What is the best way to play roulette? People have been asking that for well over a hundred years, and I have spent a fair share of my own time asking it on cruise ships, where the casino never closes and the sea does the rocking for you.
I have seen some things at those tables. Stacks on Red a foot high. Players so far gone the dealer had to read their bets back to them, and the house only too happy to keep dealing. The cruise lines will take your money in any condition you care to hand it over. On this trip, though, I saw something new, and it is worth telling properly, because the arithmetic behind it is the arithmetic behind every high-coverage system ever sold.
The buy-in
It was late. The ship was somewhere in the dark between islands, most of the floor had drifted off to the bars, and the roulette table had gone quiet. Then a man walked up and asked for a buy-in north of six thousand dollars.
That does not happen on a cruise ship. The pit boss made calls. Someone verified something. More calls. Eventually the chips came out, black hundreds, a stack you do not often see at sea.
I will admit what I was hoping for. A buy-in like that, at that hour, with that much ceremony, meant a plan. I wanted to see a strategy I had never seen, something I could bring home and put through VIPER for everyone. So I took a seat at the end of the table and I watched.
The plan
Most ships now cap the table at $100 a bet. That cap is not a courtesy. A low maximum is the house quietly taking away your recovery options: you cannot press after a loss, you cannot double out of a hole, you cannot turn a bad run into a good one by getting bigger. You are at the ceiling before you start.
The man began laying black chips straight up, one to a number. Ten. Twenty. Thirty chips, thirty numbers, three thousand dollars on the layout for a single spin.
So there it was. Not a secret system. A high-coverage bet: thirty of the thirty-eight pockets on a double-zero wheel, leaving eight holes for the ball to find. He was at the maximum on every one of them, with nowhere to go up.
The dealer spun. The ball rattled, dropped, and landed on one of his. He was up. Surely, I thought, he pockets that and walks. Would you?
He put the same thirty chips out again. Hit again. Up twelve hundred dollars, and the pit boss had stopped pretending to look elsewhere.
Third spin. Hit. Three in a row. I felt for the dealer, a little, standing under that stare. I was rooting for him, too. Everyone at a roulette table is secretly rooting for the man who might take the house's stack for once.
But I was also counting holes.
What he actually had on his side
Here is the bet, spelled out. A $100 chip on a winning number pays 35 to 1, so the hit brings back $3,500 and keeps its own $100. The other twenty-nine chips are swept. On a double-zero wheel:
| Numbers covered | 30 of 38 |
| Chance the ball lands on one of his | 78.9% |
| Chance it finds a hole | 21.1% (about one spin in five) |
| A hit nets | +$600 |
| A miss costs | -$3,000 |
| Hits needed to pay for one miss | 5 |
That last line is the whole story. Every hit earns six hundred. Every miss takes three thousand. He needs five clean hits to pay for a single miss, and then a sixth before he is ahead at all.
Now the part nobody at the table was doing in their head. How long does the run last?
| Straight hits | Chance of getting there |
|---|---|
| 1 | 78.9% |
| 2 | 62.3% |
| 3 | 49.2% |
| 4 | 38.8% |
| 5 | 30.7% |
| 6 | 24.2% |
| 8 | 15.1% |
| 10 | 9.4% |
On average a player making this bet gets 3.75 hits before the first miss, and half of all runs are over by the third spin. Five in a row, the number that merely breaks even, happens less than one time in three. His three straight hits were not a heater. They were dead on the average.
Play the thirty-number bet until you miss and you will finish behind on average: about 3.75 hits worth $2,250, then a miss worth $3,000. There is no run length that fixes it, because the plan itself says you stop on the loss.
And that is with a cap that stops him pressing. A high-coverage system with a progression behind it can dig out of one miss; at a hundred-dollar maximum, this one could not. The eight holes were not a risk he was managing. They were an appointment he was keeping.
The holes came
Fourth spin. The ball dropped into one of the eight. Three thousand, gone in the time it takes a croupier to sweep.
He did not flinch, or if he did, he did not show it. Out came another thirty chips. Within two spins that stack was gone too. Somewhere above five thousand down, he started rebuilding with street bets, three numbers at a time, the way a man bails a boat with a coffee cup.
