The most important part of being a successful sports bettor may be the ability to find value on a consistent basis over the course of a season. Behind that though, is the ability to properly manage your bankroll (the amount of money you have set aside for sports betting). Bankroll management is all about ensuring that you’ll be a profitable sports bettor if you’re making good picks. If you do not have a system in place and tend to put an arbitrary amount on each and every bet, you could end up losing money despite a solid winning percentage. This happens when the bets you’re putting more money into end up losing more often than the bets you end up putting less money into.

**Using Units**

Before jumping into the various bankroll management systems, let’s talk about how we define a Unit. FantasyPros has an article that breaks down units in more detail, but let’s cover the basics here. A Unit is the baseline amount of money you’ll be wagering on a bet-to-bet basis. Beyond this, units are the best way to track your profitability, as it goes beyond just the amount of money you’ve won/lost and gives you an easy way to compare yourself to other sports bettors. Generally speaking, a Unit should be one, maybe two percent of your total bankroll. Any more than this and a losing streak can do some real damage to your total bankroll.

**Tracking your Bets**

There’s a lot you can do with tracking your bets, but at the very least you should use a spreadsheet to track the date, who you’re betting for and against, the spread/total/line, the number of units in play and the number of units won or lost. Of course, if you want to go full tilt with it (and have the time to do it) you can track everything from weather, time of the game, when you placed the bet compared to game start time, etc. The reason for tracking all of this is to build a database of picks that you can use to pinpoint where you’re winning and where you’re losing. Maybe you don’t do well betting against a specific team, maybe you’re doing better at specific O/U totals, maybe you’re hitting on favorites at a more profitable rate than underdogs. All of this data gives you an insight into your strengths and weaknesses. It gives you the chance to dig into why you’re missing more on those specific picks compared to others and in turn become a better sports bettor. Even if you don’t do all this, at the very least you can accurately track your profits/losses.

**Bankroll Management Strategies**

With some of the groundwork laid, let’s look into the various management strategies that you can use. Each of these has its own pros and cons. I’ve arranged them starting with the most basic strategies, then building up to the more complex. Ultimately, you should go with the option that works best for you and fits your own betting style. If you’re using a bankroll management system, and stick with it, then you should find yourself being a more profitable sports bettor, as long as you’re winning above the break-even rates.

*Fixed Unit Model*

If you’ve never done sports betting or you’ve only dabbled with it in the past, but have never used a bankroll management system, then the fixed unit model is the recommended approach to start off with. It doesn’t require any extreme math, it has little in terms of variables, and is the most consistent approach. Essentially, with this method, you’ll always wager 1 unit and 1 unit only on each and every pick regardless of the odds, your perceived confidence, your recent win/loss streak, etc.

As mentioned above, it’s recommended to have a unit size at one or two percent of your total bankroll. So let’s say you have about $1,000 to set aside for sports betting this football season (as a quick aside of responsibility, please remember that sports betting isn’t a get rich quick scheme, so you’re bankroll should only contain an amount of money you’re willing and financially able to lose). At a unit size of one percent, then you’re looking at a unit of $10 (1000 * .01 = 10). With this, you’ll be betting $10 on each and every wager.

It’s recommended to set some milestones on when you’ll reassess what your unit will be. Perhaps it’s every time your bankroll grows by $100, or maybe it’s every $250. Whatever the number is, be consistent and stick with your system. It’s easy to get on a good winning streak, feel unbeatable, and then increase your unit size, only for the inevitable losing streak to begin shortly after this. That’s one way good sports projectors end up losing money despite being profitable on a unit basis.

**Pros**

- Easy to track wins/losses
- Simple and to the point
- If you win at a greater than break-even percentage, you’ll be profitable

**Cons**

- If you’re strong at determining confidence then you may be missing out on some profit
- It doesn’t take into account how much you’ve won or lost, meaning your future wagers may be more or less than the one or two percent of your bankroll that you started with

*Percentage Model*

The percentage model is similar to the fixed model, but instead of the unit dollar amount being fixed, it’s the percentage of your bankroll that is fixed. Much like with the fixed method, let’s say you start with a $1,000 bankroll and have chosen one unit to equal one percent of your total bankroll. This means your starting wager is $10. However, this will change on a day to day basis. Let’s say after one round of betting you’ve won 1.5 units or $15. Now you’re bankroll is $1015 so on day 2 you’re new unit size is $10.15 (1015 * .01 = 10.15). Let’s say this time you’ve lost 2 units this time around or $20.30. Your new bankroll amount is now $994.70 and you’re new unit size would be $9.95 (994.70 * .01 = 9.947). So in this scenario, you’ve lost 0.53% of your bankroll despite only losing 0.5 of a unit. Now this example could have easily gone the other way where you’d have made more with the percentage model over the fixed model.

**Pros**

- You’re always betting one or two percent of your bankroll regardless of what you’ve won or lost
- It takes advantage of a winning streak as you wager more

**Cons**

- It’ll take longer to recover from a losing streak as you’ll be wagering less
- The timing of your wins has a bigger effect on your overall profitability
- The extra variables put more luck into how profitable you’ll be

*Potential Return Model*

The fixed and percentage models take only your bankroll into account when determining a wager. The potential return model is a variable model option that takes odds into account. With this method, you’re making wagers in order to win one unit as opposed to risking one unit. Let’s look at a few examples to see how this works. If you’re betting against the spread or over/under for a football game, you’re likely to see odds of -110 or to win .91 of your wager amount (100/110 = .91). So with this model, instead of wagering 1 unit to win .91, you’d wager 1.10 to win 1 unit (1/.91 = 1.099). Now let’s say you’re betting on an underdog at +120 or to win 1.20 of your wager. You’d now be wagering 0.83 units to win 1 unit (1/1.2 = 0.83). This model is taking into account that favorites should win more often, and therefore are less risky, while underdogs will win less often and are inherently riskier.

**Pros**

- Level of risk is taken into consideration
- If you tend to fair better with picking favorites, then this model will be more profitable than a fixed method

**Cons**

- Your wager amount varies from game to game, regardless of your confidence level
- If you tend to fair better with picking underdogs, then you’ll be leaving money on the table compared to a fixed method

*Confidence Model*

The confidence model is a variable model that adds a new wrinkle to the equation. Based on how confident you are in a particular wager, you may risk multiple units (or bet to win multiple units if you’d like to combine this with the potential return model). Based on your level of level risk aversion you could choose to use a 1-3 unit scale or a 1-5 unit scale (I’d recommend the 1-3 scale personally, unless your confidence levels have shown proven accuracy, as even at a unit of 1% if you lose a few 5 unit wagers, you’ll be eating through a lot of bankroll). If you are eventually interested in using the confidence model, I’d recommend tracking your confidence level along with everything else even if you’re not betting multiple units. This way you can track how accurate your confidence levels are prior to deciding whether or not to go this route.

To use the confidence model simply come up with your confidence in a bet, if it’s low then stick with 1 unit, if it’s medium go with 2 units, and if it’s high go with 3 units. You can also use a 2 unit confidence scale if you prefer to be conservative (I’d generally recommend this until your confidence levels have been proven over a large number of wagers). Lastly, you also have the option of betting a half unit. This would be reserved for when you like a very heavy underdog who is extremely unlikely to win.

A very important note: If you choose to begin using confidence levels, then you need to remain disciplined! It’s easy to go on a winning streak and all of a sudden all you’re wagers are 2, 3, 4+ units. This is a recipe for disaster as all sports bettors (no matter how good) will have losing streaks, and if you’ve gotten into the habit of betting multiple units, you’ll burn through our bankroll quickly. In fact, if you do go this route, then copy and paste that note into a monthly reminder to keep you honest.

**Pros**

- If your confidence levels are accurate, you’ll be more profitable compared to the fixed/percentage models on their own

**Cons**

- If your confidence levels are inaccurate, you’ll be burning through your bankroll considerably faster.

*Kelly Criterion Model*

Finally, there’s the Kelly Criterion Model. It takes everything into account and a little extra. Before I break this model down I want to add a caveat, this model should be reserved for only the best of the best, as if you’re not close to spot on with your projections, then you’re very likely to lose a lot of money very quickly. So this last one is more for academic purposes than a recommendation to use this thing anytime soon. For this model, you’ll need to figure out an exact percentage of the time you believe a wager will win (meaning if you are considering the model, I’d suggest doing this exercise and tracking it even if you’re not using this method yet). The exact equation will look something like this:

(Decimal odds of your potential return*win probability – loss probability) / Decimal odds of your potential return = the percentage of your bankroll you should wager

Let’s look at a normal pick against the spread where your odds are -110 or .91 in decimal format (100/110). Let’s say you expect your wager to win 55% of the time. Then the Kelly Criterion Model would give you 0.055 or 5.5%.

(.91*.55 – .45) / .91 = 0.055

That’s right, at just 55% of a spread bet this model is saying to go for over 5% of your bankroll. Remember that even the best of sports bettors are completely happy at hitting 55% accuracy on picks against the spread. The other part of this model is that it can tell you to stake some pretty crazy amounts. For example, let’s say you see a match-up as a 50/50 toss up, but you can get +125 odds on the underdog. The Kelley Criterion Model would tell you to wager 10% of your total bankroll.

(1.25*.50 – .50) / 1.25 = .10

Because this model tends to spit out some numbers on the higher side, more so than what most expert sports bettors would ever recommend wagering, there are various adaptations you can take to this model. You can cut the suggested wager amount in half or by quarters (my recommendation would be to go with the quarters method until you’ve shown profitability over the long-term). Using these methods with the against the spread pick would have recommended 2.75% with the Half-Kelley and 1.88% with the Quarter-Kelley. With the underdog pick, you’re still looking at 5% with the Half-Kelley and 2.5% with the Quarter-Kelley.

Now let’s do some math to illustrate how bad the Kelley Criterion Model can go. You can use the following equation to see the average amount you’d win on a bet at any specific percentage:

(Potential Win * Win Probability) – (Amount Bet * Loss Probability) = Average units won

So if you were to win that bet 55% of the time like you predicted, you’d make 0.27775 units on average or 27.775 units after making that bet 100 times at 55% accuracy (5.005 * .55) – (5.5 * .45) = .27775. This is a pretty solid return on investment and you’d be happy with having risked the extra amount. However, let’s say on that pick against the spread, you’re only accurate on that pick 52% of the time. Then you’d end up losing an average of .0374 units per bet or 3.74 units over 100 bets at 52% accuracy (5.005 * .52) – (5.5 *.48) = -0.0374. Think about that for a second, being even 3 percent off on a very common bet can be the difference between a nice profit and a small loss. Had you accurately projected it at 52% the Kelly Criterion would have told you not make any wager at all on that match-up. Now let’s take this one step further and see what can happen when you’re way off. Now let’s assume you only hit that wager at 45% instead of 55%. You’d end up losing an average of .77275 units per bet or 77.275 units over the course of 100 bets at 45% accuracy (5.005 * .45) – (5.5 * .55) = -.77275. That’s right, if you’re considerably off in projecting win percentages, you could very well end up losing a considerable chunk of your bankroll.

**Pros**

- If you’re practically spot on or very conservative with your win percentages, you can be very profitable by risking more than a unit or two.

**Cons**

- It’s a complex system
- There are a ton of variables that go into this
- If you’re not spot on with your projected win percentages and/or are a little liberal with them, then you can end up eating through your bankroll faster than any other method
- It takes extreme accuracy to be profitable

**In Conclusion**

There are various ways to manage your bankroll. It’s highly recommended that you start with the most basic models until you’ve truly mastered your strategies and then build from there. Don’t feel like you need to go straight to any specific model. On a personal note, I prefer a Fixed-Unit Confidence Model, meaning I always bet in whole units, but I’ll occasionally bet 2 or even 3 units on rare occasion. With that said I’m very conservative with my approach and rarely risk more than 1 unit, especially against the spread or on over/unders as those bets, while they have value, tend to be closer to even than what you can sometimes get on the money line. I bring this up to remind you that if you find a system that’s working for you, then stick with it unless you have the data to prove wholeheartedly that another system would be more profitable for you.

*Kyle Kontos is a featured writer at BettingPros. For more from Kyle, check out his archive.*