Using Time Stops?

Does anyone here use time stops?
By that I mean exiting a trade that either immediatley goes into consolidation or one that goes into consolidation after moving in your favor or not.
....1) Do you feel, after an extended perod of little movement, that your capital could be used better elsewhere?
....2) Do you feel at some point your analysis is no longer valid?
....3) If you use time stops, do you have a firm rule that you apply without question? Such as:
.........a. number of candles elapsed?
.........b. give weight to time of day, or day of week?

Obviously there are many variable such as timeframes, rangebound or trending conditions.
Do you use time stops in every condition, or use them in some but not others?

If you use time stops let us know how you apply them. There is no right or wrong answer.

Thanks
Bill

Or...

Do you use a time based exit on your statedgy for profit: at first glance that may not make sense, but if you think about the undelying dynamics its not much different than holding for a specific price point (somewhere in time).

So lets say your system generated 1000 signals in your historical testing. In addition to trying to determine what a good exit level (price point is) go back and back test the exit from a structured hold period. 1h 10h 24h... 240h, 480h etc. You may find there is a time element that can be capitalized.

Time also has impact on merket entry, Go through you trades and catagorize them by the hour they were entered and the run a p/l on them and see if there is a specific window of entry where the model (or the trader) is doing better.

Time based optimization is a very interesting thing; I'm only just starting to scratch the surface of it but I've made some interesting finds with regards to my own models as well as some third party applications.

Thanks FXA

This is interesting. My first exposure to Time Stops was in stock trading. I knew of swing traders who would start gradually tightening their stops on positions that weren't doing much. They didn't want to tie up their capital on a stale position so they would squeeze themselves out with their price stop--but it was essentially a Time Stop because the stop tightening was prompted by time.

My example is simplistic compared to what you are suggesting--actually exploiting time for profit. It would be great to hear from anyone who has studied exploiting time for profit.

Well let me chime back in on this again...

I was recently involved in a project where I did some time optimization on an application and trade model for an outside company.

We broke their signals down into several catagories and then tested them for n hours after entry. There were no stops or limits, it was simply taking the trades as they came (increasing the same position and hedging were allowed) and held them for a specific period of time. The test itself was statistically significant (eight years of data and over 3500 trades for the individual catagory listed below).

Because I was paid for SOME of the work I'm not going to disclose too much (as I may attempt to recover other $$ later).

The best trigger we tested at the following hold periods produced the follwoing results (many more hold periods and triggers were also tested - there were about 17 profitable scenarios that would have out performed Barclays CTA Index):

Hold Period #1
Annual Return on Average: 21.15%
Sharpe: 1.053
Max Closed DD: 10% (Open draw down would have been higher)

Hold Period #2
Annual Return on Average: 26.90%
Sharpe: 0.997
Max Closed DD: 7% (Open draw down would have been higher)

Hold Period #3
Annual Return on Average: 29.24%
Sharpe: 0.914
Max Closed DD: 13% (Open draw down would have been higher)

What I discovered in my own trading and held true in this application as well (which was completely different than my approach to entry) was, hold periods tend to be longer than most people would begin to believe. A model that signals once or twice a day, may find its most profitable period four or five TRADING days out (maybe longer).

Even if its not the right approach for the average trader, it's worth exploring as it can add a completely different dimension to your trading.

Thanks again FXA

I've never considered this in anything other than its simplest form.

I use time stops a few months ago when I felt that one of the pairs I was trading was going absolutely no where. I like trends and the pair was going sideways for two days. I looked at my other currency pairs and picked up more lots which ended up being a great move since I was trading the 4 hour TF.

I scalp a lot more now which means less time stops.

I use time stops. They are usually somewhere between 5 and 10 candles. When I enter a position I presumably have an edge. Over time that edge errodes to nothing better than 50:50 and I no longer choose to risk my money on nothing more than a gamble.

I will say that in a lot of testing I've done timed exits never improved overall equity but sometimes that's not the only priority in trading.

IMHO there's a very important time/price relation and something which might be called "ticks per hour"...perhaps something like a speedometer for the chart...

Basically mean: if price moves too fast too far, it will snap back more viciously than in a more measured decline. (I've often wondered whether charts couldn't be reinterpreted with price as a rolling / falling rubber ball and then letting a physicist loose on it to calculate vectors, speed, angles, reflection...might actually prove interesting..line charts would perhaps make this possible as trajectories / path traveled.)

