PokerStove

This Poker Stove article discusses intermediate-level concepts that requires knowledge about equity and how hand ranges work. The last part is about finding out how often your opponent has a certain hand.

Introduction

PokerStove (www.pokerstove.com) is a program for calculating hot-and-cold equity, that is to say, it calculates your exact chance of winning a certain hand at showdown. For instance, you can plug in 72o vs. AA and find out that 72o only has an 11.8% chance to win if all the cards are dealt and neither of the two players fold. It's a program that you download and run directly on your computer, as opposed to online odds calculators. The upside to running a program locally on your computer is that it's faster, really.

So what do you use it for? Well, Poker Stove is a very useful tool for analyzing hands and situations away from the tables. For instance, if UTG raises in a 6-max limit ring game, and I call from the BB with JsTs, how am I doing on a Jh-7s-7d flop if my opponent is-

a) squeaky tight (raises 3% of hands UTG)
b) average (raises 10% of hands UTG)
c) loose/aggressive (raises 20% of hands UTG), or
d) maniac (raises at least 50% of his hands UTG)

I can plug these ranges and my exact hand and this exact flop into PokerStove and find out that my chance of winning vs. each respective opponent is,

A: 41%
B: 61%
C: 67%
D: 73%

Is this useful? Yes, very! Knowing your hot-and-cold equity is a great first step in being able to figure out the best course of action. Whether you should call or raise the flop can be debated, but we've at the very least established that you're doing well enough that you shouldn't fold, at least not on the flop.

The Basic Controls

Now open up PokerStove. If you haven't already, download it from here: http://www.pokerstove.com/download/

We're going to perform some simple experiments to get us started.

Selecting a range for an opponent

The first thing we want to do is pit a few preflop hands versus some ranges. To do that, we're going to first see how "ranges"- or "Hand Distributions" - work in PokerStove.

There are a few different ways of setting your opponent's range, but the easiest one is to just put in a percentage*. So, for example, if you know from PokerTracker that your opponent raises preflop with 10% of his hands in this position, you put that 10% in. As soon as you shift focus (e.g., press the Tab-button), PokerStove will convert that to a range. Try it!

Pokerstove

Putting in 10% PokerStove has now converted 10% to a range. 10%, in other words, is synonymous to putting in "77+,A9s+,KTs+,QTs+,AJo+,KQo" in the same field. "77+" means any pocket pair 77 and higher. "KTs+" means any suited king, with a ten or better kicker. Try putting in some different percentages (don't forget to append the '%') and see what ranges they correspond to.

Of course, we can put in a string like that ourselves as well if we want. Try altering the string that PokerStove produces when you put in 10%, and see what works. For instance, you will notice that if you make a string that PokerStove doesn't know how to read (like putting in "Fredrik") the background is going to turn maroon, signifying that there's a parsing error in the input - that it's for one reason or another incorrect.

poker stove

Setting the board

Now we move to the next control, the Board-input. Here, we can tell PokerStove what the board looks like. Since it's somewhat pointless to speak of "ranges" on the board, we put in what the board actually looks like. Sometimes, we can ignore what suits are on the table, but it's usually best to put in as exact of a board as we can. For instance "AhQs8d." Furthermore, you don't have to put in all the cards on the board. You can put in just the flop, or even just one or two cards. PokerStove will generate random scenarios for any and all cards that you don't explicitly put in.

www.pokerstove.com

Setting the board to Ace of hearts, Queen of diamonds and 8 of diamonds

As with the Hand Distribution input fields, the Board background will turn maroon if we put in a string that PokerStove can't understand. You may also click on the "Select" button next to the Board input field to click on the specific cards. This way, you don't have to worry about typos or other things getting in your way.

Dead cards

I've never used this feature for anything. In theory, you can use this if someone who folded tells you their hand or if someone accidently flashes their cards. In practise - and especially online - this is not a very useful feature. You may use it to find out just how bad of a beat you just got (if you after-the-fact find out that one or more of your opponent's outs were folded by someone else) but for proper analysis and learning, it's really not necessary.

"Enumerate All" vs. "Monte Carlo"

What PokerStove does is to run simulations. It doesn't calculate, it simulates. So when you run a simulation, it's going to pit the hands and ranges that you put in, on the board that you put in (if any), randomize all the unknown variables many times, and tell you how often on average the different players win. There are two ways it can do this:

"Enumerate all" goes through every possible combination, in some kind of order. For some scenarios, this is very fast since there aren't so many possible combinations. Most cases where you only have two players involved, for instance, doesn't take Poker Stove many fractions of a second to calculate. But when you have three or more players involved in a pot, the number of possible cases has grown exponentially, and it may take a long time for the program to run every single combination of possibilities. That's when using "Monte Carlo" comes in handy: It randomizes the simulations.

