The Magnificent Madness of Martingale

The Magnificent Madness of Martingale

May 24, 2017

It first came into my consciousness a few years ago.

We were at dinner with two of our friends discussing trips we’ve taken (and would like to take), and naturally Vegas came up. Vegas always comes up when I talk about traveling.

The part that got my attention? He said he had figured out a way to go to Vegas and have gambling pay for the trip.

That sounded awesome. And it also sounded scary.

Of course, I asked him how it’s possible.

Easy, he said. All you have to do is play roulette, and commit to choosing red or black. Then only play that color for as long as you play. Then determine your bet size, keeping it at a fraction of your total playing stake. And then, if you lose, double your bet!

For example, if you have $1,000, you should risk maybe $10 per bet. If you win, you get $10. As long as you win, your money keeps growing. But if you lose–DANGER ALERT!–make the next bet twice as big.

A typical run might go like this. You win $10, win $10, win $10, then lose.

Once you lose, you bet $20 on the next spin.

If you win, you’ve made back the $10 you lost plus made $10 on top of it. It’s as if you never lost at all! Once you’ve made your money back (and a little profit to boot), you go right back to betting your original $10 per spin.

And so it goes.

Examining our example above, we won $30 on three bets and then lost $10. But we made $20 back on our double-down. So now after all those bets, we’re $40 up. That all might take 5-10 minutes.

If we made $40 in 10 minutes, we’d make roughly $240 in an hour. If we play four hours, we might make $960 and that should pay for a suite at the Aria. Mission accomplished!

Our friend said he has tried this a few times and it has always worked.


Is there a catch? Maybe.

Although roulette is almost a 50/50 game, losing streaks still can exist. Red can come up many times in a row, or vice versa.

That would make our betting look like this:

Win $30 after three bets and then lose. Then double our bet to $20 and lose. Then double our bet to $40 and lose. Then double our bet to $80 and lose. Then double our bet to $160 and lose.

You see where this is going.

Instead of comfortably growing our pot, in that scenario we’d find ourselves down $280 with a $320 bet on the table for the next roll. If we lose two more times, we’d be scheduled to bet over $1,000 on the next spin (more than we came to the table with).

Boy, that escalated quickly!

It would seem that a game we have a 47% chance of winning wouldn’t go through prolonged losing streaks, but it happens.

We just don’t know if it will happen today, tomorrow, or never.

The truth is: we could play a long time and never have an account-ruining streak. This type of betting (gamblers call it Martingale) can be really successful a lot of the time, but the losing streak is always out there waiting.

To bet this way in real life, you have to ask yourself one question: Do I feel lucky?

For a long time, I’ve wondered how Martingale would work with trading.

In casino roulette, the House has the edge. That goes without saying.

But in trading, we can have an edge. So does that make a Martingale strategy a smart bet ?

For fun, I coded up the Heron robot with the Martingale betting technique. Here’s what happened.

I first used a small trade size of 0.1 lots. If the trade won, the robot banked the profit and nothing was changed.

If it lost a trade, the robot doubled the trade size to 0.2 lots. If it won the next trade, it went right back to the original 0.1 lots.

If it lost two trades in a row, it doubled the trade size twice (going from 0.1 to 0.2 and then from 0.2 to 0.4). If it lost again, it doubled size again, and so on. Once it had a winner, it went back to the original size of 0.1 lots and started over.

How did it do? Let’s hypothetically compare this style to the regular Heron.

Regular Heron profit: $6,750
Martingale Heron profit: $9,413
Look at that! By using a double-down strategy, we increased our profits by almost 40%!

How about max drawdown?

Regular Heron max drawdown: -$280
Martingale Heron max drawdown: -$1,141
And there it is.

By increasing trade size after losers, there’s no doubt we’ll make more money. But, because losing streaks happen, we also will drastically increase our max drawdown.

I don’t know about you, but it would be frightening to keep doubling down after a series of losers.

So it’s pretty clear the best long-term strategy is to keep our trade size stable for every trade. Yes, we won’t grow as fast, but we also won’t lose as fast.

That doesn’t mean I’m not going to keep testing it, though. Martingale has got me curious, and we’ll talk more about it tomorrow.