Blackjack Tournament

[muffdiver]

I have no idea how many tournaments most of the posters here play in but I do know that both gronbog and myself participate in many (some of the same) and I'm sure that he concurs that the "luckbox factor" (as we call it) is HUGE and needless to say uncontrollable.
Beyond luck, we do have total control of bet sizes and strategic moves and it's imperative to put that control to it's optimum use. Most of the tourney experts will tell you that the most important things to monitor are your BR as compared to others at the table and the probable last hand betting positions. After those things come the possible familiarity with opponents tendencies and (in my case anyway) knowing the makeup of undealt cards.
Many times the BR comparison dictates that strategic bets are required well before the last hand so you must be very aware of that. It's a real bonus to get on a table with a bunch of what I call "dinkers and dunkers". These are the people that figure they're due to win the next hand even if it's the first hand. I'll never forget a tourney at the Frontier (Vegas) when I was dealt a deuce and a trey on the last hand and doubled it down ($100 live money) because it was my only way of advancing to the semifinal. Luckily for me, the dealer busted and the fellow that I beat out started ranting about how he lost out to a player that was dumb enough to double down such a hand . I was able to suppress the laugh but not the smile. Please put those players on my table every time!

[mhb]

I have no idea how many tournaments most of the posters here play in but I do know that both gronbog and myself participate in many (some of the same) and I'm sure that he concurs that the "luckbox factor" (as we call it) is HUGE and needless to say uncontrollable.
Beyond luck, we do have total control of bet sizes and strategic moves and it's imperative to put that control to it's optimum use. Most of the tourney experts will tell you that the most important things to monitor are your BR as compared to others at the table and the probable last hand betting positions. After those things come the possible familiarity with opponents tendencies and (in my case anyway) knowing the makeup of undealt cards.
Many times the BR comparison dictates that strategic bets are required well before the last hand so you must be very aware of that. It's a real bonus to get on a table with a bunch of what I call "dinkers and dunkers". These are the people that figure they're due to win the next hand even if it's the first hand. I'll never forget a tourney at the Frontier (Vegas) when I was dealt a deuce and a trey on the last hand and doubled it down ($300 live money) because it was my only way of advancing to the semifinal. Luckily for me, the dealer busted and the fellow that I beat out started ranting about how he lost out to a player that was dumb enough to double down such a hand . I was able to suppress the laugh but not the smile. Please put those players on my table every time!

you did a very advanced complicated deviation on the last hand.
did you consider the following factors:
other players' hands.
other players' bets.
probability of dealer busting.
probability of other players' busting or making hands or making winning or losing hands.
other players' possible deviations such as doubling spitting etc that could give you edge.
etc etc all in comparison if you were to play the hand according to count.
with all of these factors and possibilities and possible damage to your br, ev and ror how sure are you that you maximized your ev etc.

[muffdiver]

you did a very advanced complicated deviation on the last hand.
did you consider the following factors:
other players' hands.
other players' bets.
probability of dealer busting.
probability of other players' busting or making hands or making winning or losing hands.
other players' possible deviations such as doubling spitting etc that could give you edge.
etc etc all in comparison if you were to play the hand according to count.
with all of these factors and possibilities and possible damage to your br, ev and ror how sure are you that you maximized your ev etc.

It was an easy decision because it was the last hand of the session and it was the ONLY way that I could surpass the person I needed to for advancement in to the semifinal. Even though it was only for the semi, first place in this event paid 30k and the total purse was 50k which easily justified the decision.
The move put me in the semis but I wasn't able to get to the final table that time.

[mhb]

It was an easy decision because it was the last hand of the session and it was the ONLY way that I could surpass the person I needed to for advancement in to the semifinal. Even though it was only for the semi, first place in this event paid 30k and the total purse was 50k which easily justified the decision.
The move put me in the semis but I wasn't able to get to the final table that time.