I left before the end. I hope he got some of it back. But I know what the numbers say about the second three thousand: exactly what they said about the first. The wheel did not notice that he was owed anything.
The dangers of high coverage
High-coverage bets are the most seductive thing on the layout, and they are seductive for a reason that has nothing to do with the maths. They win almost every spin. Four hits out of five feels like skill. It feels like a system that works. And that feeling is precisely the danger, because every part of the bet that makes it feel safe is the part that hurts you.
The payoff is upside down. You are risking thirty units to win six. Most bets on the table risk a little to win a lot; this one risks a lot to win a little. One miss undoes five hits, and the miss is due, on average, every fifth spin.
The win rate hides the cost. Three thousand dollars crosses the table every spin, and on a double-zero wheel the house keeps 5.26% of it whether the spin hits or misses. That is about $158 a spin, and it is invisible because the chips come back more often than not. A player who is "up" after three spins has already paid nearly five hundred dollars in edge and felt none of it.
There is no recovery road. A coverage bet is already most of your bankroll on the felt. The usual escape from a loss, pressing the stake, needs room to grow, and at a table maximum or with a normal bankroll that room does not exist. The player above went from "three hits, up eighteen hundred" to "one miss, down twelve hundred" and had nothing to do about it but load up again.
It makes the losses cluster. Because the hit rate is so high, the misses feel like bad luck rather than the schedule. Two misses in three spins, which happens about one time in nine, wipes out ten hits of work. When it arrives it does not feel like the bet's normal behaviour. It feels like an injustice, and injustice is what makes people reload.
It is built to keep you seated. Nothing about a coverage bet tells you when to stop. There is no natural end to a run of hits, and each one makes the next spin feel more like a formality. The bet's greatest strength, the constant small win, is the thing that walks you into the miss.
None of that makes coverage bets unplayable. It makes them the single easiest bet on the table to play wrong.
The lesson, without the sermon
High-coverage bets do exactly what they look like they do: they buy a very high chance of a small win, and pay for it with a small chance of a large loss. On a single-zero wheel the same bet is a shade kinder, 30 of 37 with seven holes, but the shape is identical.
That shape only works one way. Hit and run. Decide before the first chip goes down how many spins you are buying, one or two, and leave on the number whether the ball has been kind or not. Played that way the thirty-number bet is a quick strike with the odds of the strike honestly in your favour. Played until the wheel catches up, it is a loan from the house at a rate you would never sign for on land.
Every one of the man's three hits was a chance to walk. What he was waiting for, I think, was a feeling: the moment it would obviously be time. That moment never announces itself. At the fourth spin the decision was made for him.
Where VIPER comes in
Everything that went wrong at that table went wrong between the ears, not on the felt. The bet was fine. The exit was missing.
That is the problem our software exists for. VIPER's Bankroll Protection takes the decisions you would make calmly in your cabin and enforces them at the table when you no longer feel like keeping them:
- A take-profit that fires on its own. Set "up $600 and out" before the session and VIPER ends it on the first hit. This is the setting most players never touch, and it is the one that would have sent that man to the bar twelve hundred dollars ahead.
- A stop-loss that fires on its own. One miss, session over, no second stack of thirty.
- Stake sized against the bankroll you actually brought, so a single ordinary miss is a dip rather than half your night.
- A spin cap, so "one or two spins" cannot quietly become "until it turns".
If you want to see what a coverage bet really does over enough spins for the eight holes to show up on schedule, build it in the AI Wizard and run it over the free million-spin dataset. The backtest will show you the same run he had, a few thousand times, with the same ending.
A high-coverage system does not shrink the house edge by a cent. On a double-zero wheel every one of those thirty chips is giving up 5.26%, and covering more numbers only spreads the tax across more chips. What coverage buys you is a high hit rate, and a high hit rate is exactly the thing that keeps a player in the chair one spin too long.
Somewhere out there is a man who, for about four minutes, was the most interesting player on the ship. If you ever find yourself with thirty black chips on the felt and a pit boss staring at the dealer, you already know the best move. It is the one that involves standing up.
— The Professor