well it should change because adding the random condition reduces the probability of exit in 2. so 2xholding time is expected with random vs. without random
If I buy when price crosses above 50 ema and sell when it crosses below the 50 ema would this be considered a random entry also? If not , Why?
Bill, Fine, you never attacked Mike, sure, got it. Amazing that you cannot seem to just ignore Mike. I'm sure a psychologist has a theory about such responses--bullied as a child, something. My two cents, move on or contribute. You didn't see me continue to bicker with goodgoing now did you??? As for your analogy, the car, don't really see how it applies, but no worries. Please feel free to contribute a new analogy. Now, there is a very important distinction between Mike's contributions and my own. Mike is actually looking at applying the random # generator to make trades. I am looking at comparing a system's trading results to a random distribution. I agree, Mike's is a bit 'out there', but the simulations speak for themselves...just wanted to clarify... As for this statement: You can distinguish those two by the generation processs but not from secondary processes like an equity curve. . A) What is the 'generation process'? Is it the idea? I have plenty of ideas, but that does not preclude the possiblity that the idea is not trading markets noise. (I personally equate noise with randomness, may be wrong). And how is the equity curve not a 1st order product of the system. What is? Win%, profit factor? Those are intimately related the equity curve and, in my mind, are 1st order products. Bill, you seem to have some knowledge, some abilities, so please share. masterjaz
Blox, This is where my thinking on randomness/noise vs. edge comes in. First, you have a theortical reason for such a trade such that I, personally, would not call your entry random. Now, does the system have an edge? This is where i would look at the system stats (win% and avg win/avg loss) and run simulations to look at possible equity curves (see prior post for website or I'd use my spreadsheet). I would also run a simulation looking at random entries on the same instrument using some statistic from the system (avg holding time...) to see if there is a difference between the fundamental/theoritical system and a the random system. My ultimate goal is to determine if the system I created trades a true edge or did the system just produce a decent equity curve from market noise (randomness) ala the results Mike posted a while back (before adding conditions). masterjaz
I like to use monte carlo . Even with some systems having good results, I always have the " when will the system stop working" question in my head . When a system stops working does this prove that their success was random? Is it possible for a nice 10 year equity curve with relatively low DD be due to randomness or the "edge" ceasing to work further?
Just go back in this thread and look who called me a douche. I was not even participating at that point. I saw his post by chance. You are not going to tell me what I should do or what I must do. Please refrain from making suggestions. You have no authority for that. You asked a question, which in my opinion is a valid one but has an easy answer. You asked what we can learn from coin flipping as related to trading. My response to you was clear. Nothing. If you do not understand this is because you have little or no trading experience and obviously no formal education in probability theory. I will try to help you though, despite the fact that you are too, way too, arrogant. Coin flipping is an experiment in probability where the outcome of each trial, the tossing of a coin, has a probability P that is independent of the outcome of previous trials. This is in page 1 of every reputable probability book, obviously you have read none. Had you read one, you would be able to realize that markets have no relation to coin tossing. This has been proven over and over again. Market movements have memory because traders have memory. Their decisions depend on previous decisions. Traders sell to exit long positions. They buy to exit short positions. They buy more because they accumulate during momentum. there is nothing random about (most) traderâs actions. There is always some type of decision process involved. Each day when the market opens, it is not like a new trial in an experiment of tossing a coin. Bayesian probability plays an important role. Traders use new information to re-evaluate the probability of the market rising or falling. Their actions, whether to buy or sell, depend on that re-evaluation. Therefore, market movements and coin flipping have nothing in common. There is nothing you can learn from coin flipping about the market. You only learn about the market, when you trade. Obviously, your experience is limited in this way. This is fine. One could use though coin flipping to trade. Fine. You can do that. This is democracy. Now, this is what the idiot Mike does, he does this on purpose because he has malicious intents. First, you do not realize that he generates noise to the thread with his stupid examples, literally. His objective is to attract some people he hates and insult them. This is what he does. As a result, a lot of people do not participate. He presents himself as an authority but in reality he is probably one of those rednecks who live in trailers some place in obamaland and he is so frustrated he is using the web to attack people. Go back in the thread to see that he made an unprovoked attack insulting me.
intradaybill and Mike, please stop fighting. Who cares who started? I'm really interested in this thread and every second post has nothing to do with the topic. I've read here a lot about random conditions, previous day's close etc, etc. Where can I learn these concepts and how Mike runs this test? Maybe someone can explain it in English to me? Thank you.
