BY 
I lost my first bankroll because I never learned how to size a bet. The picks were mediocre, sure, but what actually killed me was putting $50 on games I liked and $300 on games I loved, with no math behind either number. A few of the $300 specials lost in a row, because a few always do, and the roll was gone. It took two more bankroll funerals before somebody explained units to me in a way that stuck.
This is the explanation I wish I'd gotten at the start, with the arithmetic attached.
A unit is a fixed slice of your bankroll
A unit is a set percentage of the money you've put aside for betting, and the standard range is 1 to 2 percent. That's the whole definition. Nothing mystical. If your bankroll is $1,000 and you choose 1.5 percent, one unit equals $15.
The percentage stays fixed while the dollar amount floats. Grow the roll to $1,400 and your unit drifts up to $21. Bleed down to $800 and it shrinks to $12. Your stakes breathe with your bankroll automatically, which is the discipline most of us can't manage on willpower alone.
Why units exist at all
Units solve two problems at once. The first is language. When a capper says she has two units on an under, you know exactly how much conviction that stake represents relative to her roll, whether her unit is $10 or $1,000. My record and yours become comparable even though our bank accounts aren't.
The second problem is proportion. Flat dollar amounts ignore the state of your roll, so betting $100 a game feels identical whether you're up big or down bad. Right up until the drawdown where $100 is suddenly a tenth of what you have left, anyway. Percentage sizing makes that mistake structurally impossible, because the stake shrinks as the roll does.
The worked example, start to finish
Say you set aside $1,000 you can genuinely afford to lose, money that pays for nothing else in your life. You pick 1.5 percent as your unit size, right in the middle of the standard range.
- One unit: $1,000 x 0.015 = $15
- A standard play is 1 unit, so $15
- A strong play might be 2 units, so $30
- A max play caps at 3 units, so $45
Notice the ceiling. Even your most confident bet risks only 4.5 percent of the roll. You can be wrong ten times in a row at max confidence, which happens to real bettors more often than anyone admits, and you'd still have well over half your bankroll left. Survival is the entire point.
Units won is the honest scoreboard
Win percentage is the most abused number in this business. A capper hitting 65 percent sounds unbeatable until you notice every pick was a -300 favorite, and at -300 you need 75 percent just to break even. He can post that shiny record and still be lighting money on fire.
Raw dollar totals mislead in the other direction. Someone who won $5,000 while risking $2,000 a bet had a far rougher run than someone who won $800 risking $50 a bet. Units won fixes both distortions because it bakes the odds and the stake into one number. Risk 1 unit at -110 and a win gains you 0.91 units, while a loss drops exactly 1.
Add that up across a season and you get a number no marketing department can dress up. Positive units won over a real sample means the picks beat the vig. Everything else is noise.
What a 10 unit confidence scale really tells you
Multi-unit plays are fine when they're capped sensibly, somewhere in the 1 to 3 unit range. Past that, the math gets ugly fast. A tout advertising a 10 unit lock is risking 10 to 20 percent of a properly built bankroll on a single game, and no edge in sports betting justifies that exposure.
In practice, giant unit scales are marketing. The 10 unit play makes the winning weeks sound heroic, and when it loses, the record tends to quietly reset. If someone's confidence scale runs to double digits, read it as a signal about the seller rather than the game.
How CAPTRACKER keeps units honest
All of that is why the CAPTRACKER leaderboard ranks cappers by units won and ROI instead of win percentage. Every pick on the platform gets timestamped and locked, with a two minute grace window and no edits or deletes afterward. Results settle automatically against ESPN data, so the record is simply the record. A grinder sitting on a modest pile of units at small stakes outranks a screamer with a gaudy win rate and a negative ledger, which is how it should work.
With 900+ tracked handicapper profiles, the sample sizes get real in a hurry. If the vocabulary still feels foreign, the glossary covers the rest of it, and our bankroll management guide goes deeper on choosing a percentage that fits your temperament.
Start smaller than feels natural
My bias, earned the expensive way, runs toward the bottom of the range. A 1 percent unit feels almost insultingly small when you start. Good. Boring stakes let you survive long enough to find out whether you can actually pick winners, and most of us discover the answer is humbling.
These days I bet mostly unders at one unit a pop, and my inner accountant sleeps fine. Pick a percentage and write it down. Then let the unit do the discipline your gut never will.