015 at 9:14 am

    There is a small coterie of A.P.’s who plunder this game.
    WE have our own software and use our own counts.
    We do not share because it is the only form of BJ where
    the heat is low and, when played correctly, is very much
    a fountainhead of income for players on the East Coast
    and in the MidWest.
    As playable BJ becomes ever more scarce and hi-tech
    surveillance becomes ever more paranoid, the greater
    the pressure we receive to hold our secrets very very
    “close to the vest.”

  34. Wight_Gypsy said, on June 2, 2015 at 12:57 pm

    Hello Stephen-

    Could you explain the bs discrepancies for 88vT and 33v6 compared to that presented by the Wizard and K. Walker? Seems more than peculiar.

  35. ZenMaster_Flash said, on June 5, 2015 at 9:11 am

    “Discrepencies” is a vague term.

    Be specific and I will enlighten you.

    • Wight_Gypsy said, on June 5, 2015 at 6:45 pm

      Two BS differences noted. The Wizard and Walker’s BS for 88vT is split but his advice is to hit. The second is the strategy conflict for 33v6. Hit as opposed to split. This is what I want to have explained or is it a typo?

      • Wight_Gypsy said, on June 8, 2015 at 9:15 am

        At this point, I am forced to presume my question is moot.

  36. Catdata1 said, on June 9, 2015 at 3:24 pm

    Both are fully correct pair splits !

    • Wight_Gypsy said, on July 2, 2015 at 12:08 pm

      Ok but that is NOT consistent with the chart entries for the H17 RD game..

  37. Catdata1 said, on June 9, 2015 at 3:55 pm

  38. ZenMaster_Flash said, on July 2, 2015 at 12:22 pm

    The H17 RD game is a rarity and I was not considering same.

    • Wight_Gypsy said, on July 3, 2015 at 2:43 pm

      Glad you made that clarification.

  39. ZenMaster_Flash said, on August 14, 2016 at 10:48 pm

    What casinos would they be ?

  40. ZenMaster_Flash said, on August 16, 2016 at 7:35 am

    Sorry, but I was responding to a post about playing Spanish21 somewhere, where it is alleged
    that the player can play as many hands as s/he sees fit to play. I was asking for casino names !

  41. Frank in South Florida said, on December 13, 2016 at 9:17 am

    What does the exclamation (!) point next to the P (split) in your SP21 chart for Player 7 against Dealer 7 represent/mean???

  42. Frank in South Florida said, on December 14, 2016 at 8:53 am

    Perhaps H 7-7 against Dealer’s 7, if your 7’s are suited??? Trying to answer my own question, subject to your confirmation or contrary explanation. Thank you for your consideration.

  43. Wight_Gypsy said, on January 14, 2020 at 10:15 pm

    Have you analyzed the edge on a 5 deck H17 Spanish 21 game where the House pays 10-1 on the insurance bet versus a suited natural? I’ve found this game in Ontario.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: