A path towards profitability

Discussion in 'Professional Trading' started by garachen, Feb 7, 2012.

  1. garachen


    Trading is a negative sum game. Poker is a negative some game. Starcraft 2 is a zero sum game.

    It is interesting that people will use the 'zero sum' argument in order to show that traders can't be consistently profitable. But I can prove that any South Korean over the age of 5 can beat me at Starcraft (even though it's a zero sum game).

    So, while there are lots of opinions and conjecture out there I would like to offer mine and tell a little bit of my story.

    I did a BS in Mathematics and a Masters in Mathematical Finance.

    I worked for a bank, hedge fund, mutual fund and trading company before going out on my own.

    I have once lost my entire life savings - not due to trading but due to broker fraud.

    I have taught several people to trade. They all work for me. They still work for me. Nobody has ever quit or stolen from me.

    Now, I'm not certain that everyone can be taught to trade. Neither do I buy in too much to the overly hyped theme of hard work, persistence, grind it to death attitude that seems so espoused. I've seen many of those fail. Nor do I agree with the 'you need to have what it takes' attitude either. Somehow suggesting that one can be consistently profitable by pure force of will. Neither can you be a dilettante.

    The key things I look for when hiring a new trader are these:
    (roughly in order of importance)

    1) Ability to take risk: avoid people with excessive debt or obligations who are desperate

    2) Obedience: they have to be humble enough to obey without question. I rarely (1-2 times a year) have an opinion about anyone's position but when I do I want them out. Immediately.

    3) Intuition. I measure this by playing certain board games that require intuition. I watch if they can internalize a long list of rules and play rationally.

    4) Calmness: You swear. You're gone. You scream yell and hit things. Gone. You need to be calm to do the right thing. When you are upset you can no longer accurately assess probabilities which is the heart of trading.

    In my early years I had a pretty bad day loosing 100K+ of my own money. I was in a room with other traders and the guy next to me was throwing a tantrum about his $300 loss while I sat calmly eating my yogurt. I don't need to be around people who don't have self control.

    If you have those abilities, I believe you can learn to trade - at least the way I trade. Note what is missing. Math. Education. Persistence. Experience. Yes, I have these, but they are not essential - or even important.

    I'm completely aware that this is my own opinion for my own path and that others have achieved success in other ways.

    The first thing to do is watch a contract. Focused. Every day or night for the same hours every time. Two different contracts is probably too much unless the second provides additional information to your main contract of study. Watch the market depth. How it moves. When/why it moves. Try to gain intuition to determine if there is selling pressure or buying pressure. Charts can distract you. Don't use them at this time (or maybe ever). My most consistent trader never looks at a chart.

    This is all best done with a contract that is not super news sensitive and where there are a good proportion of manual traders vs algo traders.

    After 2 weeks of just watching for several hours a day start trading with a 1 lot. Then we do some training on following intuition (what you *know* is going to happen) and not dreams (what you *want* to happen) while the trade is going bad. This is what differentiates a mediocre trader from an amazing trader who can consistently make several million a year. How fast I can let go of my 'dream', control my ego and follow my intuition is the single area of constant focus throughout my career as a manual trader. It will have the single largest impact on profitability.

    After 1 month a trader will be profitable. Consistently. Maybe not every day but certainly every week. Growing that profit from an initial $500 - $1000 a day to $5000-$10,000/day takes about a year and is a very incremental building process. They need to fear losses. And losses need to be commensurate with your goal. If your goal is 10K/day a loss of 10-12K is acceptable. If your goal is 1K/day that loss is not acceptable. So, a gradual increase in dollar value loss tolerance MUST be accompanied by a strong fear of disproportionate losses. If it is not. You will blow up - badly. Maybe you can recover. Maybe not.

    I was going to write more, but I can't remember what...

    No doubt I'm going to take some grief/hate from the community here. I don't much care. But I would be curious if anyone were to follow my method to fruition.
    userque and Raphael like this.
  2. it is an excellent thread. My impression is that you are an algorithm trader since you care about a few milliseconds.

    I almost never traded manually. I am a system trader, develop and back test, then let the robot trade.

    It is interesting that you do not look at chart. then how long do you hold your position?

    I heard lots of wall street professionals do not look at charts either, is that true?

    besides, do you know if james simons medallion trade on chart or something else? One of my friends doubt James Simons' model is based on chart. I have no idea.

  3. sidm



    I like your post. A couple of questions.

    (1) What are the board games that you use to test intuition?

    (2) What contracts do you suggest to start studying in the beginning?
    Can you give some examples? You said don't look at charts. Are you saying to focus strictly on analytics that can be represented as numbers, parameters etc.? Why the aversion to the visual cue?

    Some observations.

    (1) You use the world "intuition". But I think what you are really implying is deliberate, calm, rational thinking. I know this just semantics and my interpretation of "intuition" is different from yours. But as a suggestion, much more precise terminology is available in books like Thinking Fast and Slow by Daniel Kahnemann.

