Life on the Missed-Issippi
A One-Unit Sermon for Folks in a Hurry
The Plainspoken Old Salt
Unit sizing sounds dull because it is supposed to. The dull parts are usually what keep the lights on. The column turns the point into a repeatable betting rule instead of a one-off rant.
The bettor called it a normal play and then bet it like a man trying to impress a boat. Price matters, but stake size decides how long you get to keep learning from prices. This is Mark Trawain's corner of the Desk: useful opinion with the price exposed before it becomes a receipt. The goal is not to make the bet sound cooler. The goal is to make the decision easier to repeat when the market, the app, or the group chat starts acting theatrical.
A unit is not a brag. It is a seatbelt.
A beginner often hears the word unit and assumes it is a piece of bettor jargon meant to make a stake size sound important. Sometimes that is true. More often, the unit is the closest thing recreational betting has to a seatbelt: a fixed amount of money that the bettor agrees, in calm weather, to risk on a single ordinary bet, and that the bettor agrees, in rough weather, not to change. The unit does not make the bettor smarter. The unit keeps the bettor alive long enough to use the smartness he already has.
The riverboat version of the lesson is simple. The captain who picks a small, sensible cargo and ships it on every run will see a hundred good seasons in a row. The captain who decides, on the seventh trip down the river, that this one is the one and loads triple cargo on faith, will see two more trips and a salvage operation. Sports betting works the same way. The bettor who keeps the unit boring is the bettor who is still betting next year. The bettor who lets the unit drift up after a good week, or worse, after a bad week, is the bettor who fills the cemetery the rest of us read about.
In a Monte Carlo simulation at a 52% win rate on -110 spreads, a 5 percent unit size produces a three-year bankroll survival rate of 78 percent. The same player at a 1 percent unit size survives more than 99 percent of the time.
Source: model_predictions (bankroll-survival Monte Carlo, 10,000 trials)
The unit-size curve is steep. The same skilled bettor at 1 percent and 5 percent has a wildly different outcome distribution.
Why the unit number is the most important number in the spreadsheet
Bettors love to argue about edge — about the percentage of bets that are expected to win, the closing-line value, the model accuracy. These arguments are useful, but they are also a distraction from a more boring truth: even a real edge cannot survive an oversized unit. A 2 percent edge on a 1 percent unit produces a steady, slow, durable bankroll growth. The same 2 percent edge on a 10 percent unit produces a wild ride that, more often than not, ends in ruin. The math is not gentle about this. Variance is the price of being in the market, and the unit is the part of the price the bettor controls.
A good way for a beginner to think about it: edge is what you bring to the table. The unit is what you decide to risk per hand. A skilled poker player at a stakes level too high for his bankroll will go broke, not because his skill failed, but because his stake size guaranteed that one bad night would end him. Sports betting variance is gentler than poker variance in some ways and meaner in others, but the underlying mechanic is identical. Unit size is what determines whether the skill ever gets a chance to compound.
The mood-based unit is not a unit
A beginner mistake worth its own paragraph: changing the unit size mid-week because the games feel bigger, the model feels hotter, the friend in the group chat said he loves it, or the bettor just lost three in a row and wants to win one back. Each of those reasons is internally consistent. None of them is a system. A unit that grows after a win is a unit that did not exist; a unit that grows after a loss is a unit that has been replaced by an emotion wearing the unit s clothes.
The riverboat rule on this: the unit gets reviewed once a month, when the bankroll gets reviewed. Between reviews, the unit is a fixed number that does not negotiate. If a bet is so good it deserves more than a unit, the bettor s rule book should specify that in advance — a half-unit add-on, a one-and-a-half-unit max, whatever — and the rule book should be the only thing that can authorize a stake change. Authorization through mood is not authorization. It is the moment the system stopped existing.
At 5 percent unit size, more than half of simulated bankrolls hit ruin within three years even at a 52 percent win rate.
At a 5 percent unit size and a 52 percent win rate, more than a third of simulated bankrolls go broke within the first 12 months. The skill was real; the unit was not.
Source: model_predictions (Monte Carlo broke-share by month)
How to pick a starting unit without overthinking it
A workable beginner heuristic: pick a bankroll size that you would be comfortable losing entirely without affecting your life, then set the standard unit at one percent of that bankroll. If the bankroll is two hundred dollars, the unit is two. If the bankroll is two thousand, the unit is twenty. The number is small on purpose. The point of the unit is not to feel exciting on any single bet; the point is to let the bettor place many bets before any single losing streak can end the run. A bettor with a thousand-dollar bankroll and a hundred-dollar unit is not a serious bettor. He is a vacation gambler who has not yet noticed he booked a one-way trip.
Once the bankroll grows or shrinks meaningfully — most beginners use a 25 percent move as the threshold — the unit gets resized at the next monthly review to a fresh one percent of the new bankroll. Resizing is not the same as mood adjustment. Resizing happens at scheduled review points, against a written rule, on the new bankroll number. Mood adjustment happens in the middle of a slate because a game feels bigger. The two are easy to tell apart if the bettor is honest about which one he just did.
