Home > General Topics > Trading > If this was futures trading, does it look like I have a (+) expectancy?

If this was futures trading, does it look like I have a (+) expectancy?

  1. Hi all

    Sorry if this is not very related to trading, but I need opinions from traders so I'm posting here.

    The attached picture shows my performance at the blackjack tables, plotted with Excel. Each dot represents my "P/L" at the end of each session.

    This represents 4 months of cardplay and about 300 sessions.
    Net balance is + $12,178

    I use a default bet size of $5. And I increase it when I have a winning streak. The largest bet size was $250 on a very good streak.

    Comments are welcome. I'm hoping to hear from experienced system testers.
  2. I think a cumulative equity chart would give a clearer picture. For example if your equity chart constantly see-sawed between up a few grand and down a few grand and you happened to present your data on an upswing then i would say it would show no edge. But if there is a smooth upward slope with hundreds of data points it would be interesting to see.

    but that is speaking as trader. I know little about blackjack but unless youre counting cards the house has a statistical edge over you and bet sizing wouldnt change that... well thats my understanding.
  3. If this was futures trading and you increase your size based on a "winning streak" (or a "losing streak", for that matter), any positive expectancy inherent in the remaining facets of the system will likely be erased.

    There is a random distribution between wins and losses for any given set of variables that define an edge (Mark Douglas, Trading In The Zone), making "streaks" irrelevant.
  4. In my personal opinion, games of chance are not relevant to forward projections about the possibilities for sustained success trading markets for a living over the long term.

    Now, there are most certainly professional BlackJack players and professional Poker players on this earth. One could argue that "chance" is more applicable to BlackJack than Poker.

    I don't think it is particularly helpful to compare gaming outcomes to trading outcomes because the nature of the winning and losing streaks and the ability for the gambler or trader to make strategic adjustments independent of chance is not comparible in my personal opinion.

    A successful legitimate career trader will have a trading system where the entry signals and position management rules are not critically dependent upon chance. Again, that is just my own personal opinion. In fact, I would argue that position management is much more critical to successful trading than having a positive expectancy per se. You can actually be a profitable trader utilizing a system with a fairly mediocre statistical expectancy rate.

    Every clear thinking soul knows that if you really need absolutes in your life, then trading is not the business for you. Go get a government job. But a good trader will not endure a lengthy losing streak without making some sort of adjustment to his trading system design - or at the very least, sit on his hands and be way more selective. In this respect, I would opine that a trader has a bit more flexibility and room for maneuver compared to a professional gambler. The professional gambler does have some options and he can make some changes - but in my personal opinion the trader has alot more flexibility and tools available and that is the defining difference. Just my 2 cents.
  5. then that would mean nobody should ever add to a position
  6. :confused:

    What does adding to a position have to do with strings of consecutive winners or losers? The trade isn't a winner or a loser until the entire position is closed, no?
  7. I don't see why.

    Suppose i toss a coin where if heads comes up, i win 2x my bet but lose if tails comes up. A bet with positive expectation.

    I bet a few times, have a streak of 3 heads (winners) and decide to double my bet next time. I still have a positive expectation bet.
  8. semantics I suppose. I consider each tick in my favor a winner
  9. It would be interesting and may provide evidence of an edge but not necessarily proof.

    Yes, this is correct.
  10. A (fair and balanced) coin toss isn't a positive expectancy system. It's 50/50. The distribution of heads and tails can be very random; you can, for example have quite a few heads in a row (a "winning streak") but the expectancy is still 50/50.

    Now, let's say the coin is rigged in a way that it will land, on average, heads 60% of the time and tails 40% of the time over a particular number of tosses, and for the sake of illustration we'll say that we need at least 100 tosses (trades) to see the approximate 60/40 result.

    Since the distribution of these wins and losses is random, a streak of heads or a streak of tails is meaningless in the 100-toss scheme of things. If you're the casino who wants to take advantage of this excellent positive expectancy "edge", you have to play every toss (take every trade) with an equal bet. If you believe that a streak of heads or tails somehow means the odds of the game have shifted away from 60/40, and decide to abstain from some of the tosses or bet larger on some of them, you become the gambler who does not have an edge in the casino unless by cheating.
  11. the question is, if the rigged coin comes up 60 tails in a row would you increase your bet on heads?
  12. are you doing a reverse dalembert?

    if you are when you lose do u reset to $5? or go down by 1 bet size 45 40 35 30 ?
  13. You can flip that coin for years and years, and all the experience in the world, the odds in your favor will never be more than 50/50, whereas in trading, in time, one learns how to trade and decrease losses, backtests and finds out what the median wins/losses are so as when to increase or decrease size. I would believe it to be impossible for me to increase bet size based on median for betting heads/tails.

