OmniVest Forum
-
User-Created Portfolios
What's your favorite risk-return measure? |
^ Top | |||
Steve Mayo![]() Posts: 414 Joined: 10/11/2012 Location: Austin, TX ![]() | As discussed in the OV-Pro seminar today, Mark Holstius and I (Steve Mayo) are working with Ed to map-out the design for a portfolio switcher. Part of that specification is what "evaluation functions" should we target. Here's the list we are currently considering: CAR MDD mCalmar = CAR/MDD, over the test period (not a fixed 3-years as normally done) mSortino = CAR/downside volatility, sans the adjustment for risk-free treasury rate mSharpe = avg period return / stdev of period returns, sans the adjustment for risk-free treasury rate Beta = COVAR(Port CAR, Market CAR)/VAR(Market CAR) mAlpha = Port CAR - (Market CAR * Port Beta), sans the adjustment for risk-free treasury rate Information Ratio = (Port CAR - Market CAR)/STDev of that Excess Return Expectation of Loss= %Losses (1- hit rate) * average of the period (monthly/weekly) losses Expectation of Gain = %Wins (hit rate) * average of the period gains Volatility-adjusted return{ = avg. period return / stdev of avg period return Risk-adjusted return = avg. period return /% losses * avg period loss | ||
^ Top | |||
George White![]() Posts: 5 Joined: 10/11/2012 ![]() | HOW ABOUT SLOPE OF ema(8) | ||
^ Top | |||
Steve Mayo![]() Posts: 414 Joined: 10/11/2012 Location: Austin, TX ![]() | Thanks for the suggestion, George. Hmm. Wouldn't that be better for the "Scoring" method instead? I'm basically asking what you would like to measure on the OUTPUT, that being the "switched" portolio's equity curve (a/k/a, the backtest). I don't think the recent slope of that equity curve would show if it is a good investment. We are planning to use most of the standard "indicators" like EMA, RSI, periodic return, etc. to score and rank portfolios to determine which one to switch to for the forthcoming month. There are hundred to thousands of possible permuations there. So, the evaluation function that I'm describing applies once you generate all those permutations. Obviously, we can't graph every one, so this evaluation function would be a way to sort the output table into something more reasonable. In that light, I think we would want something that evaluates the overall "goodness" of the resulting equity curves, probably something that balances risk and return like Sharpe or the others I listed, although some might just want the lowest possible volatility/risk or the highest possible return without regard to the other. | ||
^ Top | |||
Jim Dean![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() | I know this response will not likely be "popular" with the developers because of the data requirements, but I think it's IMPORTANT, so here goes. We need, in a lot of places in OV and OT, etc, a statistical measure of the RISK EXPOSURE that is taken, as a percentage of Entry price, for the various strategies in the various portfolios. That is, if two ports or strats have roughly similar CAR and MDD, I for one would be drawn to use the one that has me exposed to the least degree of true money-at-risk during the trade. The difficulty is of course that this can only be measured by looking at the price dataseries during the trade ... versus the fixed or trailing stop logic in each strategy. And afaik OVest is not "built" to track that kind of thing. Yet. So ... I don't expect this will be done soon ... but I think it's REALLY important, so I wanted to add it in response to your request for ideas. | ||
^ Top | |||
Steve Mayo![]() Posts: 414 Joined: 10/11/2012 Location: Austin, TX ![]() | Does anyone have any experience with the Modigliani ratio? First calculate "Excess Return" of the portfolio over the market's return for each period, say monthly. Excess Return = sum of (port month return - market monthly return) M2 = Sharpe x STDev of (Port Monthly Returns - Market Monthly Returns) M2 is basically the risk-adjusted return of a portfolio relative to the market. It is similar to Sharpe and would rank a series of ports the same way but rather than some abstract number, it is measured in units of percent return so it is more understandable. | ||
^ Top | |||