What this also means is that if you choose your entry at a point in time after which the main impulse is already over, you WILL get these "timeouts" simply because the market goes into the resting or consolidating phase.

It is however imperative to figure out whether this means anything or whether it's just more delay before the continuation.

IMHO most of the time if one manually intervenes into a trade it's for the worse unless price has shown clear counter-indications of directional movement.

A lot of times I think you need to allow "price normalization" before the new prices are accepted and things move on.

So in short what I am saying is: time stops? Bollocks.

Price should kick you out, not time

Move your stops in closer candlewise if you're that anxious to leave the trade
At least give it that 50% chance of going where it should!
(Hey, 50% win is already quite an edge as long as you close no earlier than say at least 1.5 x your risk)

Now, who's got the brains and boredom to pretend the chart is really a monitor of a physical object with unseen elements (s/r) of obstruction? ;p

CHFJunkie PM'd this as he doesn't have access to the TR. Since I haven't seen much discussion on time stops I thought I could pass this along as CHF's reference is an undeniably good source:


Sorry that I send u in PM, as I don't have the privilege to post in the trading room.

Regarding ur thread about timestops,
I like Joe DiNapoli's rule about using time stops , he uses a rule that he calls the rule of three.
The rule of three he uses for exits depends on the timeframe u used mainly to enter the trade.
so if u r a 4H traders and u executed ur trade. after three consecutive bars or 12 hours , if the trade didnt move in ur favor , the odds decrease dramatically and u should exit the trade, I have experienced this so much.
I consider my entry as a poor one if the rule of three is applied. as A poor entry is not only an entry in a wrong direction but a trade that put u under stress before moving in ur favored direction.

Another approach to time stops, is using slopping angles that combine time/price , if price drifted away from the slopping line with angle , odds decrease dramatically.
I hope that helped.

Thanks, CHFJunkie

The reference to stress is interesting.

CHF reminded me to differentiate between time stops (Phil and WHTenn's responses) and trade timing (FXA's contribution).

Trade timing deserves a thread of its own; I suspect almost everyone has some basic rules about when to fish and when to cut bait. And a few who've made a science of it.

Thanks for the response, SL. Good points on time/price relation.

I look at moving stops in closer candlewise as a form of a time stop if you do it because your position has stalled for a significant period of time. Everyone's experience will be different depending on their trading style. My experience has been this: When I tighten my stops to x number of pips, I usually end up handing that amount back to the market.

I use what you call "time stops" in that my definition of a successful trade is one that reaches its predefined target before (timerunsout OR defined loss incurred).

The logic behind this is that you want to withstand whipsaws in the extreme short term, but even if the market never reaches a more moderate SL level after a reasonably long time (say 72 hours) and STILL your target has not been reached, then something fishy is going on, and as Phil said your edge has deteriorated towards uniform probability (or worse).

Sadly, life is more complicated for me because if I see clear reversal patterns before either part of the logical OR statement is satisfied, I will bail out. Price action got me in, it is more than capable of convincing me that it is time to get out. This why I choose to close my trades manually - the trader's mind and experience are indispensable.

some trader choose to trade in a volatile market session and stop the trade at end of the session,

some check the trade in the morning and dawn, and take action promptly,

i guess timed stop is indeed being done consciously or not. but again that is not the main reason the trade is closed, no?

the ussual thing to trigger the time framing would be news release, i guess,

That's been my experience as well. If you don't want to trade exit your position rather than handing more pips back to the market. The market wouldn't do the same for you if the situation were reversed.

Many systems have an extra stop built in that reflects this. When price is stagnant, this particular stop is tightened (and keeps tightening over time) since the trade has lost momentum and momentum is a key component in a good trade. This is simply recognizing that the odds of success have decreased as you pointed out.

I use fixed time stop on automatic trading for certain pairs in 2 ways:
don't open new trades after time X
closed all opened trades at time Y

The reason is easy to understand, the market volatility changes a lot during the different market sessions, and some methods are better for slow markets and others for volatile markets.

It is just to avoid not favorable conditions.

But I don't use anything like "if in the trade for X time close it" because for my strategies usually if I stay very long in a trade then it is a very profitable one.