    Actually, I believe it is 49.5/49.5/1 as the one is when it lands on the edge.
  14. for those of us that are trend traders, trading is a lot more like playing a slot machine, because occcassionaly we hit a jackpot
  15. When I lose after upsizing, I half my size on the next bet. 2 losses in a row and I'm back to $5.

    This is an arbitrary rule I enacted without technical research. Seems to work for me.
  16. you're betting on a trend in negative expectency game.

    to make it more like futures, I would be betting on what I think you are going to do
  17. Sorry, NoDoji, you need to reread what i wrote. I said you win 2x your bet if it lands on heads, which makes the wager a positive expectation one.

    It doesn't matter what the wager amount is, whether you increase or decrease after a win or not, it remains positive.
  18. True story: I played online BJ last month. My balance went from £800 to £1400. I was flat betting £10 per hand and did approx 3000 hands.

    Does this mean i have an edge?

    Of course not, I was lucky!!! If i kept on playing, ultimately I would lose, guaranteed.
  19. In this case bet size does not matter as every bet carries a positive expectation.

    There is no way to achieve a positive expectation in BlackJack by varying bet size alone. If a player's game does not carry a positive expectation then increasing bet size either with winners or losers will not achieve a positive expectation.

    The only way to achieve a positive expectation in BlackJack is by counting cards. It is true that while basic blackjack strategy will get a player very close to a positive expectation - it is still negative without some knowledge of the makeup of the remaining cards in the deck/shoe (achieved through counting) and nothing that you do with bet size will change that.

  20. hmmm looks like many dont know what expectancy is about.

    say, in blackjack, if you're NOT counting cards, you've negative expectancy, while the house has positive expectancy.

    if you're counting cards, then, you reverse the roles, thus the player reading cards has positive expectancy and the house has negative expectancy.

    there's nothing else that can do to turn the odds around for the player than counting cards. obviously, no casino ever allow card counting, therefore, in normal circumstances, you always have negative expectancy in blackjack.

    any betting system you use will never revert the negative expectancy, unless you'd be allowed to, for example, increase the bet geometrically every time you lose, until you finally win, but that's not allowed in any casino and even if it would, you'd have to have an infinite bankroll to play.

    now, you can play blackjack lots of time without counting cards and still make money. why is that possible, if you have negative expectancy? well, it is because you got lucky.

    so, you can make lots of money in a negative expectancy game or in a positive expectancy game. the main difference is while in a negative expectancy you only make money if you get lucky, in a positive expectancy game, you'll make money no matter what, it is a question of playing it enough times until the law of large numbers kicks in.

    in trading, the above principles apply like a glove. find yourself a way of counting cards in trading, and you'll be long term winner no matter what happens in short term.
  21. you sure about that? With a rigged coin that on a 100 tosses flips heads 60 vs tails 40, you sure there is no advantage to bet sizing? Many mathmaticians have tried to explain this to me, but it defies my simple common sense.

    especially like I said, when it comes up tails 40 times in a row.
  22. Getting an exploting an edge at gambling is unfathomably easier than doing the same in trading.

    Any idiot can put out a slot machine with a Negative EV pay structure and make consistent money as long as he has players and a bankroll.

    Any highly intelligent, quick thinking, well capitalized black jack player can create scenarios, depending on house rules, that are considerably +EV.

    Very well connected Sports Bettors with good consultants and the ability to shop for lines and move real volume are so devastating to LV sportsbooks that the casinos have passed laws making it harder for them to get their plays down. Just try making a 2-4 dime play on a NCAA game in Vegas, if you are considered the least bit sharp. If you spread it out over several casinos, they just start calling each other and you are chasing the number before you know it. Basically, if they let you move real volume in LV, over 10K college hoops/football, over 25K NFL, it is because you are a sucker and they know it. In fact, when I worked for some of these groups, we used to try to find/get friendly with bigger gamblers who played pit games, so that we could have them get down action for us.

    In comparison to trading, where the distributions are infinite, gambling is child's play.
  23. Your simple common sense doesn't make sense statistically.

    The issue is - "Can bet sizing change a negative expectation to a positive expectation?" and the answer is NO.