DBokhart![]() Posts: 3 Joined: 10/11/2012 Location: Dallas ![]() | Steve I prefer Ulcer Index and Ulcer Performance Index, similar to Sharpe ratio but it only measures the top half of the standard deviation ---essentially. Things that go up but not down, obviously everything goes down sometime, but it measures how close it comes to the ideal. The problem with just the UI is that a bank CD is perfect but returns are low, so you put in returns also to get the UPI as a measure of persistent moderate to high growth. Ulcer = \sqrt { R_1^2 + R_2^2 + \cdots R_N^2 \over N } UPI = { Return - RiskFreeReturn \over ulcer \, index } http://en.wikipedia.org/wiki/Ulcer_index After Peter Martin wrote his book in 1988 Fast TRack Software was developed to use it to trade Fidelity select funds as I understand it. It is widely used in ETF and mutual fund trading. Its slow, but a great measure of longer term persistency. http://www.fasttrack.net/newhelp/ulcer.asp http://stockcharts.com/help/doku.php?id=chart_school:technical_indicators:ulcer_index http://www.tangotools.com/ui/ui.htm I have the .pdf of Dr Gary Elsner's article and S & C magazine from about 10 years ago if you would like it, he's probably the current expert on it. This article has a better method of calculating than the original Peter Martin version. Ed has committed to putting it into OT when he gets time, (hint Ed) And Thanks for all the work you have done on OV. email me at dbokhart@aol.com if you would like a copy of the .pdf | ||
^ Top | |||
gbarber![]() Posts: 282 Joined: 12/30/2012 Location: Pearland, TX ![]() | I would like to see something along the lines of those ideas from Jim Dean and Dbokhart. average expected reward/risk at trade entry and average achieved reward/risk after the trade is closed if you then compute the ratio of achieved versus expected reward/risk numbers and pick the ones that are greater than one, you should have a high likelihood of making money. | ||
^ Top | |||
Eric Severance![]() Posts: 23 Joined: 10/11/2012 Location: Incline Viallge, NV ![]() | I’ve been working on OV tests as well as rigorous backtesting of strategies (in OT, Tradestation, etc.). Steve, I was happy to hear you brought this (measurement topic) up (again) and created a section in the Forum to get a discussion going. A recent performance measurement article I read (re: Measurement, Alpha Ratio, attached) may be relevant to this discussion IMHO. More modern, advanced measurement metrics like this may help us even better evaluate, compare, and rank, (switch/balance) various strategies and portfolios - to have an even better tool in OV. See what you think. Regards, Eric Severance ![]() | ||
^ Top | |||
Eric Severance![]() Posts: 23 Joined: 10/11/2012 Location: Incline Viallge, NV ![]() | A few years ago, in my backtesting, etc., I had been looking for a good way to summarize performance. Sunny Harris, a real backtesting guru, developed what she called the CPC Index - a good tool IMHO, that I’ve been using ever since. The CPC Index (from Sunny Harris’ book Tradestation Made Easy): I attached the chapter (4th page has formula, etc.) from her book where she talks about this… But here’s the formula: CPC = P * PF * R Where: P = Percent Profitable PF = Profit Factor R = Ratio Avg Win/Avg Loss Profit Factor is calculated by dividing Gross Profit by Gross Loss. By definition, a value greater than 1 means the trades have a positive net profit. A CPC score above 1.2 is considered “safe”. Over 2, better. As an example, I ran the calculation on a test of SPY (long only) on a RTM strategy of mine (below): came out well (CPC score of 2.98), which correlates well with this strat’s equity curve, etc.. *** Multiple ways to use this. For one, imagine if we could have the CPC score as a metric. as a tool to rank Portfolios and strategies on, etc. Maybe it could also be run, say once a week or month, on each symbol in a standard list (i.e. SP500), for XYZ strategy, then take the top CPC scoring ones and use those as a dynamic list! See what you think. Another useful performance measurement and comparison tool is the “Student T” test. Perry Kaufman speaks of it in his book on Trading Systems. Can send if interested. Regards, Eric Severance ![]() ![]() | ||
^ Top | |||
Jim Dean![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() | Thanks, Eric - I'd be very pleased to see CPC added to the metrics. Also, I can see some nifty uses for the "dependency" metric mentioned in the article. | ||
^ Top | |||
DBokhart![]() Posts: 3 Joined: 10/11/2012 Location: Dallas ![]() | As a forum newby how do I attach files to a post, cant' find info Thanks Duane Bokhart | ||
^ Top | |||
Jim Dean![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() | Down near the bottom of the post-creation window, you'll see a bolded-message checkbox "Attach a file to this thread or posting" Check it, then when you click Submit, follow the instructions browse, submit, then back. | ||
^ Top | |||
