If you want to make money in poker it is possible, but it should never be the goal. Money is only the possible result. It is only a probable side effect of making good decisions. Having fun is a side effect of playing well and making money. Therefore the goal of poker is to make correct decisions. Money is the probable side effect of correct decisions. Let me repeat for emphasis. The goal of poker is to make the correct decisions. If you succeed in this goal, the probable side effect will be more money.
So if opponent is tight and plays JJ+,AQ+ in a spot, your Ace King or QQ is no good, fold. If opponent plays 88+,AQ in a certain spot your AQ suited is no good and your ace king suited is a coinflip and ace King offsuit is actually slightly behind. On the other hand, these hands can still be called if you are getting the right price so you win say twice the amount you risk. But each of these decisions require a lot of knowledge about what's good in certain spots or at least a good feel based upon doing the math. Experience at the table will never get you enough repetition to know with as much precision as those who do put in the work.
Expected Value
Good decisions are quantified as decisions that after examining all possible choices result in a positive expected value or a positive EV which is often denoted as +EV. But they also must include bankroll decisions.If you flip a coin and you lose $1 when you win, and lose $1 when you lose the expected value is ZERO. This can be calculated as probability times result plus probability times result.
(.50*1)+(.50*-1)=(.50)+(-.50)=0
Positive expected value calculations determine when the probable side effect over the long run is more money assuming good bankroll management. It can be calculated by the sum of the probability of each result multiplied by the corresponding result.
So let's say you have AQo and opponent only has one of a few cards. A pocket pair of 88 or better or AQ and he always will play any card in this range. You are 39.55% to win vs this range. Intuition should at least get you to estimate close to this number if you have done the work. say 40%. If your opponent has moved all in for $40 into a pot of $10 after you raised $4 and everyone else folds, should you call? You have to call $30 more. You win $50 when you are right, lose $30 when you lose. You are getting 5:3 or 1.667 to 1. This means you need to win at least 1/2.667 or 37.5% to win after rake in order to call.
Your expected value is about:
(.40*50)+(.60*-30)=$2. Assuming your raise attempt to risk $3 was a break even decision, the result of both raising and then calling profits $1 less than it would in situations when you get callers or opponents to fold but it profits $2 more than folding. In some cases, you make an unprofitable decision to raise in the first place, so the EV calculation of +$2 is actually only gaining back some of what you lose from raising. In other words if you always are raised and usually fold, you usually lose $3, but in this one instance, you are able to gain $2 of that $3 you usually lose back. So just because this is "+EV" or better than folding, doesn't mean the decision to get into the spot was necessarily profitable*.
This point of emphasis is made so you understand that the EV of any particular decision is often connected with the prior decisions. You can make a correct decision in one spot, but be incorrect to get into that spot in the first place.
*(To be clear, it is unlikely that raising Ace Queen is ever a mistake but if a hand like A2 against a reckless opponent yielded $2 to call their all in while the individual mistake after you raised may make money in most cases you usually will fold to everyone else's raise and usually you lose money so that $2 may not be enough to make up for where you lose a portion of that $3 you risk initially).
Every decision should be analyzed including what hands you play preflop.
Over the very long run with consistent bankroll management, your bankroll is the starting bankroll (and any additions or subtractions to it) plus or minus the sum of the EV of all decisions you make.
If 50 hands yield $1 each but 11 hands after that yields -$5, you may be capable of being a winning player, but overall you will not be because of tilt being the biggest leak in your game.
Unfortunately, break even decisions while in "expected value" are neutral, in terms of bankroll decisions, they are negative. Even positive decisions with poor bankroll management can lead to the probable outcome over a long period of time being losing money. The goal therefore cannot be solely based upon "expected value". It is vital to understand why.
The reason break even decisions lose money is because any bet has both the possibility of having a positive result and a negative result. However, several negative results strung together require a disproportional number of positive results to break even. To illustrate this start with $100 and imagine losing 50% then gaining 50%. Think you are back to even? Think again. A 50% loss from $100 takes you to $50. A 50% gain from $50 only takes you to $75. So a break even proposition has actually cost you 25% of your bankroll given this amount of volatility over this amount of time.