    As to 40 in a row - the statistical term is clumping and is a pehenomenon that can be seen in tests that are not of adequate sample size.

    Even with 40 winning hands in a row in a BJ game w/o counting you will still wind up a loser once you have played enough hands to normalize the data.

    It is just like at the wheel or the crap table - nobody ever makes a long term living at either because every bet carries a negative expectation and no manner of bet sizing will change that.

    The logic is in the posts below. As your simple common sense is not getting the job done, maybe you should do the math instead.


  24. yeah, you still don't get it. The coin is rigged, it comes up heads 60 times out of 100.
  25. So what? that probability doesn't change every time you toss a coin. There's an optimal amount you are supposed to bet everything time, unvaryingly. This is one of the easiest problem to solve in statistics beyond 101-level.

    Why do you think your 'common-sense' is better than an actual proveable mathematically result?

  26. because most traders are not so sure about their 60% "edge" and don't bet the optimum amount hoping their backtested statistics prove true.
  27. at anyrate, you all bet on your coin tosses, I'm about tired of it all, argue all you want, I'll bet on whether I think you are going to bet on heads or tails

    and I'll adjust my bet size depending on how right or how wrong I was

  28. But there is no such thing as a "trend" in a series of outcomes in a card game. We gamblers call them "streaks" and many believe that they can ride on them.

    It is a negative expectancy game in theory. However, in the real world, some gamblers can just defy the odds consistently.
  29. In that particular case - the greater the bet size, the greater the win. The is a manifestation of a positive expectation.

    The OP was about using bet size alone to convert a negative expectation to a positive expectation - which is, of course, impossible.

  30. ... we are talking about coin tosses, not trading, right?

    Are you shifting the subject because you realize that you can't really defend your position anymore?

    You specifically talked about a coin toss with 60% advantage. Not traders wondering if there's an edge upon seeing a 60% win rate.

    Are you yet another dishonest and scummy poster who has no integrity whatsoever?

  31. yes, I have no integrity whatsoever. Go back and get an education and read the original post. First of all, even the coin tosses can't be discussed intelligently since one put up the idea of a rigged coin that flipped heads 60% of the time. And everyone who has no reading comprehension just started spewing forth everything they know about coin tosses. And then small minded posters couldn't understand if we were talking about Black Jack or trading, even though that is what opie asked about. So it is a convulted mess which attracts many to state the long known mathmatical truth that they think is so new, even though it has been known since the invention of the game (until they invented card counting) that you can't beat the dealer. Even if you adjust bet size. But that hardly addresses the question of how that relates to futures trading, which if you had read the OP was the original question.

  32. blackjack or any other game of that sort or trading is pretty much the same, as far as expectation is concerned, that is, negative expectation.

    betting strategies don't alter the expectation of the game. the only way to make money in the long run, is by reverting the built in negative expectancy, which in blackjack is only possible, when counting cards is a viable option.

    in trading, reverting negative expectation is hard but has been done many times through i.e. inside information, arbitrage, or systematic trading. again, betting strategies in trading don't alter the expectancy, but can change the smoothness of the equity curve.
  33. You are just rambling now... you are the one who stated that you don't believe mathematicians because they disagree with your common sense.

    A biased toss coin is the simplest device of testing simple statistical hypothesis. You failed to understand even that.

    Now you are posting barely comprehensible blobs trying to save yourself from looking foolish.

    You have failed, oldtime.

  34. Careful... this is only true if the game exhibits indepdence between plays. If it doesn't (like the market), then betting strategies will change the expectation.


  35. in trading, market occurrences are fully independent from each other, which, i know, is quite the opposite the common perception.

    the only visible, sustainable, non-random pattern is the autoregressive nature of volatility, which cannot be used in a betting system to maximize positive expectancy, because it tells us nothing about the next event either.
  36. yeah, I think we already covered that about 200 years ago
  37. Are you... joking? Almost all market prices exhibit autocorrelation at multiple time frames, especially as you decrease the time frame.

    Are you seriously suggesting this the SPX's price at 9:05 is generated from a process such that it's independent from the 9:10 price?

    Again, seriously?

  38. first of all, if you had ever learned to read, you would know I never said I don't believe them (why do I even waste my time on you) I said I don't understand it.

    and secondly, for the last time, if you haven't kept up because it is too hard to read simple ET internet posts, we are not talking about a fair coin toss like in the Superbowl, we are talking about a rigged coin which flips heads about 60% of the time.