Steve Mayo![]() Posts: 414 Joined: 10/11/2012 Location: Austin, TX ![]() | Duane, once you have saved a posting with an attachment, you can then get the link for the attachment, then edit the post and use the formatting codes to embed the image in the article text, like this. open-square-braket img=http://www.omnitrader.com/currentclients/omnivestforum/get-attachment.asp?attachmentid=xxxx closed-square-bracket It's a good idea to first downsize your image (max of about 700 pixels wide) before uploading. | ||
^ Top | |||
Pete Taylor![]() Posts: 28 Joined: 10/11/2012 Location: N Bradley, England ![]() | I have previously done some research into R multiples when reading books by Van Tharp. He basically says that a trading system or Portfolio in our example should be ranked in terms of R multiples or Risk multiples. By doing this you gain an expectancy of what your system or our portfolios should achieve over some future time. He states that the more time or number of trades that are calculated by the system or portfolio, the more accurate the R multiple expectations will be. I was wondering if we could use this to calculate that the R multiples as a way of ranking them for the switching process. As the R multiple is changing all the time as the newer trades analysed coming into the system, we could rank one portfolio against others using this. This could be done to highlight which portfolios are becoming “good” against others by analysing the change over the last few weeks or a month. Obviously we want to get in & trade portfolios that are becoming good, rather than looking back several months later & wishing that we had. There is a video link below, of him explaining this on YouTube but you will notice that he refers to you initial stop as your 1R or 1 unit of risk. However, we do not have a stop on any of the typical strategies that we use. So in his book he suggests that if you don’t have an initial stop but you have a list of previous trades, then you calculate your 1R by just dividing the total amount of losses incurred in a system by the number of losses, which we could easily do. As you will see in the video you then express all you trades, losses and profits in R multiples, for example -1R, +1.25R etc etc. You would then add up the total trades for say the last month and the ones with the highest R multiples could be the ones to use. I would suggest this could be done on a rolling weekly basis over say the last 4 weeks, rather than waiting to the end of the month, as in Ed’s example in the presentation. I know its a pretty basic and easily calculated (even for me!) but thought it would just highlight changes in performance, which all we want to do. Don’t know if this is of any interest but thought it was a little different approach from the normal ones that you immediately think of. If you think that this could be of interest I can post a real example from a portfolio. Youtube link http://www.youtube.com/watch?v=JFuF6M7z1jg | ||
^ Top | |||
Jim Dean![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() | Pete's suggestion is simple, and a good idea ... rather than a moving SMA type window, I'd suggest the averaging of losses and wins should be done using EMA iteration for efficiency and to assure adequate statistical basis. Doing this would mean that the "AvgR" would be a function of the time window used, even for overlapping windows ... but all the other metrics are similarly affected. The equation is very simple: Expectancy = EMA(all prof/loss of all trades) / EMA(all losses) ... of course the EMA's iterate only with each new data item, not with each new bar ========== This would be BETTER, if the risk-denominator was not comprised solely of the average of the Losses (when/if/as they occur ... could be zero or statistically insignificant). The BETTER method would be to track Max Adverse Excursion within any trade, vs Entry price (which quite often will NOT be zero). This is a bar-by-bar calc done during the Strat analysis batch run, which I hope would be reasonable to implement. =============== BEST method: the true "at Risk" value is not solely defined by what happens at Entry, or versus Entry (depending on what stops if any the Strat uses) ... rather, it's the Max Profitable Excursion (ie best price hit during the trade), minus the Max Adverse Excursion that occurs away from that price, SUBSEQUENT to that price. A little messier to calc but a lot better representation of how well the trade is "managed" and how the vagaries of different symbols' volatilities impact how "comfortable" they are to trade. I've always felt that this is the best possible evaluation of the true Risk during the trade, IF no stops are set. Of course, if stops are set, the closest active stop, vs the farthest the price gets from it on any given bar, is the proper distinction to use. I hope that I've not caused anyone's eyes to glaze over. THIS IS A SIMPLE CALCULATION ... I'll provide code for it (just a few lines) if needed. The "barrier" to doing it would really only be whether that bar-by-bar analysis can be shoehorned in to the existing Strategy processing. It would not eat a bunch of CPU time, if bar by bar is currently being done. I'm unclear, these days, if the Strat calc process even HAS a bar-by-bar Mark-to-Market granularity or not, but it certainly will have to if OVest is ever going to allow spec of Fixed Loss stops or other kinds of stops. And, please, even though this does not come with some big well known name or abbreviation or article or book attached to it, I hope that it will be given at least 15 min of careful consideration. | ||
