Registered User Joined: 11/16/2004 Posts: 4
|
I am a little puzzled on calculating the risk/reward ratio.
For example, I am usisng ERTS on 3/5/05 as example. If my entry price is $65.66, the previous low price is $61.88 (on 2/10/05) and my risk ratio is 5.75% [( 65.66-61.88)/65.66 = 5.75%]. What is the reward price that I can use to calculate the reward ratio?
Thanks!
loyinlan
|
|
Worden Trainer
Joined: 10/1/2004 Posts: 18,819
|
This a bit outside of what we, as Trainers, can help you with. I will move this to the Market Talk section so other user's can offer their thoughts.
- Craig Here to Help!
|
|
Registered User Joined: 10/7/2004 Posts: 2,126
|
There are many ways to calculate your risk/reward ratio depending on what kind of player you are and what kind of risk you are willing to take. Now, if you are asking this questions it seems to me that you shouldn't be out there taking risk that you cannot measure, and probably don't even know. There is a great deal of literature out there that you can check to informe you on this topic. My advise to you is not to take risk until you understand the risk you are taking. good luck.
|
|
Worden Trainer
Joined: 10/1/2004 Posts: 18,819
|
How do others out here measure their own risk/reward?
- Craig Here to Help!
|
|
Registered User Joined: 10/7/2004 Posts: 799 Location: Duluth, GA
|
Hi, BigBlock:
I've done a huge amount of work in this area of Reward/Risk, since Money Management and Risk Control are key aspects to my trading.
I do some pretty hairy prefiltering to find and isolate stocks with appropriate price-action "personalities" from which I then search for specific entry candidates. Historical Reward/Risk analysis is a big part of that.
This is all very involved - way, way too much for a single post (or even a dozen long posts). I strongly recommend Van Tharp's book "Trade Your Way to Financial Freedom" as an excellent starting point.
Nonetheless, I'll try to hit a few high points, to help you start thinking about it. These comments are based on a Long position being taken, but apply equally (in concept) to Shorts.
1) You should identify the risk you are taking, since that is the basis for the reward evaluation. That is, look for the landmines before you start chasing the rainbows.
2) Adjust your position size in inverse proportion to the risk ... have a fixed $risk you are willing to take on a trade, and religiously limit the number of shares you take by that value. #shares = $risk / (Entry-InitStop) ... there are a couple of nuances to this I will mention later.
3) Also, limit the #shares by some reasonable percentage of your overall portfolio value ... that is, don't tie up too much equity in one position.
4) Make sure there is adequate liquidity to move the shares in & out of the position *as a function* of the SIZE of your position - don't just use a fixed threshold. Refer to my Worden Reports dated 5/22/03 and 5/23/03 for a full discussion of how to automate the implementation of #2, #3 & #4.
5) The initial Risk/share = Entry price - Stop Price (ignoring commission and slippage, for simplicity). You can make a decent guess at the Entry price ... in fact, you should have a threshold ABOVE which you do NOT enter, related to the Reward potential ... more on that later.
6) The initial Stop price should be chosen as: the nearest clearly-identifiable Support level below your Entry, *OR* a price = Entry minus a reasonable multiple of an average bar's True-Range volatility (wiggliness). Use the CLOSER of those two (ie the most cautious one) as the initial stop. There are a LOT of ways to calc the volatility. Refer to my 6/6/03 Worden Report for explanations of how to calculate Average True Range. I suggest your initial Risk multiplier should be somewhere between 2x and 3x the ATR - I use 2x, but that's in conjunction with a partial-exit methodology (a bit more on that, later).
7) If you end up using a Support-level based init stop point rather than a volatility-based one (ie the support point is closer to your entry price than the volatility would imply), then STILL you should limit your position size by the (less forgiving) volatility-risk delta. Reason is simple ... AFTER your trade ensues, presuming it goes in the direction you want, the volatility-potential remains, even tho the init stop gets further and further away.