I am glad to hear that it was an easy decision on last hand against one player.

for me those are the ones I want to do.

I played the finals table at frontier tourney and I had a tough decision. on the last hand of finals table. I was contender on last hand and bet max of 300. I got a blackjack and dealer had ace showing. count said no insurance. I had to decide.
I decided to go with count and not take insurance. I could not calculate all of probabilities under the pressure let alone in hindsight.
dealer had blackjack so I pushed and came in second winning 20000, first place was 50000.
after tourney an experienced player came over to me to tell me I would have taken first place if I had taken insurance/even money.
she may have been right, but how was I to calculate all of the probabilities:
dealer having bj
other players deviating doubling splitting etc.
there were many parameters.
many players contending on last hand.
one thing for sure, I maximized ev of my br by not taking insurance. I did bet high on last hand against the count.

[RWM]

Next week on GWAE we have Ken Smith as a guest. Back in theh UBT days many considered him the best tournament player around, and he has written 2 books on BJ tournaments. Check it out.

[muffdiver]

I am glad to hear that it was an easy decision on last hand against one player.

for me those are the ones I want to do.

I played the finals table at frontier tourney and I had a tough decision. on the last hand of finals table. I was contender on last hand and bet max of 300. I got a blackjack and dealer had ace showing. count said no insurance. I had to decide.
I decided to go with count and not take insurance. I could not calculate all of probabilities under the pressure let alone in hindsight.
dealer had blackjack so I pushed and came in second winning 20000, first place was 50000.
after tourney an experienced player came over to me to tell me I would have taken first place if I had taken insurance/even money.
she may have been right, but how was I to calculate all of the probabilities:
dealer having bj
other players deviating doubling splitting etc.
there were many parameters.
many players contending on last hand.
one thing for sure, I maximized ev of my br by not taking insurance. I did bet high on last hand against the count.

I don't think your memory is correct for the prizes. I played in most all of the Frontier events and 30k was typically what they advertised for 1st place and they had a disclaimer saying "based on X amount of entrants". I'm quite sure that I'm the only one that ever got the full 30k there for first because it was the only time that they met the "full field" goal. Second time I won it I got 26k for first. As long as I'm bragging, I also got 15k there for a second place and 4k for a fourth or fifth place finish. Not quite as good as it sounds because it spanned a few years.
My success rate at other venues isn't even close to what it was at the Frontier. Hated to see that place go down!
Also, RWM is absolutely right in saying that Ken Smith is one of the top tourney players. I've been matched up against him on a few occasions and did okay but it was only because of luck. I'm not in his league and not many others are! He's also a well respected and fine gentleman.

[Gronbog]

Sorry for the long post, but there have been many issues raised in this thread and we need to address them all ...

The arguments that I see here in favour of counting are of two varieties:

1. "If I bet according to the count, my chip stack should, more often than not, drift to a higher total than those of my non-counting opponents"
   
   As Flash has pointed out, the actual increased/decreased value of a bet made during a favourable/unfavourable count is measured in small single digit percentages of the amount bet. Say you're willing to bet 1/4 of your bankroll based on a high count at a tournament (which is more than a little bit risky, in my opinion), the actual value of that bet will likely be something like 4% at best making it worth something like 1/4 x 4/100 = 1/100 of your bankroll. Say then that you get extremely lucky and you manage to make 10 of these bets during a 30 hand round. The expected value of these bets, is now 10/100 or 10% of your bankroll, and that is if you get extremely lucky. More typically you will have the opportunity to make between 0 and 5 bets with an advantage each worth 1 or 2%, making your expected drift somewhere in the range of 0 to 2.5% of your bankroll. A similar analysis applies if the count stays negative and you expect to drift into a leading position by counting on your opponents to lose. If you happen to drift into more of on advantage than this, then, of course it's not because you were so cleverly counting the cards, it is because you got lucky.
   