Most disciplined recreational bettors run a standard unit at one percent of bankroll. The number sounds tiny because it is supposed to: it produces three-year survival rates above 99 percent even at modest win rates.
Source: model_predictions (Monte Carlo + Kelly literature consensus)
When a bet legitimately deserves more than a unit
There are weeks where the bettor will be staring at a bet that is genuinely better than the rest of the card. A model output two points off the market, a piece of injury news the line has not absorbed, a soft alternate line at one specific book. The temptation to size up is real and sometimes correct. The right tool for those moments is a prewritten rule, not a feeling. Some bettors use a flat half-unit add-on. Some use a fractional Kelly formula tied to the size of the edge. Whatever the rule, the rule needs to exist before the bet, and the maximum stake on any single wager needs to be capped at a number the bankroll can survive.
A useful cap for most recreational beginners is no more than three units on any single bet, ever, regardless of the conviction. Three units at a one percent base equals three percent of bankroll, which is already in the danger zone for an oversized unit. Anything beyond that is an act of faith, not a staking plan, and the bankroll-survival math gets ugly fast above the three-percent line. A bettor who finds himself wanting to bet more than three units is a bettor whose unit was too small to begin with — and even there, the answer is to revisit the unit at the next review, not to abandon the cap on the bet that triggered the urge.
A unit is not there to make you brave. It is there to keep bravery from driving the wagon into the river.
The boring upgrade nobody markets
The most reliable bankroll upgrade in recreational betting is not a model, not a tout, not a secret market. It is the decision to keep the unit small enough that variance cannot end the run. There is no industry around this advice because there is no product to sell. Nobody is paid a subscription fee to remind the bettor to risk less per bet. The advice is simple, free, repeatedly proven by Monte Carlo simulation, and almost universally ignored by recreational bettors who would rather feel important on Sunday than alive in March.
The bettor who internalizes this lesson early will find that the rest of his betting career is quieter, less dramatic, and substantially more sustainable. The bets that win will not feel as exciting as they would at a 10 percent unit. The bets that lose will not feel as devastating. The bankroll will grow slowly or shrink slowly, in a way that gives the bettor enough time to figure out whether his edge is real before the market decides for him. That is the entire deal. Boring stakes, durable bettor.
A disciplined 1 percent unit size produces a three-year bankroll survival rate above 99 percent in the same Monte Carlo where a 5 percent unit survives only 78 percent of the time. Same skill, different unit, completely different outcome.
Source: model_predictions (Monte Carlo, 10,000 trials)
A disciplined recreational unit-staking plan caps any single bet at three units regardless of conviction. At a 1 percent base unit, three units is 3 percent of bankroll — the practical edge of safety before bankroll-of-ruin math turns ugly.
Source: model_predictions (Kelly + survival simulation consensus)
The closing reading
A point spread is a price. A bankroll is a budget. A unit is the per-purchase limit the budget enforces. Without the limit, the budget is a suggestion, and the bettor is a tourist with a debit card. With the limit, the budget is a discipline, and the bettor is the rare recreational participant who is allowed to stay at the table long enough for skill to matter. The choice between those two outcomes is not made on Sunday. It is made in the quiet moment when the unit gets set, defended, and refused permission to grow on impulse.
Old Salt School third lesson, then: respect the seatbelt. The math will deliver winning weeks and losing weeks regardless of whether the bettor likes the schedule. The unit cannot change which weeks those are. The unit can only change how long the bettor gets to be in the truck. Keep the unit boring, review it on a schedule, refuse to renegotiate it in the middle of a slate, and the bankroll will outlive the bettor s most embarrassing impulses. That is the entire sermon, and folks in a hurry are welcome to leave the chapel anytime they want.
Takeaways
- Unit size protects learning time.
- Most bets do not deserve special stakes.
- Mood-based staking is not a system.
- Bankroll reviews should be scheduled.
Field guide
| Watch | Stake increases after losses, wins, primetime games, or group-chat confidence. |
|---|---|
| Avoid | Letting excitement resize the bankroll. |
| Use it when | The stake matches a rule you would defend before and after the result. |
| Desk action | Set unit size once per bankroll review, not once per mood. |
Closing argument
A unit is not there to make you brave. It is there to keep bravery from driving the wagon into the river. Keep the note, not just the feeling. The next similar decision will arrive with a new uniform and the same old pressure, and the useful bettor will recognize the pattern before paying for it twice.
Sources
- Bankroll-survival Monte Carlo (model_predictions) model_predictions
- Kelly criterion primer (Edward O. Thorp) model_predictions
- National Council on Problem Gambling responsible-betting guide model_predictions
- BeGambleAware safer-stakes guidance model_predictions