^ Top | |||
Steve Mayo![]() Posts: 414 Joined: 10/11/2012 Location: Austin, TX ![]() | Interesting idea, Pete. Thanks for suggesting it. Its a question for the programmers, but I THINK that the only historic data that will be available to the port switcher/balancer is the mark-to-market equity curve (EOD equity), not the individual trade data. If I'm right, then it wouldn't be able to average individual trade R's as Van Tharp suggests. But, I'm certain it could do composite reward:risk measures on the mark-to-market equity. the kind of metrics someone would use to compare two mutual funds, like Calmar, Sharpe, Sortino, Ulcer Index, Info Ratio, Alpha/Beta, etc. I know Jim and others have requested intraday stops and similar risk-controls/metrics be added to the live processing cycle. The difficulty on that, I would think, is the same as going to real-time. Being able to process all the user accounts fast enough. I think Nirvana will get there, it just might take a bit more time and cash flow. :-) | ||
^ Top | |||
Jim Dean![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() | Please note ... as long as we LEAVE OUT any discussion about possible future loss-stops, etc, and just go with what is happening DURING a trade, then I believe that if the teensy extra calc was shoehorned in to the Strat loop calc, and then those results were PROPAGATED as a table, through the portfolio combination process, then all that info would be available as an "Expectancy" for the Portfolio. Thus it would be static data by the time PortWiz/Bal got its hands on it. So, maybe this is doable, without major surgery. It would be VERY powerful, if for no other reason that it's so easy to understand for non-statisticians. | ||
^ Top | |||
kmcintyre![]() Posts: 890 Joined: 10/11/2012 Location: Portland, OR ![]() | Steve, Just back from a week RVing so I'm late to this party. My input is not so much a risk/reward ratio metric, but general observations and experience. I used to use TC to rank mutual funds by Linear Regression slope. Then I would look at the volatility associated with the top 10 performers. (I used 63 bars - or about a quarter's time - for the LRSlope.) I would switch to the best performing funds that seemed relatively less volatile. So somewhat scientific, somewhat subjective. But that's not really a risk/reward measure. But it worked pretty well for keeping me in better-than-average performing funds. Lately I've been really thinking about my psychological profile as a trader. Much like planning a hike, it's not only about elevation gain and loss, but also about trip duration, miles per day, etc. Risk/Reward is an instantaneous measure. I am interested in a measure of relative difficulty trading over time. I want to account for depth of drawdowns AND durations of drawdowns. Like Pavlov's dog, I want immediate gratification, and I'm willing to give up a little return to get it. So I would like to rotate between portfolios that are giving me a minimum rate of return (as measured by LR Slope) while not making me endure too much pain for too long. (If I can't make a minimum rate of return, hold cash...) I'm really looking forward to the Excel export feature as I would like to work on a risk indicator that captures both depth and duration. I'm thinking something like - tradability index = 1 / (( MDD * X ) + ( AAD * Y ) + ( MBH * Z )) where X, Y, and Z are coefficients to balance the weighting of each factor and MBH is the Max Bars between new equity Highs. (Or maybe ABH for Average Bars between new equity Highs...) [KCM 4/12/14 - the above could also include a measure of volatility, such as the MAX(Bolinger Band width[nbars]). Also, MDD could be over a long period while AAD[nbars] could be used.] Then I would solve for LRSlope * tradability index [KCM 4/12/14 - changed the formula a bit, inverting tradability index so it reflects a positive inclination (vs an untradability index). Used LRSlope versus CAR...] Again, not risk/reward, but an "ulcer index" that meets my trading profile. Not sure if that helps at all... Cheers Keith | ||