    Might I suggest the Sylvan Learniing Centers, since it seems public education has failed you when it comes to reading comprehension.

  39. no autocorrelation whatsoever, especially in shorter time frames.

    in wider time frames, i.e. multi year, you can see autoregression due to known externalities, such as inflation, quantitative easing, survivorship bias, etc.

    looks like you still have a lot to learn.
  40. Um.... hence, I said multiple horizons. Shorter time-frame == tick level - go ahead and run the autocorrelation between last traded price for IBM in the last 3 days... tell me what it is, because I see around ~70%.

    And, as you say, on the longer time frame as well.

    And no, survivorship bias doesn't induce autocorrelation in the time series of price.....

    You don't appear to know what you are talking about.

  41. You don't understand... and you are unable to understand the math? Is that what you are staying? Just to be sure?

    FInally, if you notice, I too am talking about a freaking rigged coin toss.... you are the one keep pulling the conversation away. A rigged coin toss is still freaking independent between tosses.... <- yes or no?

    Why is ET full of crank-worthy old man like Jack Hershey, Joe Doeks, and you....


  42. market occurrences are fully independent. what IBM stock price has done in the last three days is irrelevant. i am not going to discuss statistical irrelevant samples. just come back with a sample of +5000 price points and let me know how you uncovered price dependency there? then we'll use you mathematical formulation to predict prices in the next 5000 price points and if there's positive expectancy, i will mention you for nobel in economics.
  43. Dude... having autocorrelation or serial dependence doesn't imply predictability - see ar(1) processes for example.

    Also, IBM stock prices at very short horizons (or any stock prices at tick level) IS indeed autocorrelated in the long run. Bid/offer effect at play.

    Again, you don't appear to know what you are talking about. Please stop.

    Finally, you keep using this odd term 'market occurrences'. I get it: you use it so that when you get catch out, you can always claim 'price' or 'ticks' or whatever is not what you mean by 'market occurences' like you've already excluded volatility. Want to define it for us here?

    (I didn't forget: did I catch you out using words you don't understand? You didn't respond to the fact that I pointed out, contrary to your claim, survivorship bias doesn't create autocorrelation in prices).

  44. yes, and the question remains, if it trends more than expected for you should you reduce size? or if it it trends more than expected against you should you increase size?

    of course this is of no concern to you youngsters, because you always bet optimum size. After all, it's not like you are betting your life savings. If you're wrong you can always make it up later.
  45. Hold on... we are talking about independent tosses right?... what does it matter if it 'trends' or not???

    Or are you guy who bets large at the roulette wheel when a number becomes 'hot'? Am I to understand that you do?


  46. i now realize you're the typical multi alias ET troll that don't have anything else to do other than hijacking threads and disinform the readers.

    keep on with the good work.
  47. What are you talking about? Who am I?

    I think you are just really mad that it's clear you have no idea what you are talking about when it comes to technical subjects. So you resort to ad hominem attacks.

    Why don't you point out why I'm wrong? My statements are factual and emperical. Yours are baseless.

    Are AR(1)s forecastable because they are serially dependant? Show me why I'm wrong - rather just throwing around accusations?

    Typically ET trash.

  48. I'm the guy that bets at the roulette wheel or black jack or craps a constant bet with a limit and either has a lucky night or loses my limit and enjoys an evening of entertainment.

    when it comes to trading I know when I am winning more than I should, and that is when I reduce size. I don't want to be George Soros, I'm happy just being profitable. I already risked it all. It was fun and I'm glad that I won, but I don't want to bet I can do it again.

    And when it is going much worse than any reasonable man could expect, I increase size. If not size then time.
  49. the difference is, at the casino I am betting against math. In trading I am betting against you.
  50. Christ... we are talking about a biased coin toss (that you proposed)... again, you change the subject like a slippery flish.

    Your views are right if we are talking about trading. It's wrong if we are talking about a biased coin toss.

    Is this fair?

  51. well, you're starting to make some progress.

    I'm not the one making the point. I'm the one asking the question.

    again the question remains

    "On a biased coin toss, can you enhance returns with bet sizing?"

    The mathmaticians tell me no. But they don't bet their own money and hold back, just in case they are wrong.
  52. And in a circle we go.

    Okay okay, I get it, you are just asking questions.

    I'm beginning to get the image that you are just a lonely old man wanting some company... some human contact. You don't care about the answers, you just want someone to reply to you - ET is like a nursing home social room for you.