Basically, this means that instead of following a pattern and grinding its way through every possible holding, it's going to just randomly run simulation after simulation. Because computers are so fast these days, we're going to get a huge number of samples (millions) in just about a second. It's true that we've substituted precision for speed, but if you let Monte Carlo run for awhile it's going to pretty quickly stabilize towards the true value. Often, these approximations are good enough.

Examples

So let's put in some ranges and boards and see what happens.

For instance, if you have a nut flush draw with A9s and your opponent has one of the top 10% range (let's say he raised under-the-gun and you called from the big blind). How big is your chance of winning if the board is paired, versus if it's not paired? I seem to remember hearing that we must be careful with flushes on paired boards. So let's put in the following values:

  • Player 1, Ah9h
  • Player 2, 10%
  • Board: Kh Kd 4h ... and run "Enumerate All."

A9s is 45.508% to win. Alright. What happens if one of the kings was a queen instead? Change the board to "KhQd4h" and run "enumerate all" again. See what happened?

Was it what you expected?

One more example: How about JJ vs. a player whose range is confined to pocket pairs and AKs? Try it! What happens if you add AKo to his range? What happens if you remove some of the bottom pocket pairs? Play around, and Evaluate!

Selecting specific ranges

By now you may have noticed that you can click on "Player 1," "Player 2" etc. But if you haven't, go ahead now and click on "Player 1" to the left.

In this new window, you have two different tabs. The first one, "Cards," lets you select two specific cards for the player. I don't usually use this, since it's just as easy (easier, really) for me to just type the two cards. Choosing the ace of hearts and the nine of hearts from this table is no faster than just writing "Ah9h" in the input field. But if you dislike keyboards or for any other reason feel more comfortable picking them from a list, this is where you do it. You can only pick two cards in this first tab though.

But if you want to select a specific range, you can switch to the "Preflop" tab in Poker Stove. This tab, unlike the last one, is extremely useful for everyone. It lets you do two things: It lets you (easily) input specific ranges, and it lets you play with the slider and see more specifically - and visually - which hands get included as you increase someone's range. So you have someone at your table who plays half of his hands. Have you ever seen just how wide 50% of all hands is? Go ahead and draw the slider slowly to the right until you reach about 50%.

Now clear it again (click the "Clear" button). Now you can select specific starting hands to add to your opponent's range. Try clicking a few. see how the percentage counter to the right of the slider changes? The cards included in your selected range are marked yellow, plus the specific hand that you've currently selected which is purple. To de-select a yellow hand, click on it again. If you want to undo your current selection, press "shift" and click it again. This will make it yellow. Now release shift and click on the hand one more time to remove it from your selected range.

Now, I want you to try two more tricks to add hands to your range:

1. Hold down the Alt-key and press any hand. See what happened? This is the same as taking that hand - let's say it's 84o - and adding the "+" in the input field, e.g. 84o+. This means any 8 with an offsuit kicker better than 4 (but lower than 8), i.e. 84o, 85o, 86o and 87o. As you can see, everything above the diagonal (marked in maroon - apparently one of the programmer's favorite colors) are suited hands, and everything below it (gray) unsuited. Since 85o and 85s are different types of hands, you will find that alt-clicking on these hands will lead them up to the diagonal, but won't included their corresponding suited/unsuited respectives. So to speak.

2. If I raise, he calls and he checkraises an A-8-2 flop - which he would with any pair - how good are my pocket kings likely to be? I go the Preflop tab, Ctrl-click on "AA," "88," and "22" (which will select all the hands containing aces, eights and deuces), click "Apply" and then OK. Then I evaluate this range vs. KK on the above mentioned flop and find out that I'm 52% to win. Does the result surprise you?

"How Often Does He Have a Flush Draw?"

A cool trick that can be done using PokerStove is finding out the likelyhood of a certain type of hand. This is done by putting in your opponent's standard range, and from there select your own cards in a way that makes your hand just one step weaker than the hand you're interested in. Sound complicated? Let's look at an example.

Let's say that you want to know how often your opponent is drawing to a flush, if his range is 30% of all hands, and the flop is 3h-Ts-Kh. What you do then, is put in your opponent's range, and that board, into Stove, and then invent a turn and a river that would give him a flush, but give yourself the second best hand. "Second best hand" depends on the board, of course, but it should be either a straight (and then it should be the nut straight) or top set. For example:

Poker Stove

Here, I've given myself QJ, which would make the nut straight, and then filled the board in a way that the only way my opponent can beat me is if he has a flush, i.e. invented a turn and a river. Be careful that you don't accidently make some other hand the nuts other than the one you want to check if he's drawing to. In the example I used above, be careful not to pair the board!

