For instance I was backtesting one model and after 40 trades (N=40), the win ratio was a horrible 42% and did not fit into my criteria of a 60% win ratio. So no point of continuing to manually (read: painful) backtest another 50 trades right?
Think about it this way. If your strategy took 100 trades and 42 were winners but your winners made 5 points each and 58 of your losers lost 1 point each. Thats a great strategy even though only 42% won. So its not just about with win %.
Some general rules that I keep in mind when developing my models, hopefully these will help you:
Strategies that trade less, capture more of a move are preferable over strategies that scalp. So focus on strategies that trade for longer durations, higher time frames etc. I would prefer a strategy that takes 1 trade a day and goes for 10 ES points to a strategy that takes 10 trades and goes for 1 ES point/trade any day.
All backtests should assume slippage. Especially in thinner markets. For instance, in CL, you want to assume 2 ticks of slippage on each side - entry and exit. For ES, assume 1 tick slippage each side. If your backtests with these slippages included are successful, that model is generally a good candidate for forward testing.
Even when you include all the suggestions that everyone has given you regarding backtesting, your strategies will not perform the same way in live markets as they did in backtests. So the next step is to run it in SIM, validate it works properly and validate that each entry/exit got filled for either a sufficient period of time or number of trades. If after this validation, the results are satisfactory, then go live with it. @kevinkdog has an entire thread on this.
This is the single most important point: Think about your competition and what is it that you bring to the table vs what they do. Is your model something special that just you or very few people might have? If not, what makes you think it will give you an edge in these markets. Please remember that this is a highly competitive edge we are talking about. If it wasnt so competitive and difficult, everyone would be running a model trading ES and/or CL from their basements. You have to have something special, a niche, that no one else has. If the markets are Gotham city, you have to be Batman otherwise you dont stand a chance. Poor analogy, I know but this is really important to think about. This stuff isnt easy.
Last edited by Hulk; October 26th, 2014 at 12:59 AM.
The following 3 users say Thank You to Hulk for this post:
I assume you're asking about forex brokers that use NT? I'm not sure now except for MBTrading. Ninjatrader since starting their own brokerage don't list that many on their site info now:
(probably this question can be re-asked on Ninjatrader support forums)
a trader goes into a bar and orders a drink. as he sits there sipping his single-malt, a young lady sits down next to him. she says, 'I'm a lesbian” i spend my whole day thinking about women!” a little while later, a couple sits down next to the old trader and asks him, 'are you a futures trader? 'he replies, “i always thought I was, but I just found out i'm a lesbian."the mistake the trader makes is that he assumes that the definition of a lesbian is somebody who spends the "whole day thinking about women." the reason the joke works is because in a certain way that definition could apply to lesbians, but it fails to address the point that a lesbian is a gay female. the trader is neither gay nor female; therefore, he is obviously, not a lesbian.
this underscores the basic epistemological problem of establishing causal relationships and is one of the predominant reasons traders are unsuccessful. false premises often lead to bad results; but, what is even worse is that a strategy based on a premise that is false, can in fact, result in a positive outcome. in the former case, most new traders tend to make excuses and place the blame on external factors and not themselves or their strategy, e.g., my premise is right, but the market isn’t making sense or is wrong, the market is choppy, the market is manipulated, the algos are hunting my stops, the market is too volatile, the market is directionless, etc.. there are 2 variables in trading; the trader and the market he trades. improvement must come from the trader and not the tradee.
in the latter case, the random success of of a pattern, set-up, or formation, is often mistaken as a reproducible edge. even the fair toss of a coin has a 77% probability that a streak of 5 consecutive heads will occur. the false confidence gained from random success often distorts the naive trader’s perception of his methodology. the set-up may have been successful purely by chance, or may have worked well recently, or under certain conditions, but not provide a well-defined historical edge. it is extremely difficult to differentiate between a deterministic edge and random results. this is why it is important to question your successes as often as your failures. this is why it is important to question your strategy/methodology, and this is why you must question yourself.
traders who have capital, confidence, commitment, and learn to adapt their techniques to varying market environments and to current price drivers, can and will succeed. however, this requires continuous incremental improvements and new insights. it means knowing when to improve performance by further mastering existing tools, and when to invent/learn new tools. this kind of expertise can take years to achieve, yet most new traders believe the process is easily acquired and mastered -false premise #1.
"contradictions do not exist. whenever you think that you are facing a contradiction, check your premises. you will find that one of them is wrong." false premise #2 - i am objective and unbiased about the market’s direction. most traders are guided by confirmation bias, and use their chart patterns and set-ups to convert their bias into a logical argument.
false premise #3 - a good entry and tight stops reduces my risk and trailing stops protect my profits. arguably, the most stultifying practice in trading, it is predicated on the totally subjective decision of risk/reward and where the stops s/b placed. price change is not linear, and while a price chart models relevant aspects of price change, such models invariably support relationships that do not correspond to the original process. so, a while a trendline or a swing high/low may present a visual representation of a viable place to place one’s stop, it has no intrinsic meaning for the price series, and only serves as target for algos.
which of course, leads to false premise #4, which is charts and technical analysis, and the patterns, formations and set-ups they produce, are forward looking and can predict the market's direction. charts no longer play the role they once played and are simply artifacts of random price action. we are now in the post-technical analysis era where hfts, central bank policy, and sovereign agendas drive the markets. traditional technical analysis is like a religion, and for most traders, logic and reason does not play a role, in the practice of their religion.
carbon-copy trading, puts one in the herd and in a very crowded space and only sets one up to be disappointed, while a realistic approach allows one to act uniquely, and sets one up to be creative - markets begins to take on nuance and progressively more meaning. recent moves in the spx and crude have been veritable treasure troves for short term traders, where big cycle moves meant that traders willing to trade the market won big. at first domestic qe drove equity prices higher, and now global qe is the driver behind the equity move. the economic and monetary divergence that favors a strong us dollar, along with saudi policy is the driver behind the large linear moves in crude.
the opportunities are there. even on the rangebound days the market will cycle from low to high and from high to low. and one just needs to trade the the market to make money. but, take it a step further and use your favorite reversion strategy to buy/sell pullbacks from the trend. now use 2 independent edges for confirmation and to give you confidence in the trade. you are trading in direction of the trend, but now you are using a strategy with a bigger edge – because you are trading w/ the drift and you can give your trade a little more rope. MOST IMPORTANT - be patient and adjust trade size and stop-loss levels to prevent getting stopped-out of a potential winner by managing expected value along with p&l, while allowing for a margin of error, so that you may stay-in-the-trade, and then hold on to the trade.
The following 15 users say Thank You to tigertrader for this post: