@garachen: How do you define the term convexity? Can you please elaborate? Just interested in your thoughts. Regards, Monoid.
I guess this merits explanation. The concept is simple but the implementations are varied and complex. I'll hit it from a few angles. Convexity is positive gamma. An example is the upward curvature of an option payout function. When you buy an option you are always long convexity and short theta (time). You lose money every day that you are supposed to be able to pay for by extracting value from the gamma by delta hedging. Ok. That's technical but should make sense to an option trader. Here's another way. Imagine that everyday there is a product with this payout: if ES is down at the close you lose 8c. If ES is up you make 10c. Assume ES follows a random walk. How much should this product cost? 1c. Trading isn't really about guessing overall market direction, it's about synthetically constructing scenarios where you can buy the above payout structure for 0c or less. There are many different ways to apply this but they should all force you to closely examine the reason for why you are getting paid. Manual traders should be expert at doing this. People can spot 'movement dislocations' across products much easier than a computer can. It's too much a priori knowledge to program into a trading strategy. Charts: I know I've voiced my opinion on this before. Charts have the following problems. 1) They cause your brain to see patterns that aren't really there and that have no predictive power. 2) They cause people to start to personify the product they are trading. Saying it 'wants' to go up or down. This distorts your perception and corrupts your intuition. 3) People start to get overconfident in their trading by looking at a chart and thinking in terms of 'good setup' or 'great setup'. Then they take too much risk. I'm not saying it's impossible to trade off charts. So far, the academic literature has shown that chart reading with standard indicators when applied systematically does not provide any value. But academics don't really know or care about the depth of the problem space and make LOTS of simplifying assumptions that have meaningful impact. For example, they are not going to want to account for the changing relationship dynamics between different products or news events or anything else that is messy and subjective. But this is my standard. Suppose you are a decently paid engineer making $150k. Given the risk of trading and the disruption of your career and the higher incremental taxes, you should be making at least 2x trading than you were in your job. That's $300k. And you've got to get to that number rather quickly. I'd suggest that's a very tall order for a chart reader. Almost nothing I do and nothing I care about watching is going to show up on a chart.
But what happens Forty draws...when your techinque is defined stop+ 65% defined entry + undefined exit? I do what you say ( take the average's calculation)...but do you think the "edge" calulation is still correct?...or is it just "close enough"???
@garachen: Thank you for elaborating and sharing your thoughts. With regard to Charts: Looks like the reasons you provide are problems with the people who use charts --- not with charts themselves. We would not blame a hammer if a person uses it to kill another. Would we? Regards, Monoid.
You need your win/loss percentage, avg loss and avg win: These are variables to plug into a mathematical equation. Even if your method is "winging it" you need to keep the records that will allow your to calculate your w/l % your avg amount lost and your avg amount won to calculate whether "winging it" has an edge for you.
Would that include totally random systems that made money purely by random chance ? Take a look at the quarterly NFA statistics, around 30% of retail traders are profitable over the 3 month reporting period, and thats exactly what you might expect to find by random chance. I'm not sure that equating profitability over such a limited number of trades with an edge is particularly sensible. Anyone whose done any research into TA based systems will tell you that its fairly common to find simple TA setups that are profitable for prolonged periods of time on a given market, but that the same set ups can almost instantaneously stop working, or even completely reverse. I'd probably argue that *the* edge is the capability of actually developing an edge, and in order to do that, it takes a fairly substantial committment in time and effort. Markets are non stationary time series, and the point made earlier regarding the ability to spot when an edge is no longer producing the results expected, and the ability to act on that information is about as close as you'll get to practically defining what a trading edge involves.