^ Top | |||
Mark Holstius![]() Posts: 744 Joined: 10/11/2012 Location: Sleepy Hollow, IL ![]() | Good points Keith... and this market (and year coming up?) will probably test that "pain threshold". I suspect you'd need to use the ADD and ABH so you're not thrown off by outliers, but I look forward to your results when the download capability is available. Please keep us posted. Mark | ||
^ Top | |||
kmcintyre![]() Posts: 890 Joined: 10/11/2012 Location: Portland, OR ![]() | It dawned on me that I didn't explain why I prefer Linear Regression Slope to CAR or a % gain measurement... CAR (or % gain) is an instantaneous measure that relies on a single start and end point. The LR slope is a measure of trend that relies on (and reflects) all points within the time range specified. Points seldom lie on the LR line. If the start point is below the line and the end point in above the line, CAR is overstated. If the start point is above the line and the end point is below the line, CAR is understated. CAR will seldom reflect the true trend. The shorter the time frame, the more volatile the market, the more chance there is that CAR (or % change will misrepresent the trend. LR Slope is far better IMO. (Use log scale so the slope represents % change, not absolute change...) Cheers! Keith | ||
^ Top | |||
Steve Mayo![]() Posts: 414 Joined: 10/11/2012 Location: Austin, TX ![]() | The "Scoring Function" that determines which port to use for the next period will definitely allow applying a linreg -- it was a separate step that I called score transformation in our Excel mockup but it is a native part of Omnilanguage so much easier to implement. After you compute a bunch of possible stitched ports, using a regression line to "evaluate" the permuted stitched ports (to determine which to use in your "switching system") might be a good idea too. Thanks for suggesting that one, Keith. The issue, however, might be how long do you lookback and do you need to compensate for low R-squared values? Looking a a port at the end of, say, 2007-2009, you might have a "U" shape. Depending on what day you measure, you might get a good slope but the correlation will be close to zero. Wouldn't this get into, therefore, needing to give uses a lookback and R2 threshold setting -- that gets too complicated for the causal user, I fear. | ||
^ Top | |||
kmcintyre![]() Posts: 890 Joined: 10/11/2012 Location: Portland, OR ![]() | You are correct in sighting that LR would not be a great tool for determining the trend between 2007-2009. But I think the issue is not with LR, but the fact that there was no single trend to be determined between 2007 and 2009. LR analysis would be very useful in determining that the trend was changing from a strong downtrend to an strong uptrend. The LR slope changed radically during that period. The problem you have sighted is due to sample rate, which I have pointed out in a previous post (Nyquist theorem...) You can't detect a two year trend with a sample rate of two years. Antialiasing occurs. [KCM - put another way, one can't determine what happened during a cycle with a 2 year period by sampling at a rate of 2 years. If one wants to detect 2 year cycles a minimum sample rate would be 1 year. That would not provide much resolution, but would capture the U trajectory the market took at a very gross level.] Anyway, my preference is to use LR slope verses a point-to-point measure such as CAR or % change. Looking forward to Portfolio Switching. Sounds very powerful! Cheers Keith | ||
^ Top | |||
Jez![]() Posts: 7 Joined: 3/3/2014 ![]() | Guys, I can't help much with risk-return measures but a few things I'd like to know about a strategy is what is the expected outcome, worst outcome, best outcome and the probability of a large string of losses. | ||
^ Top | |||
Steve Mayo![]() Posts: 414 Joined: 10/11/2012 Location: Austin, TX ![]() | Thanks Jez, Yes, that's along what I was thinking too. Perhaps rolling 1|3|6|12 month returns with min/max and standard deviation....or a frequency distribution on the monthly returns, or maybe a box plot using quartiles or standard deviations. | ||
^ Top | |||
Ger![]() Posts: 62 Joined: 10/11/2012 Location: Austin, TX ![]() | Question: Has anyone considered an exponential moving average EMA for ranking portfolios or even for strategies. For Portfolios, the input data would be the percent profit (loss) of every completed trade taken be the sum of all strategies. For multiple closed trades on any given day, one may have to calculate an average profit or loss for that day. For strategy ranking, one could use the same type of data, ie percent profit (loss) for calculating an EMA for the individual strategy. Advantage would be the the higher weighting applied to the nearest data. Any thoughts? Gerry | ||