   The point I want to make here is that even with an extreme max bet and an extreme amount of luck with respect to how much you'll be able to leverage it, you can at best expect to drift into a lead of something between 0 and 10% of your bankroll due to counting. While any lead is a starting point, I will tell you now that this lead is insignificant in a blackjack tournament. The first really significant lead in a blackjack tournament is 1/2 max bet minus the min bet plus a chip (It's a good exercise, if you're an inexperienced tournament player to figure out why, but I will explain it if necessary). I can tell you that there is no tournament in which 10% of your bankroll reaches this threshold, so this lead you're hoping to drift into is of almost no significance.
2. "Sometimes blackjack pays a premium and, if I time my a big bet to correspond to the count, I increase my odds of hitting one with a big bet out"
   
   This strategy should also be evaluated based on its expected value. Once again, using a fairly risky 1/4 of your bankroll as your big bet, look at the difference between the rate of winning blackjacks at different counts. For a neutral count, you get a blackjack about 4.8% of the time but only 95.2% of these are winning blackjacks so you can expect a winning blackjack about 4.6% of the time. Similarly at a high true true count of +5 you can expect a blackjack about 6% of time time and a winning one about 5.6% of the time, an increase of 1%. So even if you were fortunate enough to place 10 big bets during the round, the expected gain from 2/1 blackjacks would be 1/4 x .01 x 10 or 2.5% of your bankroll. Once again, an insignificant amount in any blackjack tournament. If you actually hit even one blackjack, you would, of course do much better, but at only a 1% increase in the rate, the odds are not significantly higher that you will hit one even at a high count of +5 and, if you do, it's because you got lucky, not because you were counting.

The next most common argument in favour of counting has been that you should bring all of your weapons to the battle and be prepared to use them. The question then becomes one of the relative effectiveness of each weapon and in which situations each would become the weapon of choice (if any). Now, in terms of bet sizing, the strategic weapons available to a tournament blackjack player are all based on win/lose/push/swing percentages. The key thing to realize is that these percentages change very little with the count; less than 1% even at extreme true counts. The percentages are:

Percentage of hands won: approximately 43%
Percentage of hands lost:: approximately 48%
Percentage of hands pushed: approximately 9%
Full Swing (i.e. one player wins while the other loses): approximately 12%
Negative Half Swing (i.e. one player loses while the other pushes): approximately 5%
Both Push: approximately 1%
Positive Half Swing: (i.e. one player wins while the other pushes): approximately 2.5%
Success Rate winning two bets using a double/split anything strategy: approximately 30% (are you surprised?)

Now let's look at a few common situations and what the weapon of choice would be for maximizing their values:

• You have a lead. The value of your lead depends on how big it is and how many hands remain in the round. Let's look at the possibility that your opponent will take the lead from you on the next hand.
  
  • Betting first with a lead of less than a minimum bet.
    
    • The strategic bet is to bet minimum. Your opponent should respond by betting so that he takes the lead by winning his hand. You successfully defend your lead if he loses or you win or push while he pushes. You have a 51.5% chance of defending your lead.
    • If you bet with the count, then you will bet more than minimum in a good count and your opponent can respond by betting so that he takes the lead if you lose or he wins while you push. Your chances of defending you lead are now only 49.5%.
    
    
  • Betting first with a lead of less than or equal to 1/2 max
    
    • The Strategic bet is to bet your lead less a chip or less. If you can do this, the your opponent must bet to win in order to take the lead. If he loses or pushes, you still have the lead. You have a 57% chance of defending your lead.
    • If you bet with a favourable count and bet more than the strategic bet above, you will find your self in the position as with the previous lead. Your opponent can bet for the low leaving you with a reduced 51.5% of 49.5% chance of defending your lead.
    
