Episode 196: The Magical Potion of Probability and Why Trading Is Gambling

Returning again after their Cigar Shop debut in Episode 179, Shonn Campbell and Matt LaCoco host The Traders Podcast again to talk about the magical potion of probability and to what extent trading is, in fact, gambling. Matt and Shonn discuss how traders are just in denial about being gamblers, asserting that traders aren’t much different from professional poker players or card counters who use statistics to make their money. Shonn and Matt suggest that trading is merely card counting — with candles! They also talk about “the gut feeling” and how it’s essentially the intuition of probability.

So, in Episode 196, the Cigar Shop guys will answer the questions: Is there a way to count cards in trading? And is there a way to gamble for a living? You definitely don’t want to miss this one! Thanks for listening to The Traders Podcast.

Links for this episode:

Follow Matt on Twitter: @MattLaCoco

Follow Shonn on Twitter: @PlateauTrader


Rob on Twitter: @RobBooker

The Traders Podcast on Twitter: @TradersPodcast

E-mail us! Producer@TradersPodcast.com

Trader Interviews.com

11 comments on Episode 196: The Magical Potion of Probability and Why Trading Is Gambling

  1. Darth Vader says:

    hey this is cool. I actually have an indicator that counts candles. My favorite is a 7 bar fade but sometimes it wants to turn into an 8 bar fade then a 9 or 10 bar fade but then sometimes it hits… but then you cant take profit because it hit your area of profit while you were taking out the trash and you forgot to set a take profit order lol love this episode!

    1. Shonn says:


      would you ever consider sending that indicator to us?


  2. Serge says:


    Thanks for a great podcast! Interesting to see hear different thoughts being spoken about, keep up the good work.
    Now, it sounds like you guys have a little technical background and hence you may be able to use some of the following tools I have been fooling around with.
    For a lot of the analysis that I do I use RapidMiner (free :)) to mine a lot of data and to visualize the results.
    I think, specially as you are looking at massive amounts of data that pre-programmed drag and drop blocks are an enormous help!
    In addition, there is a very comprehensive video training set of resources (including a guy running through a predicting the market example.

    Thanks again!

  3. Matt LaCoco says:

    Wow! Thanks Serge! I’ll definitely be checking that out!

  4. Chris says:

    Great show guys.

    I’d just like to point out that the doubling up roulette strategy that was mentioned is known as the Martingale system (see: http://en.wikipedia.org/wiki/Martingale_(betting_system)). The odds of you losing 10 times in a row is higher than you think and once that happens you’re broke – not a good idea!

    1. Shonn says:


      thanks for listening!

      I completely agree with you that it is a bad idea when probability is not in your favor as at a casino. I think it is being used in the case of roulette to actually combat starting from a statistical disadvantage. But what if you could find a probable edge in a game, would that make it more interesting to employ? I would love to hear some thoughts on this!


      1. Chris says:

        This got me thinking so I wrote a small java program to get an idea of how many losing trades in a row you can expect for a given win/loss ratio. I ran the test using randomly generated numbers for 1 million times for each win rate.

        Here are the results:
        Win rate MaxConsecutiveLosses
        50% 21
        60% 17
        70% 15
        80% 12
        90% 7

        So, if your strategy has a win rate of 50% (you win half the time and lose half the time), you can expect at most 21 losses in a row. Similarly if the win rate is 90%, your expected maximum string of losses is 7. I ran the test a few times and the numbers varied slightly (usually 1 more or less than above).

        How good/bad are these numbers? Well lets say our starting trade risks $1 (stop loss & lot size set so that if it goes against us, our account is down by $1). Martingale strategy says we have to double up on each loss to make our initial risk amount ($1). So assuming our first trade is stopped out, we have to risk $2 on our next trade. If that bombs out, we’re betting is $4, etc.

        Here’s a table of the number of consecutive losses and the money that has to be put down:

        Loss Trade-risk
        1 $1
        2 $2
        3 $4
        4 $8
        5 $16
        6 $32
        7 $64
        8 $128
        9 $256
        10 $512
        11 $1,024
        12 $2,048
        13 $4,096
        14 $8,192
        15 $16,384
        16 $32,768
        17 $65,536
        18 $131,072
        19 $262,144
        20 $524,288
        21 $1,048,576

        So if your win rate is 50%, we have to be prepared to ensure 21 losses in a row, so to be assured of making $1, we looking at trading over $1m on the last trade and would need an account size of over $2m.