^ Top | |||
Steve Mayo![]() Posts: 414 Joined: 10/11/2012 Location: Austin, TX ![]() | You lost me Gerry. Can you explain? How does that differ from these? Expectation of Loss= %Losses (1- hit rate) * average of the period (monthly/weekly) losses Expectation of Gain = %Wins (hit rate) * average of the period gains Volatility-adjusted return{ = avg. period return / stdev of avg period return Risk-adjusted return = avg. period return /% losses * avg period loss Remember, this is the EVALUATION function -- for deciding which permutation to select for walk-forward testing...and it only has the EOD equity to work with. Conversely, the SWITCHING function is standard OmniScript and it can do much more complex calculations. | ||
^ Top | |||
gbarber![]() Posts: 282 Joined: 12/30/2012 Location: Pearland, TX ![]() | I still like expected outcome (risk/reward) versus achieved outcome. This seems to me to be the gold standard of measuring a system. It shows you what was actually achieved versus what was promised. That is the end game. We all want to make money when the signal says that is supposed to be accomplished. I think that overrides all the other derivative measures that relate to the end game but don't state it as forthrightly as expected versus achieved. | ||
^ Top | |||
Ger![]() Posts: 62 Joined: 10/11/2012 Location: Austin, TX ![]() | Steve, Apologize for confusing note. My thoughts were very similar to that in your note, except I was looking for a way to give a weighting to the data on the hard right edge. That is, some type of time weighting like EMA rather than a simple average applied over the period. This would probably only work for the switching function evaluation since this requires more than just the EOD equity data. just a thought, Gerry | ||
^ Top | |||
Jim Dean![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() | Ger - I think your point is well-taken. An EMA or a WMA or a hybrid (bounded EMA or WMA with lowest wtg >1) might be more representative than an SMA. You can apply any of the averaging methods easily to Equity data. | ||
^ Top | |||
OCC![]() Posts: 85 Joined: 3/7/2013 ![]() | Answer: SQN (System Quality Number) Van Tharpe's SQN encompasses Return/Risk * Opportunity! Comments: Unfortunately, too many academics and pundits forget the basic concept of "opportunity" when measuring reward/risk. Essentially, a fantastic reward/risk system, perhaps with awesome %Win rate, that rarely trades, might not produce anywhere near the same profit rewards as a lower reward/risk system with a greater number of opportunities to trade. I'm a "total absolute return" thinker--and willing to make compromises to realize maximum profits. So, for sake of continuity, I'll repeat here some ideas posted about a year ago on this Forum: SQN (System Quality Number) together with its subset, risk-adjusted EXPECTANCY, would resolve many of the dilemmas about choosing and combining Strategies and Portfolios that bother others who have posted in the OV Forum. Van Tharp developed these metrics, thoroughly explained in his book Definitive Guide to Position Sizing. (Every trader should own it, and reread it at least annually.) For OmniVest, SQN and Expectancy have two excellent applications: 1) RANKING STRATEGIES, PORTFOLIOS In his books, courses, and seminars to develop Super Traders, Van Tharp teaches using SQN to evaluate and rank trading systems. It is a well-conceived statistical tool that encompasses rewards, losses, probabilities, open trade risk, frequency of trading opportunities, and sound statistical reasoning. 2) DEFINING MARKET TYPE Tharp’s free weekly trading newsletter presents monthly SQN rankings of key asset classes and markets worldwide, and explains his statistical methods to rank Market Type. 3) The most exciting application of SQN and Expectancy in OmniVest would be to create an automatic switch to activate (turn ON) PORTFOLIOS on when their current SQN level and trend indicates that it has moved into a favorable and promising Market Type for the purposes for which that Portfolio was designed and tested. Because of the multi-factor dimensions of SQN, such a switch might be more accurate and profitable to capture opportunities when they arise than any other method of activating Market Type customized Portfolios. (i.e., SQN and Expectancy signals could be both more timely, and more sustained, than using Equity Curve triggers.) 4) By using SQN to classify Market Type, as Van Tharp teaches, users could readily do their own research into the development of collections of Strategies into Portfolios that “fit” each environment as closely as possible. 