    
  • Betting first with a lead of 1/2 max +
    
    • Now you're strategically in in the driver's seat. You can bet 1/2 max and defend your lead successfully even if you opponent wins a max bet. Your opponent must win while you push or lose or win two bets by doubling or splitting in order to take your lead. The latter is his best option and he has about a 30% chance of succeeding and only if he realizes that this is what he must do. Your chance of defending your lead are 70% or greater.
    • Once again with a favourable count you could you end up over betting, leaving you with only a 51.5% or 49% chance of defending your lead. You could also, in a negative count not bet enough allowing your opponent to take the lead by simply winning his hand. Your chances of defending your lead then are only about 57%.
    
    
  • Betting first with a lead of max +
    
    • At this point you're really in control. You can make a max bet and defend your lead with a win even if your opponent wins a double. He needs you to lose while he wins. You have about an 88% chance of defending your lead.
    • In an unfavourable count you might bet too small to cover your opponent's double lowering you chances of defending your lead back into the 70% range.
    
    
  • Betting last with any lead
    
    • When betting last with any lead, you can defend your lead with a probability ranging from 81% to 100% simply by matching your opponent's bet, however big it is or by locking him out, when possible if he bets small.
    • If you bet with the count you open yourself up to the same over betting and under betting possibilities as above which lower your chances into the 49% to 57% range, depending on the situation.
    
    
  
  
• You are trailing and within range.
  Let's look at your chances for taking the lead.
  
  • Betting First
    
    • When trailing and betting first, you're in a tough spot because, as we've seen above, your opponent can defend his lead with a high probability of success by matching your bet if you go big or by locking you out if you go small. This is why position is so important.
    • The count is meaningless here. Your opponent will do one of the above no matter what bet you choose.
    
    
  • Betting last
    
    • The strategic bet is to bet the opposite of your opponent. If he goes big, you go small. If he goes small you go big. Of course you hope he goes big regardless of the count. This is because he has about a 48% chance of losing his hand in any count and that's your best chance take the lead. If he goes small, you go big and you have a 43% chance of taking the lead regardless of the count.
    • If you bet with the count, then you may make a similar bet to that of your opponent and we've seen how disastrous that can be.
    
    
  
  
• Timing
  The value of a lead also depends on how many hands remain in the round. Clearly the fewer hands that remain, the less chance your opponents have to catch you. The ultimate is the case of taking the lead on the final hand. In this case, it's game over. Above I discussed the percentages for defending or acquiring a lead on a single given hand. In almost every situation, there is some chance that you will retain your lead on a given hand and some chance that you will lose it. The probability that we will defend our lead over more than one hand is (the probability of defending one hand) ^ (number of hands needed to defend it) -- where ^ is "to the power of". The chances of defending your lead decrease exponentially with the number of hands needed to defend it. This is why it is so important to strike late in the round in addition to striking from good position. The exponential nature of this relationship and the fixed nature of the percentages involved make taking the lead orders of magnitude more valuable near the end of the round, and at any count, than for taking it at some random point in the round during a good count with an arbitrary number of hands remaining to defend it.
  
  Similarly, if you are are behind and not in range, then you are going to need to win more than one big bet to get back into it. The count is not a good predictor of getting ahead by more then one bet from your current chip total. If you do manage to get two bets ahead in a short period of time, it will not be because of a good count, it will be cause you got lucky. The more hands you leave yourself to get lucky, the better. If you wait for a favourable count to make your contrary bet(s), you're simply wasting time. Of course once you're within range again, you would revert back to waiting for the proper time and position to strike.
• Winning Multiple Bets
  
  • Premium hands
    Some mention has been made of using the count to optimize the chance of capitalizing on a hand which is paid at a premium, such as a 2/1 blackjack. We went through the percentages for a winning blackjack above found found that even at a count of +5, we only get a winning blackjack about 5.6% of the time. Compare this to the chances of winning two bets by doubling/splitting anything. The success rate for that was around 30%. That's almost 5 times more likely than depending on getting a blackjack, in any count. If I needed to win two big bets in a hurry, I would certainly choose the latter strategy.
  • Two in a row
    Another strategy to be aware of, regardless of the count is that you are more likely to win two big bets by doubling just about anything than by winning two hands in a row. So if the are only two or three hands remaining and you can only take the lead by winning two big bets, you're better off doubling/splitting the first one than trying to win two in a row. This result is in Ken Smith's latest book and holds true for just about any two card hand, including a few that will surprise you).
  