Think people are getting a little hung up on the charts point. Another angle: one should have a model which describes a market based on testable parameters this then gives a baseline / estimated fair value then you trade deviations from that value (again if you can understand why the deviations should occur) so the price compared to your model gives whether there is a signal to trade whereas people go wrong not having a model and so just start comparing price to itself or price in the recent past . Charts are just a simple/lazy way to do make this unproductive pseudo-comparison. But they also have the additional disadvantages garachen described. So in addition to not having a proper strategy in the first place you now have something which is likely to trigger incorrect decisions. Charts give an approximation/summary - to compensate for the fact that the human brain cannot process every price change. So lets ignore the inside market and focus on trades only, ignore the sizes being transacted, and pick out e.g. 4 transactions essentially at random over an interval of time, and lets use the same time interval as a constant despite the obvious fact that the pace of a product changes throughout the trading session. -not logical You are preserving something far less than 1% of the data and you've chosen which to discard essentially at random : " " Inability to think abstractly and the tendency for retail to see everything as directional in a single product stumbling points here. Perhaps a familiar simplistic concept is a spread - so similar to a situation where you can leg into a spread at better than fair value and then liquidate into e.g. exchange supported spread - but how you manage your orders and the risk of being filled on one side needs to be accounted for. And it is constructed as a payout structure rather than a single position so you may be doing e.g. high correlation microstructure work in a product complex where you don't get much net directional exposure (delta) compared to the convexity you are capturing. What you wrote earlier about high correlation microstructure relationships and looking for short term disparities (e.g. stops/ icebergs) comes to mind. When I first started in trading I found one of the true price edges - printing/stealing money - that rallymode refers to. Obviously this depended on someone making a mistake which didn't go on forever. Instead of seeing this as a windfall profit I started to seriously explore whether this could be an escape from the employment matrix which serves to capture most. I also by chance did some computer work for a successful manual trader and was exposed to / taught a few concepts which had value. Per rallymode's post above, I then went through the difficult and expensive cycle of learning that strategies / market approach need to be adapted. In hindsight I overvalued my ability, undervalued my time, and my lack of professional/institutional background was a definite disadvantage. Mathematical intuition too. So somewhere between not having done badly enough to quit but not well enough to retire - but we can only spend our time once. I can understand your position of discouraging most if they cannot come up with a viable plan to make that e.g. $300k and in hindsight I'd have made different decisions if I'd had access to correct information in the beginning. So certainly appreciate your contributions here on ET. I think perhaps my biggest meta-mistake was the failure to plan early enough to be able to provide value to others in the trading industry in order that in turn they could help me bridge areas in which I am weak. I've found some things which are legit and persistent in places where nearly nobody else seems to be looking but which I don't have the skill set to maximally exploit. (and paradoxically I probably never would have found these things if I'd followed a more traditional path)As you have mentioned elsewhere people tend to overvalue their strategy and undervalue the other components to fully leveraging that strategy. And simply didn't have the necessary perspective to start - yet was aggressive and trusted in my ability to make it work. A false sense of certainty or near certainty far more dangerous than uncertainty.
@Ghost_of_Blotto: I assume that the people you referred to is me (for I was the only one who made a comment on @garachen 's thoughts on charts) . I am not hung up on the chart point. I did not made a comment on @garachen 's thoughts on "edge" 'cos I consider it correct from an options market maker perspective. I don't know him, and, I assumed he is an options market maker; here, I could be wrong. However, his comment did imply that his was the only game to be played. The word speculation implies one takes a bet, and directional bets are also part of the game! I also had reservations about his "general opinion" #2 -- trading plan 'shouldn't it be make money' (we never say the business plan is 'to make money!' ... a trading plan like a business plan is about the 'how to' part of making money), #5, and #13. But any comment by me in that direction would not be in OP's intended direction for this thread. However, his comments on charts caught my attention and I was interested in understanding his point of view as none was given. As I pointed out earlier, his reasons are not problems with charts but with perspectives people bring with them when viewing charts. A chart is nothing but a visualization tool. A visualization tool can be used to visualize data within the confines of any model. One builds a visualization tool for a model not the other way around -- stating that the data points does not show in a chart means that the visualization tool is not properly defined for the said model. I agree with your statement: Charts give an approximation/summary. But as I mentioned in the above paragraph, data summarization is not the only use for charts. I also agree with the other point you made: using an artificial mechanism to segment data -- be it time based, volume, etc -- only adds more noise (I don't want to use the word randomness here). Given such noisy data making sound judgements become more difficult. I have said it else where and will repeat again: Statistics are good at giving us summary, but to trade we cannot rely on summary data; The more the information one acquires, the better the decision. This applies equally well to any holding period. Now you can see why I was interested in knowing a fellow trader's perspective on charts. A final point: A model can be developed based on past prices alone. Your claim that such models are invalid or that only models based on comparative value are valid is just not true. If your claim were true, then short-term trading of outrights cannot be profitably done, and would rekindle the great debate in statistics all over again! Experience has taught me that whenever I make sweeping statements that generalizes an entire subject, I am usually wrong for I have most likely not considered the positions of others who might have a different yet valid points of view. Further, having experience trading OTC fixed income, credit derivatives, and exchange traded options and futures, and having been exposed to some great minds, I have seen and do understand how people trading different product classes with different motives within each product class can and do think differently -- saying one way of modeling is the right way is just a shame! All the best. Regards, Monoid.
Blah blah = mean reversion. Good luck "estimating fair value" in a run-away trend. You'll get raped....