5) Therefore, Nirvana would not have to divert resources to attempt historical testing of Market Type. 6) Users could have constantly updated, current evaluation tools. Thus, users would not have to rely on past fixed periods of research that will eventually become out of date and of diminishing usefulness. Furthermore, they would not have to accept somebody else’s definition of Market Type: they could define their own parameters or thresholds for entry/exit or strategy on/off switches. 7) EVALUATING TRADERS Since Van believes that any “holy grail” of trading (if there is one) lies inside each person, he also urges traders to use SQN and Expectancy to evaluate their real-world performance, assess systems, define trading skills, critique discipline, identify areas of errors that result in sub-optimal performance, and make necessary improvements. | ||
^ Top | |||
OCC![]() Posts: 85 Joined: 3/7/2013 ![]() | Calculating SQN and EXPECTANCY SQN (System Quality Number) calculation for systematic TRADES SQN = {[Expectancy / STDEVP(RM)] * SQRT(Number of Trades per Year)} where EXPECTANCY is another useful statistic to add to OmniVest (see below) STDEVP = Standard Deviation of a Population SQRT = Square Root RM = R Multiple = {(Profit or Loss of each Trade) / (Risk for each Trade)} “R” Risk for each Trade = Entry Price – initial Stop Price on that Trade (or, see substitutions in Notes) Notes: i) If Stops are not used (unwise), Risk/Trade = maximum adverse excursion (loss point) between Entry and Exit of each Trade ii) If Risk/Trade is unknown, or not calculable, use absolute value of Average(Losing Events) as “R” value iii) If Number of Trades per Year is unknown, substitute 100, of which the square root equals 10 SQN calculation for a PORTFOLIO or INDEX The Equity Curve defines a Portfolio’s overall “Index” of its performance, as a Index does for the group of securities it includes. Thus, Calculating SQN involves some simple, logical substitutions for variables. [It appears that Van Tharp similarly calculates SQN of Indexes to define overall Market Type classifications.] Notes: i) Define each DAY’s change in the Portfolio’s equity curve (or Index) as a Gain (Up) or Loss (Down) “Event” ii) Use the past N Days as Number of Trades (Van Tharp calculates several, but emphasizes 100 trading days to determine Market Type classifications reported in his research papers, books, and newsletters) iii) Use absolute value of Average(Losing Events), or Down days, as the “R” value Risk/Event iii) Compute Average Winning and Losing “Events” (days), and %Win/Loss probabilities in the Expectancy formula using “Event Days” in place of “Trades”. EXPECTANCY (a/k/a Average “R” in Van Tharp’s writings) EXPECTANCY = {[Average(Winning Events) * %Wins] - [ |Average(Losing Events)| * %Losses ]} divided by Average Risk/Event Notes: i) An “Event” is a “Trade” for a trading system, and Risk/Event is the maximum loss on initial stop. ii) For an Index or Equity Curve, an “Event” is each DAY’s Gain (Up) or Loss (Down) from the prior day. ii) If Risk/Event is unknown, or not calculable, use absolute value of Average(Losing Events) as divisor “R” iii) Importantly, some other definitions of “Expectancy” elsewhere are defective and incomplete, because they omit the devisor (Average Risk/Event). Therefore, they present an unwarranted, optimistically distorted expectations of profits, because they fail to adjust for Risk. | ||
^ Top | |||
kmcintyre![]() Posts: 890 Joined: 10/11/2012 Location: Portland, OR ![]() | OCC - love it! That book has just been placed on my todo list... (Although I don't know why Van Tharp hasn't incorporated time into the calculus. It one thing we all have a limited quantity of. Time is money, as they say...) Hopefully OV Pro will give us the ability to develop and backtest SQN. Cheers! Keith $240 from Van. $285 at Amazon. No Kindle version. Ouch! | ||
^ Top | |||
Matt B![]() Posts: 105 Joined: 10/11/2012 Location: Austin ![]() | Thanks for the comprehensive posts OCC. I agree completely with using SQN. Matt | ||
^ Top | |||
Jim Dean![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() | If SQN is implemented, be SURE not to label it as "SQN" since I believe that is trademarked. It might be that Van Tharp would ask significant licensing payment for using it ... his hourly rate for consultation, outside any of the Super Trader programs, is (drum roll) $1000/hour. I kid you not. | ||
^ Top | |||
Matt B![]() Posts: 105 Joined: 10/11/2012 Location: Austin ![]() | It is indeed trademarked, but if labeled and attributed as such, perhaps that gets around the usage. There is actually another statistical name for it that I don't recall. Maybe Z score? Maybe one of the statisticians here can help out. Matt | ||