  

These are only a few of the most basic principles of tournament blackjack but they are some of the most effective. In every case above, I can't see a reason for prefer betting with the count over betting according to the situation. Of course we can't prove anything using a few examples. We can only use counter examples to disprove things, but as the situations get more complex I suspect that the trend would be even stronger in favour of strategic decisions over ones based on the count.

[muffdiver]

Very nice job gronbog. You put some work in that! I can't disagree with any of it.
I'll say that in most tourneys the strategic plays (for me anyway) aren't until the last few hands. There are exceptions to that. If two or three other players start running away from you early on, it may be wise to up your bets but if it's only one person that's leaving you behind then it could be smarter to minimize the risk and hope his luck turns as it often does. Later in the sessions, you don't have that luxury, of course.
I know that after many years of playing in tournaments, it's only been a small percentage of times that I've felt the need for what I consider real strategic bets and plays before the last 5 hands.

[Three]

If two or three other players start running away from you early on, it may be wise to up your bets but if it's only one person that's leaving you behind then it could be smarter to minimize the risk and hope his luck turns as it often does

I usually wait for someone to lower their bets unless time is getting short. As long as they keep betting big their luck will probably turn. If they are smart enough to force you to try and catch them by betting small then thats what you have to do. Position is the key here because if he is savvy enough to bet small he will likely bet as Gron suggested to protect his lead.

[mhb]

I don't think your memory is correct for the prizes. I played in most all of the Frontier events and 30k was typically what they advertised for 1st place and they had a disclaimer saying "based on X amount of entrants". I'm quite sure that I'm the only one that ever got the full 30k there for first because it was the only time that they met the "full field" goal. Second time I won it I got 26k for first. As long as I'm bragging, I also got 15k there for a second place and 4k for a fourth or fifth place finish. Not quite as good as it sounds because it spanned a few years.
My success rate at other venues isn't even close to what it was at the Frontier. Hated to see that place go down!
Also, RWM is absolutely right in saying that Ken Smith is one of the top tourney players. I've been matched up against him on a few occasions and did okay but it was only because of luck. I'm not in his league and not many others are! He's also a well respected and fine gentleman.

for the sake of the argument I am sure of the prizes. it was a long time ago, I suppose before your time. I played against Anthony Curtis and what was known as the group of 10i believe it was called. yes, the frontier had great tourneys in those days with all kinds of freebies.
eventually the casinos would not let us locals play, they wanted the tourists to win.
I played mostly a straight counting game and it really baffled Anthony Curtis and the group of 10 because so often I was seen at the final table. I was asked why I do this. I answered because I don't want to pick up any bad habits, they shrugged their shoulders. yes, maybe I was just lucky, but that is what happened.
I think the scenario could be simulated: a counter 95% of the time vs. deviation tourney play.

[Gronbog]

While re-reading this thread, I thought of a reasonable reason to count during tournament rounds. The purpose would be to detect other counters in order to predict their bets when you must bet before them. Could be useful during the middle to late hands of the round. You would have to have a very good read on someone, however, in order to assume that they would stick to it during the end-game.

[Gronbog]

I played mostly a straight counting game and it really baffled Anthony Curtis and the group of 10 because so often I was seen at the final table. I was asked why I do this. I answered because I don't want to pick up any bad habits, they shrugged their shoulders. yes, maybe I was just lucky, but that is what happened.