8) NEVER take a position without IMMEDIATELY entering a worst-case stoploss order immediately thereafter. NEVER.
9) You can estimate your Entry price, and pinpoint your init stop, BEFORE you ever take the trade. Using the difference of those two, which is your Risk, you should ALWAYS calculate a "Minimum Acceptable Profit Target" (MAPT) as a multiple of the Risk. I've seen "defensible" multipliers as low as 1.6x, and as high as 3x ... personally I use 2x. You are not "allowed" to take any "Profit Grab" exits before this Min Target is passed (but stoploss exits ARE ok to use :~)
10) Once you have calculated the MAPT, BEFORE entering the trade, check out the chart to see how likely it looks to you that the stock's price might actually REACH that point, without hitting a clear Resistance point first, or without straining the bounds of credibility in light of the price-patterns of that stock in the fairly recent past. If it looks unlikely that it can reach the target within those limitations, then DON'T TAKE THE TRADE! Find some other horse to ride ... or just keep your money for another day.
11) Calculate all this stuff, do all the Sup/Res searches and decision making THE NIGHT BEFORE you take the trade. This is basic stuff. Don't put yourself in a position to have to make flash decisions during the "crazy hour" when the market opens. (This presumes you are not a died in the wool ballsy macho daytrader, of course :~)
12) OK, you're in the position. You've got your stop set, and your target in the sights. Now, RELAX and BACK OFF. Don't worry it to death. Last minute decisions will DESTROY all the work and planning you've done ... and that Reward/Risk thing you're shooting for will be just a fantasy.
13) As the position progresses, I am a firm believer in using TRAILING STOPS. That is, you gradually move your stop up as the price moves up. There are a LOT of ways to calculate this, but almost everybody agrees to two basic, unbreakable rules: a) "Ratchet" the stop up based on the highest price reached since entry (or a reasonable substitute). You can do this once a day, or as the day progresses ... but DO it. If you don't, your Risk will increase as the price moves in your direction, which is ungood. b) NEVER lower the actual price of your stop. NEVER. I don't care what wonderful news item is about to surface, or what your uncle says about his brother-in-law's inside track from the janitor that cleans the CEO's office every Tuesday ... no matter what, NEVER lower your stop.
14) Given those two rules, you can add more to make your stops "smarter". a) You can if you want adjust the volatility-risk-calculation based on each new bar, but if so, you should NEVER allow this to INCREASE the DELTA between the Ratchet High and the current stop. This might cause the delta to decrease, which pulls the stop closer. If the price is "going nowhere fast", this will get you out of the trade sooner than otherwise. b) Additionally, or instead of adjusting the volatility calc, you might want to gradually DECREASE (never increase) the ATR multiplier as the trade progresses. Again, this will tighten the stop-delta further with each passing day. If you think about ROI as well as R/R, this makes a lot of sense.
15) You could define more that one stop level per day. That is, you can have the worst-case ATR-based stop as your fail-safe, but also calc a separately-managed stop that is much tighter, which you ONLY apply in the last half hour or so of the day. The volatility calc for this would ignore intraday H-L, and just be an average of recent Close-to-Close moves. If you like that idea, and if you trade fulltime so that you can respond mid-day if necessary, you might want to also calc a third stop based on H-L volatility, and set it up as an Alert rather than a hard stop. Last (and least), you can even calc a stop related to the FIRST hour of the day. I advise against this for most people, but hey! if you can dig it, do it.
16) Almost as important as the loss-limit exits should be a set of profit-capture exit rules. I call the former exit types "stops" and the latter "grabs". I base the Alert levels for the grabs on projections the night before related to (nonstandard) Bollinger Bands plus a multiple of Average True range. As I said before, figure all this stuff out the night before, to minimize ulcers the next day. I usually set two profit-grab levels - the first one calls for me to tighten up my stops intraday, and the second one (at a really statistically-unlikely point) calls for me to take out some shares if it is hit.