    (2) You said math, education etc. are not needed. But surely they are kinda needed as they allowed you to get on to that career path, earn a high salary and thus build up a capital base to start trading on your own.

    Overall great post for newbies like me trying to learn.

  4. garachen


    I started out as pure Algo. Probably went 1.5 years without manually touching anything. But then I started to notice patterns that I couldn't easily code against. And I started trading some size in illiquid markets where I could push other manual traders into panic. Now, Algo trading accounts for about 25% of profits. Much of that occurs in 5-10 seconds scattered throughout the day. There are some seconds where I will make 200 distinct trades. You just pray you're handling context switching, garbage collection and journaling to disk in some efficient manner and nothing breaks down.

    I have not met an algo trader who looks at charts. I'm sure they probably exist. The main chart lookers are point and click spreaders. Guys trying to trade the bond basis or Eurodollars through a spreader. Come to think of it - most of the ED guys I know look at spread matrices and not charts either.
  5. that is possibly because quantifying candlestick is extremely hard. I do not know anyone successful at it. but I met lots of day traders who manually trade based on chart.

  6. you lost me at "I did a BS in Mathematics....."
  7. garachen


    I got lost reading this textbook on machine learning today. We all get lost sometimes.
  8. Thanks for your post. This kind of insight is very useful!

    My thoughts on this are likely not as useful but I'd like to share my beginning inklings of 'turning the corner' in trading. I can see similar stages in my trading except it is over a longer time period, probably fewer trades per day, and with smaller size, and not as consistent.

    For consistent weekly gains, how many trades are your traders making per day?

    My average hold time is 31 minutes, max is up to 4-5 hours - all intraday. Annualized sharpe ratio of 2.08 - nothing to write home about yet. But, there is a point in time when a few things changed for me for the better about 6 months ago.

    I went back and forth between algo and discretionary in fits and starts. Like waffling between trying to be a garage rocket scientist or to just try and be a natural through sheer will. Several mini-blowups. I decided to do the following:

    Trade 1 contract of ES per $10,000 in equity
    Not risk more than 2-3% of equity on any one trade

    At that point I started being more confident and it became easier to both exit with a loss and to enter another trade later - even after a drawdown that lasts weeks.

    As you indicated you experienced, there are parts of it I am not sure how I'd automate. This is in one sense uncomfortable, but in another way it is comforting..

    If I had to describe the _strategy_, all I can come up with is some general ideas or principles. I try to enter at value areas (looking at volume profile) and feel that helps me not get whipped out as often. I put stops outside of value, because I figure at that point the trade is a bust. I like to be in the direction of either a gap that price can zip through to another value area, or at a week or multi-week high/low where there is open air price can move through. I like to be with the sentiment/trend -but the definition of what that is has become a bit fuzzy. I can see an intraday downtrend as price getting sucked back to a value area and perhaps a buying opportunity. My P&L seems to follow a similar pattern to trend following: more small to medium losses and some larger wins that make the money.

    Much I need to improve but at least I feel like I'm in the game in a more sustainable way and my equity curve is roughly slanted in the right direction. Any suggestions on how to approach this continual improvement or insight on my style would be welcome.

    Continual improvement is an interesting problem, because it's not clear that you are screwing up when you are making some money.
  9. garachen


    Good questions!
    Board games:
    For some reason I don't want to say directly or make it searchable or too obvious. The game is German. There are lots or rules. 30 or so which I rattle off in a monotone voice. Winning requires correctly assessing several overlapping levels of probability withing a resource constrained environment and also predicting the other players choices based on their constraints. I'm not looking for them to win but like to observe how they play and if given just that brief exposure they are able to pursue a reasonable strategy.

    visual aids:
    I know this goes against the grains for many people. But when you walk into one of the large trading companies in chicago what do you see? Lots of charts full of indicators. No. Maybe one chart. Every bit of screen space for multiple monitors is crammed full of market depth traders. Prop trader groups are maybe different. Never felt the need to visit one. I feel like for me using an indicator would give me a false sense of security and I would fall into the trap of thinking trades don't work because my indicators need adjusting. I would think they would cloud the intuitive 'feel' of price action gained from looking at the market.

    I gave some thought to your suggestions and I might be at a disadvantage here because I have not read that book. I think I'll stick with intuition for now because in my mind it implies a lack of deliberation. I'm looking for the quality of being able to feel out the right answer given incomplete information. The best examples I can think of are from Abstract Algebra which is off topic. Here's another one. [Ug, I can't think of a brief way to sum this up right now...skip]

    Well, kind of. I admit that I only seem to hire from top schools because that's where I advertise. Major is not important but I'd give negative points for economics or accounting. I'm looking for smart people who have not been 'trained' by someone else before. But I don't think this is necessary for someone trading on their own.
    Large capital base in a strict sense is not really needed. Though our definition of large might be different. I'd think 50K, no debt and no need to touch trading income for a year is sufficient. It CAN be done on less capital but not on a retail platform.
  10. I am interested in your observation on visual aid. if wall street does not trade based on chart, instead on market depth, how could they hold their position for more than say 15 min?
    #10     Feb 8, 2012