There have been several posts in this thread with a similar theme. That being, "I have played <a certain way> and have had <some degree of success>". This kind of anecdotal evidence does not really carry much weight in terms of establishing the effectiveness of a particular strategy. As with straight up AP play, one must collect data on many millions (preferably billions) of samples in order to approach a result which is statistically significant. Even if someone was to play a tournament every day for their entire life they would not collect enough data to establish the effectiveness of their particular strategy.

I think the scenario could be simulated: a counter 95% of the time vs. deviation tourney play.

This is, most likely, the only way to collect a significant amount of data. Fortunately, I can shed some light on this. The end-game strategies are all pretty well documented and, for the most part, agreed upon. I began to wonder about effective strategies for the early and middle hands of blackjack tournament rounds, and I did exactly as you suggest. I created a simulator that can simulate subsets of actual tournament conditions that I have encountered, ranging from single hands to entire tournaments. Each tournament is populated by bots which employ different early/middle round strategies that I have either observed in real play or have come up with myself. All of the bots have the same level of skill when it comes to finishing. The bots play against one another and statistics are collected about their performance. The purpose is to see which tactics at which levels of aggression are most effective in setting one up for the end-game under particular tournament conditions.

The reason that this is relevant here is that some of the strategies lend themselves to variations in bet size and timing and for those I have included variants which consider the count. So there might be a bot which employs "strategy X" and another which uses "strategy X with counting". In every case, for every situation I have studied, the bot that uses counting either has the same performance as the variant which does not or performs worse. The reason is that these bots sometimes mistime their moves (in a strategic sense) because of the count. They either move too early or too late, or they make big move when out of position, so that their opponents can easily counter them.

Just some more fuel for the fire ....

[mhb]

There have been several posts in this thread with a similar theme. That being, "I have played <a certain way> and have had <some degree of success>". This kind of anecdotal evidence does not really carry much weight in terms of establishing the effectiveness of a particular strategy. As with straight up AP play, one must collect data on many millions (preferably billions) of samples in order to approach a result which is statistically significant. Even if someone was to play a tournament every day for their entire life they would not collect enough data to establish the effectiveness of their particular strategy.

This is, most likely, the only way to collect a significant amount of data. Fortunately, I can shed some light on this. The end-game strategies are all pretty well documented and, for the most part, agreed upon. I began to wonder about effective strategies for the early and middle hands of blackjack tournament rounds, and I did exactly as you suggest. I created a simulator that can simulate subsets of actual tournament conditions that I have encountered, ranging from single hands to entire tournaments. Each tournament is populated by bots which employ different early/middle round strategies that I have either observed in real play or have come up with myself. All of the bots have the same level of skill when it comes to finishing. The bots play against one another and statistics are collected about their performance. The purpose is to see which tactics at which levels of aggression are most effective in setting one up for the end-game under particular tournament conditions.

The reason that this is relevant here is that some of the strategies lend themselves to variations in bet size and timing and for those I have included variants which consider the count. So there might be a bot which employs "strategy X" and another which uses "strategy X with counting". In every case, for every situation I have studied, the bot that uses counting either has the same performance as the variant which does not or performs worse. The reason is that these bots sometimes mistime their moves (in a strategic sense) because of the count. They either move too early or too late, or they make big move when out of position, so that their opponents can easily counter them.

Just some more fuel for the fire ....

is it then fair to say that card counting is preferable during early and mid game?
it is the end game where deviations can be used when they are very simple and well documented.
in the end game when it is complex with multiple players doing large deviations and it is very difficult to
understand what is really going on that one is such as myself confused then resort to card counting
and not join the battle.
yes, if you can understand clearly the complex and see a deviation that increases your ev then go ahead and join the battle.
it sounds like the posters to this thread don't have concern of ror and br management. they have expendable funds to lose it all
at tourney -- the fees and buy ins I mean.
for me if I deviate with plays with large bets and strange plays then I am risking my br. a steep losing streak at tourneys could harm me.
fees and buy ins could cost thousands in tourneys.
a losing streak in tourneys could cost tens of thousands with live action money buy ins.