17) OK ... if you've been counting, I've defined up to SIX different exit rules. You MUST have at least one .. the ATR-based stop. I strongly recommend the Close-stop as well, if you have the time to manage it at the end of the day, or can place a Stop Market on Close order with your broker. Also I strongly recommend the use of the statisically-based extreme-grab level, ideally as an Alert.
18) So, if you've got all these stops ... plus any that your actual trading system indicators might fire at you ... won't that take you out TOO SOON? What about "let your profits run". Boy I could write a whole chapter on this. But the key is, DON'T put all your faith in ANY ONE SIGNAL. Instead, set up a lot of reasonable rules, and as each one fires, WITHOUT HESITATION, take out a PORTION of your position. If the signal was premature or just plain wrong, hey ... you've still got shares riding the wave if it's going up, and you've still got more exit points that will backstop you if it's going down. Most importantly, you don't have to agonize over things. Example: if your position size is 1000 shares, each exit signal might tell you to take out 200 shares. If you use a per-share broker like Interactive or a bunch of others, OR if you trade pretty large positions, the impact on your commission cost is negligible.
Okay. I guess that's enough for a start. It may seem that I discussed controlling Risk more than projecting Reward, but hey, that's the reality of the market. The key is to assure that you CAPTURE your Reward, whatever it turns out to be, without fooling yourself about the Risk.
Final rule. If it stops being FUN, then STOP TRADING.
I hope this helps!
Jim Dean
Copyright 2005, James D. Dean, All Rights Reserved
|
|
Registered User Joined: 10/7/2004 Posts: 2,126
|
Tanstaafl I didn't need this info, but I have to say that it is a great respond. I am sure that Loyinlan who started the topic will find your respond of great help. As for me I was familiar with all your comments, but again I think it is great the someone knoledgeable on the topic will take the time and effort to the extend you did. See ya around.
|
|
Registered User Joined: 10/7/2004 Posts: 264
|
Loy, read what Jim has to say. Then read it again and again. Study it.
My short answer is that I think it depends upon your trading strategy. ERTS looks like a breakout play which by its very nature is hard to determine reward. The first area I would look to is the high on 2/7 at 69.84. Would expect that price to be tested? If so, 69.84 - 65.66 gives you $4.18 reward. Hopefully price moves through that level.
You could also use a measured move type of concept. The move out of consolidation was back in late Jan. from 60 to 68 before the current consolidation. Your expectation here could be a similar move from 66 to 74. This would give you $8.34 reward.
You have defined your risk at $3.78 per share. So $4.18 reward / $3.78 risk = 1.1 R/R ratio. Or $8.34 reward / $3.78 risk = 2.2 R/R ratio. So the question is whether either of these R/R ratios are acceptable to you.
The bigest problem with caculating a R/R ratio is determining the reward. That number is just speculation. There is a good chance the stock never gets to the target. Thus, imo it is important only to take trades with nice and high R/R ratios. I set 2.0 as an absolute min. and perfer 3 or more. Good luck.
|
|
Registered User Joined: 12/28/2004 Posts: 11
|
I estimate risk/reward by first calculating the % risk--(my entry price - my initial exit stop)/ (my entry price)*100 Then I use the % gain from the most recent intermediate market low as my predictor of % gain. I a an Elliott fan, so I tend to think subsequent uplegs will be on teh order of the prior upleg, so there is some reasoning behind this. Since most stocks move in tandem, I calculate the gain from the date of the last intermediate low on the SPY for all stocks. What I am doing is looking for is stocks that have made a strong leg up (those that have substantially beat the market) and that hae tight stops. I generally look at stocks that have a reward/risk ratio of 3 or more. Georgia Dame Rapid Scanner
|
|
Registered User Joined: 10/7/2004 Posts: 799 Location: Duluth, GA
|
Hi, BigBlock:
Hey, sorry ... I was in a hurry and I did not look far enough back to see who the original poster was. Thanks for the gracious response!