        The 90% win rate scenario is much better with an expected worst case of 7 losses in a row, you’d need $64 dollars for the 7th trade with an account size of $128 (ignoring margin requirements).

        A win rate of 80% is a lot worse but still manageable. Anything less than 80% is not feasible.

        So, the question is: is there some correlation in your data that you can exploit that will give you a 90% win rate? Considering you’re forced to double up on each consecutive loss, this would be a highly stressful strategy.

        Your thoughts?

        1. Shonn says:


          Love this information. Very well said.

          here are my two thoughts. The martingale system is used in roulette and roulette has two very distinct disadvantages to trading:
          1. Roulette, and most forms of gambling, are almost 100% random event games. Trading is a market place that is made up of people. From my blog “coins an dice have no memory, but people do.”
          2. Change in profit. The real disadvantage to this comes in when the risk to reward is 1:1. $1 played, $1 payed. How does it calculate with a 3:1 or 4:1?

          Play this scenario out for me and I am willing to be wrong on this totally. I love this conversation.

          Let’s say my money management strategy was a condensed martingale, sort of like Lacoco’s 3.2.1. So i give myself only 3 losses per day but I have to double the risk every trade. NOT the pip stop, but the risk so .1 lots to .2 and .2 to .4 lets say. This little spreadsheet shows that if those trades can be 3:1 reward to risk or better, this could give me an edge.


          N0w, If I suck as a trader, I will lose 70 pips every day. But that will happen regardless of if I employ this system or not.

          Is this something that can be put on a robot? I couldn’t trust it. It would stress me out I agree. But can a discretionary trader use this method to make money?

          Again, thanks for the conversation.


          1. Shonn says:

            $70 dollars a day, not 70 pips.

  5. Scott says:

    Interesting topic, but this podcast was terrible. The probability premise no doubt has some merit, but the podcast itself sounded like a couple of random dudes getting stoned in their back yard and trying to carry on a conversation about trading. A lot of redundancy, a lot of overly optimistic fantasizing, and a lot of self congratulating, all over an entirely untested hypothesis.

  6. Matt LaCoco says:

    Wow this is a great conversation. Darth, please join us at our next recording. Come alone though, no Storm Troopers.

    First let me be clear. You’ll never see or hear about me martingaling my account into oblivion, because that’s exactly where it goes. Now I’m sure that there’s a way to arrive at a mathematical advantage to a Martingale approach, but to me it’s a mathematical illusion. As Darth points out, your starting trade size must be microscopic relative to your account balance in order to “ride out” the losers. So if you’ve got a $500k and you’re trading a microlot just so you can martingale when it loses, you’re less efficient PnL-wise than a $10k account fixed at 1 lot (or even a half) trades. My friends, I’m not interested in a 50 year trade plan.

    It’s fun to think about Martingale, and I think the “spirit” of the concept is certainly valid. However, basing trade size on the performance of the last trade just doesn’t sit well with me. Instead, I prefer to size my trades on a probability score. Perhaps this involves doubling the size, relative to the last trade taken, but only if this trade is twice as probable to succeed.

    I believe the root of the conversation, lies with an ongoing conversation Shonn and I have been having regarding “Risk”, and what exactly defines it. I’d be sacking groceries, or wearing an orange apron at the Home Depot if my trade sizes were some single digit percentage of my account. Remember I’m the guy that runs for the hills with 6 pips of profit. As such, those 6 better count for something measurable.

    When I am considering what size to trade, I ask my self things like:
    How long do I want to be in this trade?
    How many times have I seen this setup before?
    Is price on the “edge” of something?
    What the hell time is it anyway?
    Is there any news coming up that will impact this trade?
    Is that Bacon?

    But I never ask if my last trade was a success or not.

    In summation, I believe Martingale is a good way to quit trading fast. I am naturally quite risk averse. However, I do think most traders would benefit from drastically increasing the size they trade relative to their accounts. But I have a funny way of looking at things like that.

Leave a Reply

Your email address will not be published. Required fields are marked *