    I get it. I'll stop.

  53. ok, now you finally got it. I tried to tell you that, but you never listen.

    why the hell else would I post on ET unless I was losing my ass in some eur.usd and just wanted to talk to someone? It's lonely down here.

    But getting back to opie, the difference is between betting on a deck of cards and betting on a trader.

    but in the meantime it's fun to talk about math

    especially as to how it applies to traders
  54. at anyrate, getting back to the subject, I would say based on backtesting your system, the way to go would be play 3000 hands, and if you are up, quit

    if you are down double up and play with 20 for another 3000 hands

    continue and repeat until after 3000 hands you are finally up

    then quit
  55. This is not like a casino. This is investing money to get profit. If some one invest their money in fools' way, That person will lost their money. Do not think this is gambling. When you doing gambling, You do not know what is the result. But this trading, We think about future tradings and We guest it using data , experience , and wisdom.
  56. There are no "streaks" in trading.

    In Blackjack, you can sort of guess the odds based on cards that have previous been played before the deck is shuffled again. In other words, each hand is not independent.

    In trading, every trade is independent. Just because the contract you are trading has done something does not mean it is more or less likely to do the same thing in the future. Every trade is independent. So increasing your bet while you are on a "streak" is risky. Trading is more like roulette than Blackjack.

    At least that's my opinion.

    You may have some system which gives you a better than 50% chance of winning and can adjust your position size accordingly.

    Good luck!
  57. You are incorrect.

    we have no incentive to correct you.
  58. This coming from a man who writes hundreds of pages of complete nonsense....

    You've made my day, old man.


  59. I have a positive expectation that you do not understand me.
  60. I have doubts that you know what 'expectation' actually is.

  61. +1. each trade is an independent event. The fact that 7/10 trades with the same setup were successful does not mean the setup has a 70% success rate. Douglas talks about the "just knowing" factor and I couldn't agree with him more. There's definitely a scientific element to trading but the thing that determines profitability is talent. Are you the Miguel Cabrera of trading? In other words, you need to have some inherent talent to succeed.

  62. I have played in Vegas.

    I count cards.

    The house "knows" I count cards.

    Usually I use an ice bucket from my room.

    If friends are with me I pass chips (silver dollars in my case) to them as theirs become exhausted.

    As time passes, the house does several things:

    1. very strong free drinks.

    2. dealer changes (they get more frequent as well).

    3. They ask me to dine in a location away from the tables at their expense.

    It all seems fairly polite because I am a small fry.

    One common occurance is the displeasure of other players at the table. Before the other players figure out to change tables (their usual idea) they speak to me about how I am wrecking their playing. The house does not like this going on.

    30 or 40 years later, I noticed the BJ is played differently. they use multidecks, I didn't notice too much shuffling, and dealing is much faster. From counting, I guessed four decks were in use. I didn't notice that much had changed basically. I was slower due to lack of practice and probably my mind was slower as well.
  65. Since we are talking about Vegas, here is how to get a free drink and make some money.

    Go to one of the casino bars, ask how much money, you have to put in the machine for a free drink. It was $ 10 last time I was there. Play a game with around 50% win percentage like black jack. I won, and got a free drink.
  66. This is called Martingale (similar)- it does not work, even if you stay conservative on bet sizes.
  67. Bet sizing defintely works. However, it is not based on winning or losing.

    Reasonably, with unshuffled cards between hands, it is based on the favorability of the deck. "Bet the Dealer" explains this and the book also provites small "cheat sheet" type cards of more or less sophisticated strategies. "5's theory" is a very good starting point for gaining an understanding of bet sizing.

    Someone offered a bet to me. It may have been too time consumming for me. The time to realize the bet proceeds were much slower than trading to make money.
  69. LOL

    I have been trading for 55 years.

    1000 hands is not long but the 100,000 is a small number.
  70. I beg to differ from the largely negative replies to the OP question.

    As is known, at blackjack the expectancy varies with the composition of the cards in the deck (if I recall correctly my own studies as a kid, it's the ratio of sixies to figures and aces)

    By increasing bet size when winning, the player may be exploiting the transitory positive expectancy phases that occur; basically using the outcomes to conduct discovery on the current expectancy, under the assumption that when one is winning it is because the odds are temporarily in his favor. So, like indirectly counting cards.

    Having said that, the edge must be tiny and I suspect it's lower than cards counting.