Jim Dean
|
|
Registered User Joined: 11/16/2004 Posts: 4
|
Thank you so much for all your information and help. I appreciate a lot.
Especial many thanks to Tanstaafl, Golfman25 and Gaeagif. I will read your notes again and again and study them. Also, I will get the book "Trade Your Way to Financial Freedom" to learn more. Thanks again.
|
|
Registered User Joined: 10/7/2004 Posts: 799 Location: Duluth, GA
|
Elsewhere I've written a lot about determining and controlling Risk during a trade. Here is a very useful idea that helps you home in on symbols with good Reward/Risk ratios, using a PCF to automate the search and ranking of the symbols. Two important assumptions are the basis for this:
1) Risk can be defined, in general, as a multiple of the Average Daily Range (ADR) = Avg(H-L,n) = AvgHn-AvgLn. I've used this approximation for simplicity, instead of the more-useful Average True Range = Avg( Max(H,C1)-Min(L,C1) ,n) = Avg( (H-L+Abs(H-C1)+Abs(L-C1))/2 ,n) = a long PCF. For most (tradeable) stocks, the ATR = ADR * X, where (very approximately) X = about 1.1 to 1.4. Let's say that we want to estimate our volatility-based Risk (for setting Stops) to be 2.5xATR, which might be approximated as 3xADR = 3*(AvgHn-AvgLn).
2) Reward potential can be defined, in general, as how far the price has moved, pivot to pivot, in either direction, within a timeframe that's close to your average hold time. It's very difficult in PCF's to actually find and work with true pivots, so we approximate that idea by just asking what the MaxH to MinL move is during that period: MaxHn-MinLn. Or, if you are concerned that the H & L might be "flukes", you might prefer using MaxCn-MinCn. Personally I like picking the middle ground (long discussion as to why) - let me suggest something like MaxHn+MaxCn-MinLn-MinCn.
Let's say your typical hold time is ten (trading) days. Let's further assume that your system for picking candidates, and finding your entry point, and deciding where to exit, allows you to capture only HALF of the days in the move ... that is, in a 20-day period, you can potentially capture half of the MaxH20-MinL20 price-action. This, again, is a gross approximation, but it is not a bad starting point.
Okay we have almost all the pieces we need now. The only other key question is *which* 20-day period should be used as our metric for the Reward-Risk estimate - and the answer is that there is no such specific window that is better than any other one ... what we need is a representative average of a bunch of 20-day windows to yield a useful result.
Furthermore, the Reward potential of a particular 20-day window should be measured against the volatility-Risk that occured DURING that SAME 20-day window. That is, for the most recent 20-day window:
Reward/Risk Ratio = RRR = (MaxH20+MaxC20-MinL20-MinC20)/2/(3*(AvgH20-AvgL20)) = (MaxH20+MaxC20-MinL20-MinC20)/(AvgH20-AvgL20)/6
Almost there ... now we need to average THIS formula over a succession of prior windows. I suggest you cover an historical-lookback period of 2-4x the window-size you are checking. Ideally, since you don't really know when the pivots occurred, AND since recent action is probably more representative of what your trade will entail than past action, you should consider some kind of gentle forward-weighting of a whole bunch (40-80) of window-calculations, sliding the window back one day at a time. This requires a lot of typing, and some fancy weighting math ... best done via Excel - so here's a simpler approach ...