You may incorrectly deduce that you'll just increase the amount as you lose. But this doesn't solve the underlying problem is that you now have the same probability of an even greater loss, such that the gain once again is not enough to offset the loss. The only possible solution other than getting money from other sources (and even then sometimes it's not enough) is risking less to reduce volatility or to dramatically increase your edge so that type of drawdown doesn't happen and the gains are skewed to the upside large enough to overcome them. It's very uncommon to have that great of edge, and those with a great edge tend to be overconfident even more so and risk their bankroll more liberally than they should to their own detriment.
Money Management
Fortunately as we will illustrate below and explain later, risking 1/4th as much in the worst case scenario will still produce 40% of the results (in the long run of infinity). In the best case scenario you will go from a losing player to a winning player simply by having less to overcome on your downswings. You will also have a much greater probability of a positive outcome over a fixed period in time, and 1/4th of the volatility and emotional stress.
Greater volatility over longer periods of time will cost a fixed bankroll to go to zero, and a growing bankroll (such as a bankroll derived from excess income at a job) will continue to over time decline. Certainly it could go the other way. You COULD gain 50% first to $150 and then lose 50% to $75. The result is eventually the same over infinite time horizon. Timing it, or varying your bet size after you lose isn't going to compensate for a losing philosophy of playing a losing game.
This is what happens if you risk over 2 times the "Kelly"
However, respecting that you don't have an infinite time horizon is more reason to bet less if you are looking for the probable outcome of more money. Over a fixed time horizon the supposed "optimal" amount to bet is actually way too much unless you want a "lottery type" distribution where a very very very small minority ends up with a very large return and a large majority end up at a poor or negative return, while the upper 25% may have satisfactory returns.
Betting less instead would produce a satisfactory return maybe 40% of the time but without the absurdly unlikely outliers that will never actually occur to any one individual in the history of the next 100 generations to any one of billions attempting it. I don't think people realize that you can pretty much rule out the possibility of the "positive outlier event". The length and impact of the "infinite time horizon" cannot be overstated enough. Having the duration of time of the entire universe and all life forms ever in history of the universe would be less significant to infinity than the size of an atom to the size of the universe.
Even though it's a stock trading website, there is plenty of useful information in understanding modeling risk with a massive fixed, known edge at various risk levels produces distributions of results over a fixed time period. More about Risk Management here.
At the extreme, not even a single pixel can represent the difference between the number of people achieving outlier results and zero. At 10% most result in losing over 90% of their bankroll maybe 0.5% are outliers and 0.01% produce an average of an amazing return. At 20% risk 999.99 go effectively busto. Dream on.
You can see that with starting bankroll of $1000, larger amounts of risk produces a larger and larger percentage of people that actually lose money over 300 bets. It's difficult to really see what the actual percentage is, but the actual numbers wouldn't apply anyways. The concept that nearly ALL results effectively go bankrupt when they are betting closer to the supposed "optimal" amount to maximize long term growth goes to show you that even betting 1/3rd of this amount is risky.
In theory you might be able to eventually get your 10 cent bankroll up to a million when the 1/1000 outlier actually hits a series of several wins and eventually all those major outliers on either ide even out to the average. But we are talking about "quattroduodecatrillions" of hands until that happens and we are also talking about playing at stakes using fractions of a penny which isn't reality. Or if you have an income of $1,000, you are just going to reload and lose it all every single month or so and maybe over 1000 months or perhaps 4000 before you finally hit that one outlier month in which you pay for all the months before it and then some. The problem is a human life span is only 960 months, and only probably 480 of it if you lived and breathed poker could be spend playing poker. And how boring of life is that to be broke for most of your life?
What's worse, is you can never really retire with this strategy or play poker for a living with a high amount of risk if you don't have some income elsewhere. In order to really play poker for a living, you're going to have to play at absurdly low risk percentages such as 1/30th of the "optimal bet size" at risk at any one time (we're talking like 0.20% of your bankroll at risk) and have so much excess separate from your bankroll saved on the side that your probable result is centered right around break even anyways and most likely you are just withdrawing from 2 years worth of savings, but the overall "average" with 0.20% of your bankroll is theoretically enough to live off of if you made it each and every month with certainty.
