Betting units
Can someone please explain me the reasoning behind betting different amount of units per bet? Lets say you bet from 1 to 3 units based on your confidence level. 1 is least confident, 3 is most, and 2 is somewhere in the middle. Wouldn't you make more money in the end if you make all your bets a 3-unit bet, regardless of how confident you are? If you're a winner in the long-term, then all your bets (1-unit, 2-unit, etc) should make you money. Betting less just because you have a 53% chance of winning instead of 58%, won't make you more money as far as I understand. So why do this?
-
-
Kelly criterion.Comment -
The non-math explanation is that betting maximally all the time would be sub-optimal because the drawdowns due to overstaking on the low expectation longshots would reduce your ability to take advantage of the higher expectation opportunities more than the overstaking gains you. More mathematically, you maximize the expected rate of growth of your bankroll by maximizing the expected logarithm of your bankroll. In the context of a series of wagers this is known as the Kelly Criterion, like DJ said.
Of course if you just want to maximize your expected bankroll you should go all in whenever you spot +EV. This is why people who say they care only about EV get laughed at.Comment -
I'm not sure I understand this. Can you give me a practical example? For instance, I have $200 and would like to make 20 equally sized bets. So $10 per bet. In the end I win 11 and lose 9. Assuming they were all -110, I win $10. What can be done differently with variable bet sizes?Comment -
The general idea is that one of the factors that is relevant to optimum bet size is the likelihood of winning.
If you had a chance to bet me that at least one of the next five coin flips would be heads, you’d probably want to bet a lot on that. Not your whole bankroll certainly, since it’s still possible to lose and you don’t want to go broke, but a very sizable wager.
If instead you had a chance to bet me that at least one of the next two coin flips would be heads, that’s still a good solid opportunity where you clearly have an edge, but the risk of losing is quite a bit higher, so you’d need to be more cautious. It would still warrant a decent size bet, but it would be unwise to bet as much as in the first scenario.
In your example of the 58% play and the 53% play, if you bet the same amount on both, then you are either underbetting the 58% play, or overbetting the 53% play, or both.
A lot of people have a preference for flat betting just because it’s simpler, it’s one less thing to have to calculate and worry about, etc., and that’s fine. But in the long run, the person who does the math and is willing to vary his bet size will do better.Comment -
bet the same ammount per game sir. this feel shit on how confident you feel is for the birds. if your real confident on the game u bet and you bet 10 units and lose 3 in a row you're gonna go on tilt and get barrelled in and lose it all. find a percentage you want to bet per game and bet the same ammount sir. good luckComment -
I bet the same amount every play. Im equally confident in every play I make. Anytime I think I have an edge Ill play. I feel I can pick enough winners to make money in the long run. It works well for me. Ill change my unit size occasionally depending on bankroll growth/loss, but I flat bet. In the long term the only thing that will matter is the big unit plays. The 1 unit plays mean nothing. Why even play the smaller unit plays?Comment -
louisville kid has the right idea dat camComment -
Pat and Louisville,
Apparently you can't read the above, but please read up on the Kelly criterion before you make any more idiotic "contributions" to threads like these in the future.Comment -
You have no chance of losing in either of these 2 situations in long-term. Both situations are winners. If you gonna bet $10 on the first situation, then why not bet also $10 on the second? You would have more money in the end. If you bet $10 on the first situation, and only $5 on the second, then how do you expect to make more money after 1000 plays of each situation?
The more you bet, the more you win. Both, 58% and 53% plays, are winners. You can't lose.In your example of the 58% play and the 53% play, if you bet the same amount on both, then you are either underbetting the 58% play, or overbetting the 53% play, or both.
This is what I'm still failing to see - making more money with varying bet sizes.A lot of people have a preference for flat betting just because it’s simpler, it’s one less thing to have to calculate and worry about, etc., and that’s fine. But in the long run, the person who does the math and is willing to vary his bet size will do better.Comment -
Ah, but who says there will be a long term? That will be determined in part precisely by how you select your bet size in the here and now.
In the situation I sketched out, assuming your total net worth was $20,000, would you bet $20,000 on either of my propositions? Presumably not. And that's because if you lose that bet--which is very possible even though you have a nice edge--you are broke. And who knows if or when you'll ever build up a big enough stake to return to gambling.
So there is, then, some bet size greater than zero and less than $20,000 which best maximizes your long term bankroll growth by balancing the fact that the more you bet when you have an edge the more you stand to win long term, with the fact that the less you bet the less chance there is of destroying your bankroll and never seeing the long term.
This optimum bet amount can be determined through mathematical formulas that take into account 1) The likelihood of the bet winning, 2) The odds of the bet, and 3) The size of your bankroll. Vary any of these factors and you change the result.
So, yes, if you had an infinite bankroll and thus could safely ignore the risk of short term fluctuations, there's no reason to limit your bet size on your 53% or 58% plays, and no reason to bet different amounts. Just bet what you please and watch the money roll in long term. But if you don't have an infinite bankroll, you need to manage risk by restricting your bets, and 53% plays require a bit more restriction than 58% plays.Comment -
TLD has explained it perfectly as usual.
For example, let's say I have a bet with odds -500 (83.3% break even), that I believe will win 85% of the time. That is a 2% edge.
Let's also say I have a bet with +500 odds (16.67% break even) with a win probability of 17%, which is also a 2% edge.
So, you have two bets, same edge. Do you risk the same on each?Comment -
How many of bettors that vary their bets, actually know the exact likelihood of the bet winning? Probably very, very few, less than half percent. I think this money management sysem is for the very minority. At first one must be a winner, not a 50/50 guy. Once this is accomplished and (if) you know the exact win expectancy on each of your bets, then you can consider a more sophisticated system to maximize your growth potential without maximizing the risk factor. Therefore, as a rule, you will normally make less $ using this system, unless you have a hard-core formula that sets proper winning percentages. I think both money management systems are correct, but one applies to a much larger population of bettors, while the other to very few. A bettor should first maximize his bankroll by researching and looking for a winning formula, and only then think about how to maximize his bankroll through varying bet sizes, while keeping the risk low. It's just not logical to jump ahead of yourself.
Infinite bankroll is relative. You can have a $1 billion bankroll, but if you make 500 million bets, it's not really that infinite. In contrary, you can have a $100 bankroll and make 10 cent bets. Now you created a virtually infinite bankroll. I know 10 cents isn't much, but it shows that bettors that don't have a winning betting system, shouldn't use a finite bankroll. They should bet 1-3% of their bankroll until they figure out the exact winning percentage of a bet. I can hardly imagine more than a few people (relatively speaking) that actually have a system like that.So, yes, if you had an infinite bankroll and thus could safely ignore the risk of short term fluctuations, there's no reason to limit your bet size on your 53% or 58% plays, and no reason to bet different amounts. Just bet what you please and watch the money roll in long term. But if you don't have an infinite bankroll, you need to manage risk by restricting your bets, and 53% plays require a bit more restriction than 58% plays.Comment -
If you can't pick winners above your break even pt at whatever odds you are getting -- then your optimal bet size is 0. You should stop betting at all. These action junkies don't get this, but that 's not relevant to a serious investor.
My thoughts are that if you can in fact find a game with an advantage over the books line, then you can also quantify that advantage vs. other plays. In that case, yes you should vary your wage size based on your edge.Comment -
Are you talking about 2% edge at -500 & +500 odds type of thing, or normal pointspread & over/under odds at -105 average? Would this system apply to -105 odds?Comment -
Your edge is your edge, at -20000, or -105.Comment -
Arnold,
Just play with Ganchrow's tool for a while.
(let me rephrase...LOL)
Comment -
Well, like I said, one can certainly forego the advantages of sizing one’s bets correctly because one prefers simplicity, or for that matter because other aspects of gambling—e.g., picking winners in the first place—are more important to do well and you’d rather focus all your energies on those. But I took the original question to be asking why it’s advantageous to vary bet size when your edge varies.
Here I would have to say you’re just confused. You seem to be using “infinite” as a synonym for “large,” and—correctly—pointing out that there can still be safe and unsafe bet amounts even with a large bankroll.
Yes, there can, but that’s why I said “infinite” and not “large.” My point about an infinite bankroll was simply that if you didn’t have to worry about going broke, then you could pound your edge as hard as you please and there’d be no risk of not making it to the long run.Comment -
Listen if you knew 100% that you had a 53% edge or a 58% edge then kelly C is of course the only thing you should ever use. The problem is that handicappers fall in love with games and overrate there edge. Any time I see an edge Ill make a play. But I dont see any edge greater than another. Kelly hurts betters for several reasons. Most cappers can't even beat the vig over a large sample size of games. All kelly does is make you lose more money faster if you can't already pick 53% winners over a large sample size. The problem is that the math never lies, but the capper's ability of rating his edge does. Only pros who have beat the vig picking 53% + winners over atleast 1000 bets should use kelly. Even then pros will tell you that flat betting is the way to go. If your spot betting, like betting a large amount on one a game a week then maybe kelly could be for you. But if you a volume players picking 3+ games a night, and you can pick over 53% winners then it makes no sense not to flat bet. You can't lose, if your picking 53% flat betting. Look at people's records on this site they post. You will see shit like this. . .
108-83 (-2.6 units). That is ridic when they would be up a ton flat betting!Comment -
How do you know you have an edge at all then?
Can I book your bets?Comment -
Yes but if you look at those guys they "were" up something like +55 units before they gave out their 45 unit lock or something stupid. They aren't using kelly, but they aren't flat betting either.
Most gamblers can't look at things objectively to determine their edge. They know Boston bucket hooped them so they stay away or chase depending on the person. So when Boston finally loses 2 or 3 in a row to the west where they were previously undefeated it affects how they view the situation.
Lack of emotion when sports betting is the key. Good solid money management is 2nd. Money management also is crutial in determining the books you play at and what happens when you win. Will you get paid out? How long?
As durito said, if you can't quantify your edge, what makes you think you have an edge at all? That is a very valid point.Comment -
durito,
you can book all my bets, Im a winning capper and all my plays are posted. give me your number and we can set something up. I would love to take your money. I have an edge on every play which is why I make the play. Im done talking about the subject I was just simply telling the guy what my stance is on the topic. I've talked to many pros who have published writings and im very happy with my results.Comment -
You don’t have to “know 100%.” Like anything else in gambling (or life), you’re doing the best you can estimating it.
I doubt you “know 100%” that every bet of yours has precisely an identical edge. And yet you’re choosing bet sizes as if they do.
If someone plugs in incorrect values to the formula, then, yes, their resulting bet amount will not be optimum. And if they are more likely to overrate than underrate their edge, then they will end up overbetting more often than underbetting. Agreed.
Really? So if you “see” that you have an edge at –7, and just as you go to make your bet, you notice you can make it at another book that was slow to move its line and still has –6, you’re incapable of understanding that the –6 bet has a greater edge than the –7 bet?
Rehash of earlier point, but, yes, if the figure you’re using for your edge is incorrect, then the resulting bet size will be incorrect. And if the tendency is to overrate your edge, then the result will be betting too much rather than too little.
But flat betting also means routinely incorrect bet sizing. So it’s not as if you’re avoiding the problem by avoiding using Kelly. At best, you’re randomly underbetting and overbetting, as opposed to your hypothetical Kelly misuser who systematically overbets.
People who are betting with the worst of it are going to lose long term regardless of whether they use Kelly or not. People who are betting with an edge will maximize their advantage by basing their bet size on 1) The likelihood of their bet winning, 2) The odds of their bet, and 3) The size of their bankroll.
False.
I’m sure there’s no shortage of people hanging around sportsbooks and posting on message boards trying to come across as sharp by regurgitating common myths like “flat betting is the way to go” in an all-knowing tone. But I’m afraid what you’ve got hold of there are not “pros” but “blowhards.”
Number and frequency of bets is irrelevant. Kelly maximizes bankroll growth. Doing something else—like flat betting or misusing Kelly—does not.
The argument:
With strategy A, it’s not possible to lose if you pick 53%.
With strategy B, it is possible to lose if you pick 53%.
Therefore strategy A is superior to strategy B.
is an invalid argument.
You may find it more psychologically pleasing to always make the identical amount of money if you hit 53% over a certain number of games, but for someone whose goal instead is maximizing bankroll, Kelly is better. For some finite runs, your small but constant profit will be better than the Kelly user’s results, and for some finite runs the opposite will be true. But overall, Kelly is the superior strategy for maximizing bankroll.Comment -
Awesome post TLD I really appreciate it just looking at some posts and saw yours. I was curious what a good method you would recommend to doing the last thing you wrote, as in a person that does math is wiling to vary the bet size. How do you choose when to increase bet or when to decrease other then just a gut or feeling? I have at times varied from anywhere from $5 to $150 myself and have done fairly good besides this tough streak last week, so wondering how better I can learn to manage and maximize profits.
My feeling is better to win consistently small the lose big, as that has worked so far for me. But just curious what I can improve on. Thanks for your help!
Comment -
The so-called Kelly system is the mathematical method of translating edge and bankroll into bet size. A search should reveal many, many threads on this. Look for the ones by Ganchrow.
But remember the “garbage in-garbage out” principle. Kelly takes certain values you provide and tells you how much to bet. But Kelly of course has no idea if those values themselves are accurate. If you’re plugging in incorrect values, you’re no better off than the flat bettors.
It’s also worth noting that Kelly maximizes bankroll growth, which may or may not be your goal as a bettor. Based on your statement that you’d rather consistently win small than take bigger risks for bigger rewards and possibly lose big, you may want to modify the Kelly bet sizes accordingly.Comment -
GL to you tld, im exciting to see your picks. Do you post any?Comment -
Im dispointed I can't follow a talented capper like yourself and the rest of the stars that crazy listed. GLComment -
This is pretty laughable. I've read about 500 condescending donjuan posts and I eagerly wait his picks, but all I hear is the sound of crickets chirping. Seriously, put up or shut up.Comment -
Comment -
It would be nice to know if those players are winners in their life-time. Not over a 1-2 week period, but overall. Until then it is just talk. Would be also nice to compare their bets with flat betting and see the difference.Comment
