I am doing some statistical work on the market, specifically straight up days and straight down days with no intraday reversal to the mean or a fib 1/3 retracement. ie., last friday. Statistically these specific market days occur in the lower percentile range Are there any meaningful technical indicatiors that may give a high statisica probablity of a potential straight up move coming up. The most compelling technical stat i have come across thus far is buy/sell side volume. But its not that accurate. works better over the long term-3-7 days Any help much appreciated.
good question. i trade reversals and would like to avoid or even exploit these type of days. what do you mean by buy/sell side volume?
Yes I have reviewd option expiration days, beginning of new months from mutual fund flows and end of month markups (1 week prior). These events are usually plus or minus a couple of days. There must be an indicator that at least gives a high degree of indication of one way intraday direction. i'v been preliminarily looking at volup ticks. Buy volume over sell volume is a pretty good indicator. you just have to figure what ratio gives the best percentages of tipping off the big sell or the melt up. I'm a fader and play off reversals as well. Batting average is 80 + percent. It is the straight/down up days that stops me out. No big deal but if i can refine it and stay out of those particular days or go with the fund flows I can have a higer win rate.
How do you determine what is buy and sell volume ? such as vol on uptick/ downtick, up day , down day ? or vol on bid and ask ?and what period do you look at to determine the probability of a trend day?
Funny you mention this as I spent alot of time this weekend trying to figure out the exact same thing. I trade reversals and days like Friday usually clobber me good. I actually ended up +2 ES points on Friday, so I'm learning something apparently. The best indicator I've found so far for this is your own trades. If you take two trades in a row and they are both wrong, you are probably in a trend. Either stop trading for the day or switch to a trend following strategy. I am in the process of modifying my system to switch me back and forth between my trend and my countertrend system based on a few factors. I think it is going well, but hard to say without alot more testing, which I will be doing over the next few weeks. I know this isn't a statistical/predictive type thing, but I thought you might be interested. I am actually a reactive, not a predictive, trader.
Your two trade rule is indeed a very good rule. I usually look at time, from the opening to 10.30-11.00, if the market is straight up or down with the requisited breadth rules then its usually a decent call on direction. However, the screwiest thing I saw was this past November, where adv.decl breadth was actually negative 2-1 and the market would go up, thought it wasn;t straight up it was nearly a constant melt up. So inherently there is a flaw with any indicator. A good measure as well is usually big moves are preceded by expanding volatility.
------------- Jrmarketwise; FRI was a bit tricky in SPY due to above average % sell of in bull market & didnt apper to be very extended, open price fri or thur eod, which hint reversals many times; some strong stocks did avoid it , closed up. Another thing that was consistant about friday, unusual % red sell price volume hourly candle charts ; on hourly charts better have a good reason to go against them. In other words dont know if you are ignoring hourly charts;helps in multiday swing /position trade entry , ALL hourly charts made lower lows,lower closes fri. Some like smaller intraday charts , sure; but more noise , like 5 minute chart is plenty of noise/slop/chop/slop. Stuff like cci hits about 80%; but usually wouldnt act on any signal that contradicts red candle price-sell volume like fri.[entry,exit swing position]