So if you made $50 an hour at $2/5 live games buying in for $500, you'd need a bankroll of more than $250,000. Well, maybe earning $10 an hour is enough at $0.50/$1.00. You'd need $50,000. net worth. If you always had a way to make ends meet if you absolutely had to and were okay with having a risk of ruin and living life in fear of going broke, you could use the standard measurements of risk management. Maybe you are okay with knowing that you can always sell your house or take out a mortgage if you have your home paid off. Either way there are very few people in the world with
1)The skill to play poker
2)The understanding of bankroll management
3)The willingness to grind at less conservative stakes as a side job until capital has been built up
4)The discipline to stick to insanely conservative bankroll management strategies
5)The passion for the game to play for thousands of hours each year.
6)The discipline to not spend their earnings at at a higher rate than their "expected" earnings.
Realistically there are plenty of people that can earn professionally if they have sources to be staked and endorsement deals and/or business income, or book deals and poker coaching income, but the days of easy endorsements and free capital in the poker community is gone and in a bad economy your edge will likely go from positive to negative and the difficulty of games and easy money floating around will evaporate.
Furthermore, if you're an online player, you'd have a higher average and lower volatility multitabling at 2% risk with 5 tables than 10% risk at one assuming the edge was the same so risking more has no practical advantage in any normal form.
The principal of volatility hurting more than it helps is true even with an edge if you do not manage bankroll properly. While some people may chime in about compound interest and how several wins in a row will compound as well, that may be true to a very small degree, but the benefit is smaller than the drawback of risking more. It is not worth it and you will not benefit to a strong enough degree to compensate for the cost of volatility if you risk too much. The amount you can risk with a break even proposition to have positive results is zero. Not to mention it's not worth your time.
The casino will always have a larger bankroll than you, and if they don't, they will kick you out because they don't want that kind of risk. Phil Ivey was kicked out for playing craps, a game in which he had the worst of it from very small casinos because he was placing very large 100k bets. The casinos don't gamble, they play mathematical certainties over enough times so that they make money, and they didn't want to allow to put their entire operation at a risk of ruin. They will turn down your "business" if either they know you have the best of it at a small stakes or you are playing at such large stakes that their "probable outcome" is not more money over a very long term.
For similar reasons, decisions with positive EV can still have negative long term expected growth rates. That may sound like a contradiction, but it's not. The "Kelly criterion" is like "EV" but for your bankroll over an unlimited time horizon. It already compensates for the fact that a 20% drawdown in your account requires a 25% growth rate to get back to even and a 50% drawdown requires a 100% gain. But realize that over an infinite time horizon a 50% loss can eventually be recovered, even if for all practical purposes it cannot easily be in real life in a finite timeframe.
While an individual decision repeated for the same bet amount an infinite amount of times may be positive, repeated as a percentage of bankroll or within the realistic parameters of actually having a limited bankroll, the long term result is losing money.
When Winning With Bad Management Still Results In A Loss
Take for example a winning poker player who wins 10% of all tournaments. 10% of the time he wins an average of 20 times his buy in. Clearly this is tremendously profitable in terms of EV. Let's calculate how much in terms of buy in after 10 tournaments. (1win*20buy ins)+(9wins*-1buy ins)=11 buy ins profit every 10 or gain over cost*100 is 11/10=1.1*100 110% net ROI.However, the odds of going 6 tournaments in a row without a win are greater than 50% .9^6=53.14% If this player risks 100% of his bankroll he has a 90% chance of going bust. If he risks 50% of his bankroll, EVEN if he adjusts his bet every tournament so he only bets 50% of his remaining bankroll after each loss, he has an over 50% chance of having less than 2% of his bankroll before he wins. When he finally does win, he will be doing so at such small stakes that 20 buy in victory won't even take him remotely close to where he was. As such, he has a very good chance of not even being able to afford a low stakes buy in with this plan.
