The $10,000 Heads Up No-Limit Championship is generally a safe bet to see some top-notch play and interesting hands at the World Series of Poker, and this year was no exception. It was the young Spaniard Adrian Mateos who ultimately won it, but the hand which most impressed me was played much earlier in the tournament, by another player. The hand in question was #30 played last Thursday, on Day 2 of the tournament, between Ryan Riess and Dan Smith in the round of 16.
At this point, the 16 remaining players had all made the money, cashing for a minimum of $24,881; the gains of winning the match in question were to be an additional $22,028, and more importantly, making it through to the quarter-finals and keeping hope alive for the $324,470 top prize and the bracelet. Riess went on to finish fourth, after losing to eventual runner-up John Smith.
Riess was off to a strong start already by this point, and with the blinds at 2000/4000, enjoyed a lead of 520,000 to 280,000.
Riess and Smith made for an interesting matchup from the start because of their diametrically opposed styles of play. Smith is a very aggressive player and happy to play big pots, while Riess prefers a small ball, low-variance game and, faced with a player like Smith, is content to let his opponent do most of the betting in a lot of spots.
Naturally, the most important thing as a player in Riess’s situation in this dynamic is to correctly assess the strength of one’s own hand versus the opponent’s range and decide when the likelihood of winning a showdown is sufficient to justify the expense of getting there. This is tricky to do against a skilled aggressive player like Smith, as he’s going to be balancing his ranges pretty correctly, which makes it hard to know how many streets he’s planning on betting, and means that calls on later streets with marginal hands will tend to be close to zero EV (that is, will not be hugely better or worse than folding statistically speaking, but of course very important to Riess’s immediate results).
Riess was on the Button in this hand, which of course is also the small blind in heads-up play, making him first to act before the flop, but giving him position on Smith on later streets. As stated, Riess held a lead of 520k to 280k, or 130 BB to 70 BB.
Riess: AcQd (130 BB)
Smith: KhQh (70 BB)
From the start, it was a hand almost inevitably destined to be dramatic. Although Riess’s dominating Ace-Queen was a 70-30 favourite preflop, chances were that a flop connecting with one of them would also connect with the other somehow.
Riess raises to 10,500 (2.6 BB), Smith re-raises to 34,000 (8.5 BB), Riess calls.
Both players preflop moves are very standard, in terms both of the hands they’ve chosen to make these moves with and of bet sizing. King-Queen suited is a sufficiently monstrous hand heads up that Smith’s 3-bet is mandatory, and although Riess’s Ace-Queen is well ahead of Smith’s range, he doesn’t want to fold out a lot of the hands he beats by 4-betting.
Pot: 68,000 (17 BB)
If there was any doubt that the hand would prove a key turning point in the match, the flop quickly dispelled it. Both players flopped top pair and, although Riess had the better kicker, Smith’s flush and two pair outs meant that the players’ equities had become almost even. Furthermore, both players had, at this point, an unfoldable hand.
Smith bets 23,000, Riess calls.
There’s some question here of whether Riess should raise, and perhaps if the roles were reversed, Smith would consider it. Given the dynamic between the two of them, however, calling is almost certainly correct for Riess.
Smith’s range here will tend to contain various sorts of draws, sets, worse Queens and underpairs; given Riess’s tight image, a raise might induce Smith to fold the worse Queens, pairs and weaker draws, and go all-in with his sets and better draws, which would be the worst-case scenario for Riess. It’s an example of a spot where a tight image can work against you if you decide to raise, as you’d prefer your perceived range to contain more semi-bluffs.
Pot: 114,000 (28.5 BB)
Board: Qc8h5h + 6h
Inevitably, the Six of Hearts on the turn completed Smith’s flush, leaving Riess drawing dead. The only question remaining would be how many additional chips Riess would lose.
Smith bets 68,000. Riess folds! (After much thought.)
Here, Smith elects to increase his bet sizing to attempt to extract maximum value from Riess; if Riess calls, it creates a pot of 250,000, with Smith holding 155,000 back, just about perfect for a river all-in. This implication is not lost on Riess, of course, but Smith’s sizing is not in itself a giveaway, as it is also the sizing he would choose with most of his bluffs, in order to apply maximum pressure to the more marginal parts of Riess’s range.
The hand was aired “live” (i.e. on a delay, of course) on PokerGo, with commentary by Lon McEachern, Chad Power and Jeff Gross. During Riess’s long time in the tank, none of the three thought it very likely that he would be folding.
Although the idea was floated that Riess might be considering a shove, Power pointed out that the same problem existed as on the flop. In other words, if called, he’d usually be in bad shape, and that shoving would only make sense against a narrow range of hands, such as a naked Ace of Hearts with a King or Jack kicker.
The conclusion was therefore that a call was virtually inevitable, although Power did manage to hedge slightly just seconds before Riess pitched his cards away. Gross wondered aloud whether folding could possibly be crossing Riess’s mind, to which Power responded: “It’s crossing his mind, he’s just not doing it. [Pause]. Very often.”
How Tight is the Fold?
Depending on your level of experience with heads-up play, this fold may seem either somewhat easy or completely impossible to make. Clearly, since Riess managed it, it’s not impossible, but it’s definitely not easy by any stretch of the imagination.
In a full ring game, folding top-top on to a second barrel when a scare card falls on the turn is a tight play but not particularly unusual, but this is because the stronger preflop ranges players generally have in full ring play end up translating into stronger hands postflop as well. For instance, premium pocket pairs and suited Aces make up a larger part of a normal 3-betting range when there are more players at the table, so Smith would have overpairs and nut flushes more often on the turn. Heads up, both players’ ranges are considerably wider, so this is a hand that should rarely be folded under normal circumstances.