When you're using this technique, be careful not to "waste" too many of your opponent's outs in the process of filling out the board; it will skew the results. With flushdraws and such, it doesn't affect the outcome very much, but it will definitely have an impact on draws with few outs. For instance, if you want to know how often your opponent holds specifically AA, it's prudent to use a board that doesn't otherwise interfere with his holdings. For instance, removing all hands that contain deuces, and also specifically 3-3 from his range (which shouldn't change the percentages much) allows you to compose a board of 2-2-2-3-3 and then give yourself KK, and then run the simulation. If you have KK on a 2-2-2-3-3 board if he has no deuce and not 3-3, it's only specifically AA that can beat you.

One final example:

How often does my opponent have two-pair or better on a Kh-9d-5d flop? This is tricker than it might look, since you need to dodge the possibility of him "improving" when you add your dummy turn and river cards. This is somewhat similar to the trick of checking for AA, in that you remove cards from your opponent's range to make the simulation work. So, on a K-9-5 flop, you could again remove all hands containing deuces and treys (which should rarely be a big part of an opponent's range anyway) and make the turn and river be a deuce and a trey. Then, you give yourself TPTK, and run the simulation. Watch out for possible straights and flushes, though!

In closing, and as somewhat of a sidenote, this "trick" with PokerStove, if you can call it a trick, is very powerful in that you can start building an idea of how likely a flushdraw given a certain starting range, for instance, is. One of the things you may notice if you start playing around is that given a tight starting range, a flushdraw is only about half as likely on a broadway flop as it is on a rag flop! The reason for this, of course, is that if your opponent is tight, he's likely to play only the big suited cards. And if there's an A/K/Q of two suits on the flop, that leaves a much smaller range of his suited connectors available, so to speak.

My Experience with PokerStove

It really is a wonderful tool. It's a common human error to think that our power of estimation is good, when in fact it often sucks. We look at a situation and we try to gauge how likely we are to win, and often, we're wrong. One way to sharpen how well we estimate these things is to use PokerStove to get a feeling for different situations. In terms of hand analysis, especially in a poker forum like CardsChat or others, being able to reduce a certain problem to just a matter of what we think our opponent's range is, is invaluable. Advice is often given by people who guess what the villain in the hand has, and then guesses what our chances of winning are, and then guess what the best course of action is. Since the margin of error grows exponentially with every new operation that has an innate error, doing three operations with large margins of error means that the end result probably isn't too reliable.

Being able to reduce the margin of error for some of these operations is awesome. Being able to almost completely remove it is spectacular. If I know my opponent will push - and always push - with AA, KK, QQ, AKs/AKo and JJ after being re-raised preflop, I can run the numbers into PokerStove and get the correct answer for what to do. No more guessing.

It's very rare (I think it's happened twice) that I've used Poker Stove while playing to solve a hand that I'm currently on the clock in. In other words, I basically use it exclusively for offline analysis and study. I often look something up and find that I'm surprised by the outcome. And every time that happens, I've become a stronger player, because now I'm armed with a new piece of information. I expect that it will be quite awhile before I stop entirely being surprised by outcomes in PokerStove, and this is despite having worked with it quite a lot in the last year.

There's no doubt in my mind that it's one of the best tools available for learning to play solid poker. The only thing you need to have in order to run it is a PC, some spare time and a healthy dose of curiousity. Then start running simulations. How good IS jack-ten suited versus AA? What if you flop a pair? If your opponent defends his big blind 40% of the time, how bad shape are you in if he calls when you raise on the button with J5s?

Play around. Discover new things - become a stronger player.

* Whenever you're talking about percentages, you have to watch closely so that you're not making the easy mistake of confusing yourself or others. The percentage that you're putting in is the top number of combinations of hands. For instance, if PokerTracker tells you your opponent will raise preflop 10% of the time, that's the percentage you should put in. The difference is that there are 169 different starting hands in hold 'em, but someone who raises preflop 10% of the time, isn't raising preflop with only the 17 best starting hands. Because suited connectors such as AKs and pocket pairs are more rare combinations than offsuit hands like AJo, taking the 17 best starting hands would actually only constitute about 7.5%. For most applications, this is not that useful to keep track of, but for the sake of stringency I wanted to include it.

Return to the Poker Strategy Articles Contents page.