Mine are generally "If in trade for X time and there is not Y pips, then close." Clearly is a trade is winning there's not reason to mindlessly close it based on time.

Hello Gwan

Thank you for the comments.

When you decide to trade (or decide not to) during a certain time of day, or week, or news event we would consider that trade timing.

If.. your trade isn't moving in either direction and you decide to exit after a predetermined number of candles because:
....1) You feel your analysis may have gone stale
....2) You feel your capital isn't being used to its fullest potential
....3) The market isn't confirming your belief
Then.. you would call it a time stop. There are those of us who believe the longer our trades go without direction, the more our probability diminishes.

Here's a real life example of both: Thursday afternoon I shorted USD/CHF. My trade moved only a few pips each way for several hours on a one hour chart. After a certain amount of hours I would have closed it as the market didn't seem to be confirming my bias. If I had closed it at that point I would have considered it a time stop.

However, the trade finally went my way, but I closed it out before the market closed on Friday. The trade was going in my favor, but I choose to close my positions before noon on Fridays. So that would be considered trade timing.

So.. time stops and trade timing are two seperate things, but here comes the paradox. If you use time stops, you need to be aware of trade timing.
Say you have a firm rule about time stops and you enter a trade with a non-Japanese pair, maybe AUD/USD, a few hours into the Asian session. You shouldn't expect that pair to move much for the rest of the session. On a 4 hour chart it wouldn't mean much. On a 15 or 30 minute chart you will probably see a string of sleeping candles.

After everyone's contributions so far I've come to this conclusion:
....1) IF you are going to use time stops, especially on shorter term charts, you need to consider your trade timing.
....2) If you decide to use a time stop, it should be treated like a price stoploss and not loosened.
....3) Determine time stop rules first, then develop trade timing rules that will minimize hitting your time stop.
....4) Time stop upon entry should be less forgiving than time stop for a trade that has actually moved in your favor.
........... a. (X) number of candles upon entry for trade to prove your bias is correct
........... b. (X) number of candles allowed for stagnation after trade is underway to protect against diminishing probability.
A and B should be two seperate considerations and both should be a predetermined amount.

If nothing else, this would be a lesson in entering trades at opportune times, and avoiding entry during known times of probable non-movement on a pair-by-pair basis.

Here is a decent text that will give the basics to timing and market impact...

It's a good launching point into model optimization and should open the doors on some new thinking. It doesn't deal with time as a stop but it does cover some of the "timing" into a trade issue.

Like most timing work it focuses a lot of volatility and that can be a double edged sword: If you've done any volatility analysis at all you know that there are swings in the sessions themselves that seem to follow a cyclical pattern (meaning there will be times when the UK session will display more/less vol than the US session on average and vise versa). Undoubtedly this will have something to do with shifts in economic policies which trains the worlds attention on one clump of dirt over another for a brief period of time.

So... if you're going to use volatility for optimization you need to be able to forecast volatility, and because volatility is derived from price, underlying all of that, you need to be able to forecast price to some extent - and if we could do THAT with any accuracy, well we wouldn't need to do anything else.

To me this is the real crux of most optimization (or any pricing model for that matter); if I could accurately forecast anything, I'd put those efforts on being correct about the market, not information extracted from it... but that's just me.

The reality is every model, because it will be built in a specific manner and has certain rules, will generally perform in certain conditions better than others. This becomes the window of opportunity for best execution: this will include timing impacts that include hours of the day, locations in a volatility cycle, proximity to certain news events etc. If you have statistically significant trading data, it makes sense to pinpoint the areas of efficiency in a model. After all... it is your money (or your client’s money) and all this work may mean you generate another two to three percent a year on average... At first blush that may seem insignificant; but if you’re trading professionally, it can be the difference between raising $25m and $50m… or put differently; $500,000 and $1m in management fees every year, and that much again in performance fees.

http://www.amazon.com/Optimal-Tradin...5313314&sr=1-2

Okay I’m back to the dungeon…

Thanks FXA

I just checked it out. Though it wouldn't be everyone's choice of reading it would certainly lead one to look at things through a bigger window. When I clicked on it, a few other books cropped up that I hadn't noticed before. That's why I like turning over new stones.
I'll read a few more of excerpts and give it some consideration.

Bill