The Satellite Battle
Monday, 17 March 2014
Satellites nowadays are packed with value, but you’re a fool if you treat them like regular tournaments, says Dara O’Kearney.
Cold War
A non poker friend of mine who knows a little but not a lot about poker (but nevertheless takes a keen friendly interest in my latest career) recently asked me why I seemed to be a much better player in satellites than "actual tournaments". It seemed pointless to even bring up ICM, so to put it in layman terms, I said that if normal tournaments were war, satellites were more like the Cold War. Late on in a satellite, you needed to display to your opponents that you were willing to engage, while doing everything possible to avoid actual confrontations. Merely understanding that gave you an advantage over opponents who had never heard of MAD (mutually assured destruction).
When I was starting as a full time poker pro just over five years ago, I got quite a lot of advice from many quarters, some good, most bad. Ironically, the single most useful piece of advice came from a non pro, who told me that the softest online tournaments were satellites to live events.
Nowadays I still try to satellite into all the bigger live tournaments I play, not just because it means getting in cheap, but also because satellite tournaments tend to be great value in themselves, with a lot of essentially live players who don’t play much online. I was UKIPT Online Qualifier of the Year last year (which means I won more online satellite to the UKIPT than anyone else) and over the years I have satellited in online to several WSOP main events, EPTs, and every Irish Open I have played. With Stars leading the way by introducing tourney dollars for multiple seat winners and other sites gradually realising the advantages of incentivizing grinders to grind satellites, more and more are realizing the benefits of adding satellites to their grind.
In this piece I will look at my default strategy early in satellites, and the ICM considerations that characterize them late on.
Battle Tactics
My default strategy early on in satellites with a good structure is to play a fairly tight straightforward game. There are a few reasons for this:
1. With no antes and low blinds, there’s no real need to play a lot of pots.
2. The best strategy is generally to do the opposite of what the other players are doing. If everyone is playing loose, you can sit tight and wait for good hands and spots to chip up with little or no risk (by contrast, when everyone is playing tight, it’s more profitable to try to steal a lot of pots). As a general rule, recreational players play looser at the start (I guess they feel with so many chips relative to the blinds, they can “afford” to).
3. By playing tight early on, any other players paying attention at the table will tag me as a tight player, and later on when it becomes more important to steal pots with non premium hands, I stand a better chance of being given credit for a hand than if I have a loose image.
4. At this point in my online session, I will be peak multi-tabling, probably playing 12 tables at the same time. If I play too many hands on every table, I will quickly get overwhelmed having to make lots of quick decisions, so it’s best to stick to playing solidly. This means not only having fewer decisions to think about, but also simpler decisions and less tricky spots.
5. At the start of the tournament, I have no real reads or idea on how most of the players at my table play. By sitting back for a while, I allow them to play enough hands for the HUD I use to build up statistics of how loose they are and how they play post flop, so when I do start playing more hands, I will have hard information to base my decisions on.
Know Your Enemy
The biggest thing that distinguishes satellites (with multiple prizes of equal value) from normal tournaments (with top heavy payouts) is that ICM is a much bigger factor. There's an entire generation of MTT players on Stars with its small min cashes and top heavy payouts that never got round to learning ICM the way those of us who cut our teeth in sit n goes had to. In the normal games they play they can get away with this, as ICM is not that big a factor most of the time. However, in satellites they routinely make loose calls near the bubble that might be okay in a normal MTT, but in a satellite are the equivalent of lighting a cigar with a wad of cash in a restaurant in front of the smoke inspector.
To take a simple example, if there are ten left in a satellite that awards 9 equal prizes, and stacks are equal, then you need 90% equity against the shoving range to call (since if you fold, you have a 90% chance of a prize). This leads to the often talked about hypothetical situation where aces becomes a fold, since even against a range of any two cards, aces only has 85% equity. This in turn means that if everyone is playing optimally, UTG will shove every hand (since nobody can call), and everyone else will fold without looking at their cards. This would continue until such a time as the blinds rise to the point where the small blind and the big blind are forced all in.
Real world situations differ from this simple hypothetical example in two major ways. First, stacks will never actually be equal (meaning the actual equity required to call will vary for each shoving and calling stack), and second, not everyone will be playing optimally. At some point someone will look down at aces and call a shove. Or maybe even kings, queens, jacks, or eights, not realizing it's a mistake. This makes the maths and game theory much trickier. I recently analysed an exact bubble hand for a group of students. With 9 left, and lying 8th in chips, one of the students had shoved 53s from the cutoff. He explained that while he understood this would be a losing play in a normal MTT far from the bubble if people were calling correctly or close to correctly, in this satellite situation, he felt that the only player likely to call was the big blind (who was 9/9), and that even he would not call without a hand at least as strong as eights. Another factor was he felt that if he folded, there was a strong chance this player would get a walk, as nobody who already had a ticket locked up would want to risk shoving in to the one guy who most needed to gamble.
All good points, so using a software tool I calculated the so called Nash equilibrium (the correct shoving and calling ranges for each player if everyone was playing optimally). If ICM had not been a factor (say, far from the bubble in a normal MTT), then the Nash equilibrium solution for our hero was to shove approximately 25% of hands if folded to, a range that does not include 53s. So as suspected, in this scenario, shoving 53s would be a mistake. But obviously ICM was a huge factor, so I recalculated the Nash equilibrium to account for it. Now it became correct to shove just over 40% of hands, a range which still doesn't include 53s. However, the Nash equilibrium solution required the big blind to be willing to call off with a hand as weak as sixes, something my student was convinced he would never actually do. So I recalculated again, this time assuming the big blind would not call correctly, but only with eights or better. With this adjustment, it now becomes correct for my student to shove over 90% of hands, a range that does include 53s (and even weaker hands). So if his assumptions about calling ranges of the players behind were correct, he made a correct shove. One of the other students then pointed out that all these examples that took ICM into account assumed that all other players behind would fold aces. While they would be correct to do so, in reality it's unlikely they actually would. They might even call with weaker hands down to jacks. To illustrate how big a factor this is (having players behind who call much looser than they should), I recalculated the solution, this time assuming all the players behind would call with jacks or better. It now became profitable for my student to shove only 10% of hands!
Adapt Your Attack
I think this example of how the same stacks and blinds in different situations (ICM a factor or not, players behind calling correct ranges or not), the shoving range can go from 25% (when ICM is not a factor) to over 40% (when it is) to over 90% (when it is, and players behind call tighter than they should) all the way down to 10% (if they call too loose), shows how important it is to be aware of these factors in satellites. A top class MTT player with a good awareness of shoving ranges but not ICM sticking close to cEv Nash equilibrium ranges (25% in this case) will be making a major mistake, either shoving way too wide or way too tight depending on the specific conditions. This indicates how someone who has a good awareness of ICM can have a massive edge even against other winning regs, making satellites particularly profitable to grind. Or, as I said to my non poker friend, in the Cuban Missile Crisis that is the latter stages of a satellite, it's all about avoiding confrontations without appearing to fear them.