The theory of "adjusting downward" as you lose is reasonable, but only if you start low enough to begin with. And a fixed percentage at an amount between the "adjusting downward" amount and the starting amount probably has the same impact of adjusting downward with less volatility and more consistent gains. A "fixed percentage" actually does adjust upwards and downwards since it's a percent of your bankroll.
The theory of increasing your risk as you lose is unreasonable as this increases your chances of losing such a greater amount that you won't even have enough to continue even if you have Bill Gate's credit line, and if you did, you'd have to start at such a small bet that it wouldn't even be worth your time and you'd still put unreasonable risk of losing everything. Take a coinflip. Risk a $5 and double the risk 35 times or so and you'd lose all $85 billion of Bill Gate's bankroll. Gain $5 35 times and you've gained $175. Is $175 worth risking $85 billion? Is it worth the time? This strategy is guaranteed to lose eventually it is not nearly as logical as you'd like to wish it was.
You cannot change how you "time" the amount that you bet or to turn a profitable decision or unprofitable decision to be more profitable or less unprofitable. You can only influence the probability of a gain at the expense of the downside or probability of a huge loss. The best decision is to sit with a fixed percentage of your bankroll or lower than that fixed amount, and to determine which "fxied amount" suits your goals.
Fortunately there is a solution to increase the results if you are a winning player, and it involves understanding the nature of risk.
There is an amount to bet which over bets that approach an infinite amount, your value is maximized. Not too much, not too little... Just right. However, as you approach this "optimal" point the benefits for increasing your bets have greater and greater risk at a smaller and smaller reward.
Consider risking half the kelly has 3/4ths the return with half the volatility. This means you recover from your all time highs in bankroll much faster and make new highs much faster on average, although the upswings will not be as large and the overall trend may be slightly less strong. Risking 1/4th the kelly still has 40% of the return as the full kelly with 25% of the risk. In other words, risking less is almost always better from a practical risk/reward perspective.
This is only what risk looks like given CERTAIN outcomes. With uncertain outcomes over an infinite amount of time there is risk of betting far too much. If you bet over 2 times the kelly the eventual result is losing everything.
As such, there is no reason to overbet the kelly at all.
Over FIXED time horizon, even as long as 10 years, everything changes. Betting the full kelly over a fixed time horizon results in a probability distribution to explain the results. Out of 1,000 players who play a kelly criterion strategy a very small number of them will have extraordinary results, while the majority will have below average results and the upper 25% may have above average results but it will be skewed by the top 1%. The bottom 25% will produce break even or even a loss, while the bottom 25% (larger or smaller depending on the number of profitable "bets") will have lost 90% of their bankroll.
Since this essentially amounts to a lottery ticket chance of having enormous gains that are far beyond what you could ever spend, there isn't a lot of added value to taking on a "Kelly" worth of risk. 1/5th of the Kelly is acceptable to some, but probably not if you have to withdraw from your bankroll at regular intervals without some other means to live if you go through downswings.
However, risking less than 1/5th of the Kelly will produce both a high probability of very good results and good upside. There still will be a chance of pretty spectacular results, and still a small chance of poor results, but overall the outcome will be likely to be pretty good. There will still be a skew. Because the probable outcome is more money than less, THIS is a correct decision, and money is the probable side effect.
This amount of risk requires examination of your skill edge. That isn't something that is entirely certain so there is a slight "gamble" on your own skill. Or if you are playing an "equilibrium" strategy there is a gamble that others will be bad enough that you will beat the rake (which is a very good assumption) and that your mistakes will be small enough (playing without mistakes even with a fairly basic predetermined strategy is difficult and has mental and psychological challenges).
There is a reason that pros like Chris Ferguson and Phil Gordon and others have never gone broke and also live comfortable lives with poker playing at one time still being a reasonably significant part of it. Aside from having skill, they follow a set of bankroll management guidelines and very solid bankroll strategies. Since it is difficult to distinguish playing poorly with being unlucky and it is easy to convince yourself in defense of the ego that you were unlucky. Money doesn't care, more often than not it will flow towards those who make good decision, and some periods of time flow away from them, but overall if you do not make good decisions including bankroll management you will end up with less money. You can't afford to let ego get in the way so if your money says the right decision is to decrease the amount you risk, you should gear down.
The Non-Professsional Player
If you aren't a pro player and you don't want to make a living playing poker, but still want to give yourself a chance at getting rich, should you still care? Absolutely.Often times the way that seems like the fastest way to get rich quick ends up being the longest, as it will inhibit your ability to develop and grow when the chances fail, and when the chances succeed it will not bring you back to the same trajectory to overcome past and future failures as the result of trying to take this short cut.
In fact, the unprofessional player is less dedicated and therefore less prepared (even if they have natural talent) so they will have a smaller edge. While the non-professional may need experience to develop that edge, he needs to learn the fundamentals which requires sufficient play against enough bad players at stakes that are enough to matter, but not so much that emotion rules over logic. He also needs to not only survive long enough to develop that edge, but also maintain the ability to not have to move down in stakes or play less often.
However, if you have a separate income, and don't play as often there is SOME room for you to borrow from your future income and play in higher tournaments than your bankroll. (Please note: This is not done by going into debt, one should never practice debt as that will eliminate any bennefit having an edge has).
For example, if every weekend you like to play a little poker, and your excess cash after expenses is large proportionally to your bankroll, you may be willing to risk a bit more as long as:
1)You already have 3-6 months worth of expenses (some suggest more) on the side.
2)You don't have any outstanding debt.
3)You plan on playing for years
4)Playing at a particular stake this month won't prevent you from playing at the same stake next month without violating rules 1-3.
So if you earn $1000 a month after expenses and want to use all of it playing poker every month for 10 years you might say your future bankroll discounted for the future is maybe $50,000* (plus whatever you have as your bankroll now not including what's saved for other purposes).
(*Money now is worth more than money later, and money which requires risk must be discounted for volatility, and money used for this must be discounted each month for value that you must pass up later so rather than 10 years, you might only use 4 years worth after factoring this all in)
If playing at a particular stake prevents you from playing at that stake the next month, playing higher stake still inhibits your ability to save enough to move up, and earn something from poker (higher stakes are typically less lucrative in terms of profitability expressed in big blinds). If you want to play for say 10 years worth, take about 40% of the excess earnings over that period. Use the lesser number of either the average of what you made the last 3 years vs what you are making now.
There is a "bubble" bias of people having the most to spend right before a bubble. So it's possible you may overestimate your future earnings dramatically. Take monthly income times 12 to get yearly, then yearly excess times the number of years (in this case 4). This is your future bankroll in 10 years after being discounted. SO if it's $200*12*4=$9,600. Now 2% of that is $192.
So if you REALLY want to enter a particular $150 tournament you can borrow from your future bankroll. It comes at a pretty steep cost, but if you give yourself enough time to recoup the damages before you start again, you can give it a shot. Chris Ferguson went from 0 to $10,000 in about a year and most of it was spent at the 1cent 2cent table. That proves it is possible to be incredibly intelligent with your bankroll and still earn significant amounts, but he also took 8 of those 13 months grinding the very low stakes and freerolls and that doesn't mean it has to be done that way. He also played a few more hands than most will get to. He also earned his way to prove he is competent at higher stakes first after studying applied mathematics for nearly a decade and applying that work to poker so he knew he had a major edge.
It also certainly would help if you are FIRST able to prove your track record and profitability at a smaller stake for thousands of hands and then move up. Play is much worse live but you can compare the micros to the low stakes live and $2/$5 or $5/10 to $0.50/$1 online.
Daniel Negreanu on bankroll management and a foolproof plan to become a professional poker player.
Chris Ferguson 0 to 10k challenge.
This is why money is only a PROBABLE side effect of correct decision and it comes at the expense of a possible side effect of losing enough to want to throw up at even what many consider "conservative bankroll management". How well will you play when that happens? If you are focused on the result of money, you will tilt off and lose it all when you run badly. If you focus on making the correct decisions, and don't pay a lot of emotional attention to the money as best as possible, the money will eventually come over time. The slower path is ironically probably the more likely one to actually get to where you want to go. Study the game, EARN your advantage and go out and use that advantage over a large number of hands and continue to grow your advantage.
No comments:
Post a Comment