Just go back in this thread and look who called me a douche. I was not even participating at that point. I saw his post by chance. Fair enough, BUT, why did you feel the need to respond? Why couldn't you just go "oh Mike, you're the douche" and walk away. Point proven, you cannot let things go, apparently. You are not going to tell me what I should do or what I must do. Please refrain from making suggestions. You have no authority for that. And you have the authority to tell me for what I don't have authority, interesting... You asked a question, which in my opinion is a valid one but has an easy answer. You asked what we can learn from coin flipping as related to trading. My response to you was clear. Nothing. If you do not understand this is because you have little or no trading experience and obviously no formal education in probability theory. I will try to help you though, despite the fact that you are too, way too, arrogant. You sure know a lot about me. Wow, you are almost God-like, omniscient, omnipresent, and the rest. Must be nice, you must be pulling sh!t-tons of money out of the market knowing so much. Coin flipping is an experiment in probability where the outcome of each trial, the tossing of a coin, has a probability P that is independent of the outcome of previous trials. This is in page 1 of every reputable probability book, obviously you have read none. True, this I know, but had you been reading my posts carefully, I have never said the market is random, I am of the opinion that the markets follow a dynamic distribution such that each day may appear random, but serial patterns exist. Even if that is wrong, my main point in this thread is not proving random entries, random exits work, rather how do I prove or disprove if you like null hypothesis testing that my system is trading a true edge or just market noise. And as I stated, to me, and only me, noise = randomness, to a degree. Had you read one, you would be able to realize that markets have no relation to coin tossing. This has been proven over and over again. Market movements have memory because traders have memory. Their decisions depend on previous decisions. Traders sell to exit long positions. They buy to exit short positions. They buy more because they accumulate during momentum. there is nothing random about (most) traderâs actions. There is always some type of decision process involved. Each day when the market opens, it is not like a new trial in an experiment of tossing a coin. Bayesian probability plays an important role. Traders use new information to re-evaluate the probability of the market rising or falling. Their actions, whether to buy or sell, depend on that re-evaluation. Therefore, market movements and coin flipping have nothing in common. There is nothing you can learn from coin flipping about the market. You only learn about the market, when you trade. Obviously, your experience is limited in this way. This is fine. One could use though coin flipping to trade. Fine. You can do that. This is democracy. Now, this is what the idiot Mike does, he does this on purpose because he has malicious intents. First, you do not realize that he generates noise to the thread with his stupid examples, literally. His objective is to attract some people he hates and insult them. This is what he does. As a result, a lot of people do not participate. He presents himself as an authority but in reality he is probably one of those rednecks who live in trailers some place in obamaland and he is so frustrated he is using the web to attack people. Okay, I yield, you are right, this is exactly why Mike joined this thread. I see it now. He did not want to participate in the discussion. His only goal was to lure you from hiding and insult you. I see it now, thanks for clearing my vision. I will, ignore him from this point on... Go back in the thread to see that he made an unprovoked attack insulting me. It's true he did, but you weren't man enough to ignore him...don't know who the real douche is...have to ponder while I flip coins... Also, what about my question about "generation process" vs. "secondary process" and all the other items (conveniently) ignored. Come on man, let's discuss. Masterjaz out
I think some background info is in order. Bill is a huge fan of a type of process that is a permutation of traditional data-mining. What bill does is use a program that takes a set of desired equity curve characteristics and then performs a "pattern search" on a variety of time-series data to find results that fit the the orignal equity curve parameters. This is the "generation process" in a nutshell. This is where bill will probably tell everyone how I'm soooo clueless and idiotic with this explanantion because he thinks he's god and no one is as smart and clever as he is. While this generation process, which is a complex behind-the-back process of curve fitting, has some merit when applied correctly, it does not allow for lateral (read: creative/fundamental) edges, and, an even more frustrating result is the fact that bill will never truly know the fundamental reason why his systems may or may not work. Since, in essence, he is guessing whether his datamining produced a real edge or a random one. Bill's generation process skips the fundamental concept alltogether... that's why what we are talking about here bothers him sooo much. The blending of random and deterministic is bill's worst nightmare I hope I don't come across as dismissive of what bill does (I just have a poor personal opinion of our bipolar megalomanic). I think the procedure has potential when used correclty, BUT, and this is a huge BUT, the random element, which freaks bills out so much - and rightly so, is always lurking behind every "generation".
masterjaz, you are an idiot. Keep on flipping coins while I am not gay enough to ignore the attacks of your associates here. Stay with Mike here and continue giving head. That you know how to do well, it appears. You and Mike (again) on ignore. You are terminally ill people.