Against a strong opponent like Smith, this kind of hero fold isn’t made on the basis of one factor. A recreational player might have a big leak, like always slowing down on a flush board if he doesn’t have it, but against someone balanced, the only way this kind of move is possible is when there are several factors which, taken together, make folding advisable.
Level Two Thinking: What Does My Opponent Have?
Obviously, a player of Smith’s caliber isn’t going to have a bet-sizing tell which gives away whether he’s bluffing or going for value; whatever his bet size in a given situation, he’s going to have some hands with which he’s looking for a call, and other hands with which he’s bluffing. However, bet sizing does reveal something about how he’s constructing his range and what his plans are for future streets.
As stated earlier, the sizing on the turn is designed so as to set up a shove on the river, or at least imply that he could very well shove on the river. Conversely, the small sizing on the flop implies a range containing many hands which don’t mind seeing another card.
Taken together, these two choices of sizing make it pretty unlikely that Smith has a worse hand that he’s betting for protection or value. The turn bet is polarizing, in other words: Smith should either be bluffing or have a hand better than one pair most of the time with this combination of bets.
Level Three Thinking: What Does My Opponent Think I Have?
The fact that this is a 3-bet pot and Riess is known to play tight affects things considerably. A typical tight 3-bet calling range from the small blind in heads up play contains a lot of suited Aces (but not Ace-King), some big offsuit Aces like Ace-Queen and Ace-Jack, most suited broadway combinations, pairs up to Nines or Tens, and suited connectors down to about Seven-Six or Six-Five.
On this board texture, that means that Riess isn’t holding a lot of hands that connected only slightly with the flop. Most of his range will either have connected quite strongly, or missed. The fact that he proceeded to call the flop and hasn’t shown a tendency to float could cause Smith to take a lot of those hands which are worse than top pair out of his range. Riess might have just folded most of his non-Heart Ace High hands on this texture, not wanting to have to call multiple bets on a bad runout, and the turn would have improved many of the hands in the more speculative parts of his range.
In other words, if Smith is bluffing here, he knows that a lot of the time he’ll be trying to push Riess off a fairly decent hand; there aren’t a whole lot of easy folds in Riess’s range.
Level Four Thinking: What Does My Opponent Think I Put Him On?
The leveling war rabbit hole goes a long way down, but level four makes a good practical limit for these kinds of discussions. At this point we’re talking about table image: What our opponent thinks we’ve assumed about his play, and whether he’s going to try to switch things up on us. It’s a minor consideration in full ring play, but becomes very important heads up. It’s even more important in cases like this, with the players getting tipped off to each others’ earlier holdings by friends watching the live stream.
Smith had been bluffing a lot, and knew that Riess knew at this point that he’d been bluffing a lot. Both players therefore expect that Riess will attempt to adjust to Smith’s aggression at some point, and that Smith will be trying to guess at the timing of Riess’s adjustment, and lower his bluff frequency just as Riess raises his calling frequency.
Just before this hand, Smith had stepped away from the table to use his phone, as required by WSOP rules. This was presumably to get filled in on some of Riess’s hole cards in earlier hands, so Riess may have been thinking that Smith had his mind on the meta-game, and was contemplating a pre-emptive gear change.
So, if Smith knows that Riess is holding a lot of hands that he’ll be tempted to call with, and that Riess knows that Smith has been bluffing a lot and may be thinking about putting his foot down, that would make this a worse spot to bluff than a quick look at the board texture would lead you to believe. That, in turn, could lead Smith to be bluffing less, yet keeping his sizing large to make it look like his range is more polarized than it actually is.
These reasons, in combination, make for reasonable cause to consider a fold. But as Power commented, players will consider making an extremely tight fold a lot more often than they’ll actually do it.
However, it’s in these really borderline spots that physical reads become important, and as pointed out by the commentators, Smith accidentally gave a lot away with his behaviour, which Riess may have picked up on. Specifically, as Riess continued to tank, Smith appeared to become more relaxed, turning away from the table to take a drink of water, then settling back in his chair.
According to Zachary Elwood, who’s made a name for himself with his writings on poker tells, unusual behaviour early in a hand tends to indicate weakness, but this reverses itself later on as the pot becomes larger. Or, put the other way around, players are more careful about maintaining neutral behaviour when they’re nervous or excited, which is more likely to be with a strong hand preflop, but more likely with a bluff towards the end of the hand.
Here, Smith is making a bet which, although not all-in itself, is sized so as to make it likely that stacks will go in before showdown; relaxing as he sees that his opponent is on a difficult decision makes strong hands more likely.
Of course, good players are capable of faking physical tells at times. However, feigning relaxation at a time of high stress is considerably more difficult than other types of acting, and Riess may have noticed Smith’s body language and taken it to be genuine.
Perfect Storms as a Substitute for Randomization
Clearly, a perfect strategy would not have Riess folding in this situation very often, but it’s possible that hypothetical perfect play would have him folding some small percentage of the time. After all, it’s true that the ideal approach to most poker situations is to have a mixed strategy, that is to choose between two or more options in a probabilistic way.
This is how computers like Libratus play, but humans can’t be that random. However, the next best thing is to use these sort of external factors – tells and opponent psychology – as a way of choosing between options that should each find their way into one’s play at some point.
Here, folding top-top to a second barrel should in fact be a fairly low-percentage play. But since we can’t mentally roll dice to decide if this is the 5% (or whatever) of the time that we do so, what we do in practice is wait for this sort of perfect storm of meta-clues to induce us to do the unexpected: If you should ever be folding, even only rarely, then the time to do it is when your opponent probably thinks you’re strong and probably thinks it’s time for him to slow down, but makes a polarizing bet anyway, and then looks really relaxed about it.
Alex Weldon (@benefactumgames) is a freelance writer, game designer and semipro poker player from Montreal, Quebec, Canada.