Righto.....so if someone buys or dumps a large parcel of shares or perhaps places a large dummy order, or fat finger trade, for you or any other trader there is logic behind it? What you do, may be logical to you, but if you were to ask 1000 ET members re your placed trade, most would call it illogical. The problem is, there is herding behaviour / group think, so someone on ET decides and post on ET that every price move has logic, so all the leemings nod in agreement.
You tested, eh? You put today's price against yesterday's, and plotted errors? You then tested whatever bevy of single predictors that can be dragged up, tested them, likewise plotted their errors, and found underwhelming results for Yesterday's Price for various meaningful cuts of time, throughout whatever definition of "history" you have available? Against your population of alternatives? Really? You did that? Impressive. DO TELL. • You found that the collected error of obs{n} =/= obs{n-1] was not equal to 0 while centered on obs{n-1} AND • You found that the variance σ[obs{n}] as a predictor was larger than that for any other single predictor you evaluated. Impressive! DO tell. Please! There's a publication in here for you, just waiting to be plucked. A paper/seminar tour, for sure. (Though -- it's no great shakes -- you get a per deim that won't pay the way, and maybe a room at the university-owned hotel, usually one step up from Motel 6...) AND, if you play your cards right, a walk-one role in any re-boot of The Big Bang Theory that may come to pass.... "Sweet!" ) Sorry, ironchef, but you can't just declare "This [indicator] sucks!" at a comparative statement, without completing and then showing your own comparison study. I didn't say that Yesterday's Price is a *perfect* predictor -- I didn't say it was good at all! What I said was that it remains the single best predictor for today's price. To respond to the contrary, you must put up your own goods. You must *beat* Yesterday's Price. (And if you think it's so easy, then it won't be any burden, eh.) So again, try a little thought experiment: imagine you're given three choices for today's S&P price -- 3867, yesterday's close, or 932. You MUST choose one. Get it wrong, and horrible things happen. What's your choice going to be? WHY? ["And there's your answer????" Exactly. ]
You are way too smart and sophisticated for me. I am just a pleb, a mom and pop amateur retail. My test was simple minded: If today's price is a good predictor of tomorrow's price, then, When the market is up today, if it is a good predictor of tomorrow's price, then tomorrow should also be up??? Turned out it was ~50:50 up or down tomorrow. Please tell me how you use today's price to predict tomorrow's price: up or down? Appreciate your response.
You haven't answered my question. You can. [But if you look in my portfolio, I have no long puts below 932, nor any long calls above 3867..."Ahem!"]
@ironchef - It is not a predictor, it is the most likely outcome. The higher the volatility of the product, the less likely it will ever be the actual, maybe never, but nonetheless it is always the most likely outcome - basic statistics. Tommy boy is just stating a mathematical certainty. (cool gif - wish I was trippin')
An random walk isn't a perfect model, but is the single best model of market movement. Besides, any number of studies doing R/S analysis have shown values barely above 0.5 for daily data, which is pretty damn close to random walk. However, once you go to monthly data, long range memory starts to come into the data.
Thank you for your help. You are one of the good guys. I do know some basic statistics, I was quoting professor Tom, hoping he could show me how to predict up/down from last value.
Ironchef, not a trick question, when you are looking for trading direction, ie what and when to trade, do you think in terms of what is price doing next - what is it going to do - "I'm hoping to predict and front run the next move"?