Using our 20-day window example (for a typical 10-day hold time), create a simple average of the RRR using samples of the following overlapping periods, which gives a decent sample baseline, with greater emphasis on recent activity: Today, 5daysAgo, 10daysAgo, 20daysAgo, 30daysAgo, 45daysAgo, 60daysAgo
That is, we expand our RRR formula thus:
(MaxH20+MaxC20-MinL20-MinC20)/(AvgH20-AvgL20)/6 + (MaxH20.5+MaxC20.5-MinL20.5-MinC20.5)/(AvgH20.5-AvgL20.5)/6 + (MaxH20.10+MaxC20.10-MinL20.10-MinC20.10)/(AvgH20.10-AvgL20.10)/6 + (MaxH20.20+MaxC20.20-MinL20.20-MinC20.20)/(AvgH20.20-AvgL20.20)/6 + (MaxH20.30+MaxC20.30-MinL20.30-MinC20.30)/(AvgH20.30-AvgL20.30)/6 + (MaxH20.45+MaxC20.45-MinL20.45-MinC20.45)/(AvgH20.45-AvgL20.45)/6 + (MaxH20.60+MaxC20.60-MinL20.60-MinC20.60)/(AvgH20.60-AvgL20.60)/6 ) / 7
If you put that PCF into a Watchlist tab column, you can sort a watchlist like the Russell 3k or the SP-500 to see a wide variety of results (use Zoom 7 since it shows the past 99 bars, which encompasses the entire period of your calc's). The higher values represent trading candidates with better price-action "personalities" than the low numbers. There IS NO magic threshold ... but it gives you a good starting point to continue the selection process from.
However, having done all that, there is one GOTCHA! When you look through the list of high-RRR symbols that the formula identifies, you will see a lot of them that have one or two REALLY BIG BARS/GAPS during the lookback period. Unless you have a "jenUeyen" crystal ball, you will not be able to capture those really dramatic moves with any degree of regularity or certainty ... so I suggest simply that you REJECT stocks that act like that, just as you would not bring a rabid dog to a fancy dress ball.
It's pretty easy to do this manually, since the big jumps are readily apparent. First, decide on an RRR lower-limit that fits your requirements ... I suggest maybe 2.0x, but not lower than about 1.5x, and copy the symbols with a higher RRR than that to a new watchlist. Then quickly spacebar thru that list and flag any symbols for which you see a big jump show up - don't take a lot of time to study ... just hit the F key and move on. Maybe repeat this process a couple of times. Now, delete the flagged symbols. The ones that remain have decent RRR's, and don't have a habit of surprising you to provide them.
From there, it's up to you! These symbols are NOT necessarily candidates for entry tomorrow ... but rather they are a "circle of friends" that you want to work within for the next week or so. I suggest you do this culling process once a week ... SAVE the lists from each prior week. Once a month, use flagging and sublisting to find the symbols that have been present in EVERY ONE of that month's weekly lists, and think of these as your GOOD friends, moving forward, as long as they remain in the weekly lists.
A somewhat more robust variant of this process is one that I came up with several years ago, and have been improving on here and there ever since. I've found to be VERY useful as an important part of my pre-filtering, along with liquidity analysis (see my 5/22/03 & 5/23/03 Worden Reports), and about eight other "personality" filters. I hope you find it useful.
Jim Dean
Copyright 2005, All Rights Reserved
|
|
Registered User Joined: 10/7/2004 Posts: 799 Location: Duluth, GA
|
Oops there should have been one additional opening parenthesis at the beginning of the formula .. correct version is:
( (MaxH20+MaxC20-MinL20-MinC20)/(AvgH20-AvgL20)/6 + (MaxH20.5+MaxC20.5-MinL20.5-MinC20.5)/(AvgH20.5-AvgL20.5)/6 + (MaxH20.10+MaxC20.10-MinL20.10-MinC20.10)/(AvgH20.10-AvgL20.10)/6 + (MaxH20.20+MaxC20.20-MinL20.20-MinC20.20)/(AvgH20.20-AvgL20.20)/6 + (MaxH20.30+MaxC20.30-MinL20.30-MinC20.30)/(AvgH20.30-AvgL20.30)/6 + (MaxH20.45+MaxC20.45-MinL20.45-MinC20.45)/(AvgH20.45-AvgL20.45)/6 + (MaxH20.60+MaxC20.60-MinL20.60-MinC20.60)/(AvgH20.60-AvgL20.60)/6 ) / 7
|
|
Registered User Joined: 10/7/2004 Posts: 799 Location: Duluth, GA
|
One other note -
This RRR gives equal weight to up moves and down moves ... it does not distinguish between them. It will find stocks that have long ramps up or down, and also stocks that tend to cycle up and down (in reasonably long waves).
So, please don't use this as a tool to identify JUST "long" candidates ... again ... it is a "personality" filter, not a setup-search.
Jim Dean
|
|
Registered User Joined: 2/27/2005 Posts: 2
|
Not sure how much this will help anyone... and it may look familiar to some of you who looked into trading the ForEx market with a certain piece of software.
Basically, the method they presented was pretty simple. They recommended setting stops and limits and saying, for every 10% you want to make, you're willing to risk a loss of 5%. So If a stock or fund or currency or whatever was trading at 10.00, you'd set a limit at 11.00, and a stop at 9.50.
The theory was that you only have to be right 50% of the time to come out ahead. Sounds good in theory... but didn't turn out to be successful for me in my paper trading.
|
|
Registered User Joined: 10/7/2004 Posts: 799 Location: Duluth, GA
|
I think that most will agree that you determine your risk first, based on support levels &/or volatility, then you calculate the necessary profit to make the trade worthwhile, typically 1.5-2.5x the risk, then look for areas of likely resistance to see if the profit is obtainable. If not, look for a different opportunity.
The manner in which support & resistance levels are determined, and the volatility-based risk is evaluated is the big gray area ... lots of different recipies.
Jim Dean
|
|
Registered User Joined: 3/20/2005 Posts: 78
|
Commissions should be included in your risk/reward assessments, too, if you are trading under $5000, or paying fairly high commissions.
For example, Brown|Co has a Market Fee of $5 and a Limit Fee of $10 for Internet or Automated Telephone order entries.
If you buy at the market and sell on a stop, you pay $15 total.
With a $15 total cost, if you buy $3000 worth of stock, you need a one half percent gain in the stock price to break even; for $1500, one percent; $750, two percent; $375, four percent!
|
|
Registered User Joined: 10/7/2004 Posts: 264
|
Jeff,
The problem with the method your forex guys described was that it was completely arbitrary. It is better than nothing, but still arbitrary. It didn't take into account anything about the stock (or pair in your case) you were trading. You need to take into account the volatility of the trading vehicle and/or the location of support or resistence. As Jim says, that is the big grey area.
|
|
Platinum Customer
Joined: 1/24/2005 Posts: 195
|
I'm not real adept at programming in tc, but this is what I use and it has held up well for me, any suggestions are welcome. First I size up my risk by identifying my stop, usually a support level, a channel, whatever you like. When I use a support level though initially I will have my stop below it so as not to have an obvious stop level run by the MM. Then I'll determine roughly where I'd like my entry, usually very close to my stop. I'll calculate my risk/reward based on resistance levels and just plain eyeballing it. Then I calculate 2% of my portfolio and that determines how much I can lose on the trade if I'm wrong. So if 2% of my portfolio is $200 and I've determined my risk to be $1 a share, then 200 shares is the size of my position, so long as the position is less than 25% of my portfolio and is not highly correlated with other positions in my portfolio. As far as the risk/reward ratio, I want minimum 3:1. Often I'll only go for 6 and 7:1. Once the position starts moving in my favor I use a custom price channel:((h-l)+(h1-l1)+(h2-l2)+(h3-l3)+(h4-l4)+(h5-l5)+(h6-l6)+(h7-l7)+(h8-l8)+(h9-l9)+(h10-l10))/11 I have the width multiplier set to 10 and smoothing at 1. If I'm long and the channel is rising, I use the highest point of the lower line as my stop. Conversely, if I'm short I use the lowest point of the upper line as my stop. I like it because 1) my stops are rarely at obvious level and it continues to tighten, but allows enough room for swings without knocking me out too often. Sometimes I'll give it a little more room by just switching from a 1 to a 2 day chart.
|
|
Guest-1 |