^ Top | |||
JimB![]() Posts: 36 Joined: 3/7/2014 Location: Rogers, MN ![]() | Jim, ...and that hour will likely be the most expensive nap one will ever have! | ||
^ Top | |||
Jim Dean![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() | Yes ... but it will be a grateful, centered, self-actualizing nap! | ||
^ Top | |||
Geoff![]() Posts: 180 Joined: 12/4/2012 Location: Byron Bay NSW Australia ![]() | Calmar, MDD, CAR. In that order. | ||
^ Top | |||
Steve2![]() Posts: 750 Joined: 10/11/2012 Location: Annapolis, MD ![]() | Steve, I have a slightly different way of looking at things. I value consistency of positive returns more than anything else. I believe that if a portfolio simulation can deliver positive returns over all periods of the simulation date range then it has a higher likelihood of delivering positive returns in the future (assuming your simulation period is long enough to cover many different market conditions). To this end, I would like to see the following risk-return measure supported: The user specifies a Period (number of calendar days or number of market days) and whether or not P/L calculations should be Realized or Unrealized. The simulation date range is divided into Periods and for each Period a determination is made as to whether or not the Period is profitable. A count is made of the number of losing Periods and the maximum number of consecutive losing Periods. Portfolios are then ranked by: 1. Minimum number of losing Periods 2. Minimum number of consecutive losing Periods 3. Maximum CAR Steve | ||
^ Top | |||
Jim Dean![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() | Hi, Steve(2) I like your approach a lot. It's sensible, and psychologically satisfying. I might well choose to use a Port that has a lower CAR, if the "score" you are describing is high. Suggestion: optionally allow for the various periods-slice results to be forward-weighted somewhat, so that a case with 80% periods profitable with the 20% unprofitable near the left, will show up as worse than 80% profitable, with the 20% profitable near the right. | ||
^ Top | |||
Steve2![]() Posts: 750 Joined: 10/11/2012 Location: Annapolis, MD ![]() | Jim, I think that would be a sensible option to have although I think one of the problems we currently have with long simulation date ranges is that they get unduly influenced on the right by large account sizes. Typically, your percent invested goes down and performance changes, not because of market conditions but due to effects from trading a $100M+ portfolio. I'd really like to see a simulation option that would cap account size on an annual basis. I don't find the ending equity number very useful (except in my dreams)and would much rather see the other metrics influenced only by market conditions. So, I'm always breaking a 14 year sim up into 4 year sims (which is a pain in the neck) to eliminate the impact of account size | ||
^ Top | |||
Jim Dean![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() | Yes I *strongly* agree re the cap on buying power. I suggested this very early in the V1 game ... both for modelling purposes, AND because (probably) the majority of traders will occasionally SPEND some of their earnings (or transfer it to other types of investments such as real estate, etc ... there should be an option to cap buying power. That is, the account size can continue to grow, but the max available Buying Power is not allowed to grow past a certain point. This would effectively solve many issues in comparing strategies, forcing the trade sizes to remain in the realm of reasonableness. In fact, an interesting simulation method would be to cap buying power at the INITIAL account size, and see how things work. That would put all the years in the simulation on the same playing field, and be a much better measure of Risk of Ruin. | ||
^ Top | |||
Steve2![]() Posts: 750 Joined: 10/11/2012 Location: Annapolis, MD ![]() | Yes, capping buying power is the right way to go! Is there a feature request in for that (hint, hint, if not :-)) | ||
^ Top | |||
Jim Dean![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() | I just posted one, again. If you think it's particularly important, email Ed | ||
^ Top | |||
Steve Mayo![]() Posts: 414 Joined: 10/11/2012 Location: Austin, TX ![]() | I think you guys are talking about "rolling return". I use that a lot. See my box plot under thread 6092. Rather than slicing up the return into, say, one-month chunks over 3 years, you roll-forward day-by-day over however many years to calculate all possible one-month returns for every possible starting date. Then, you can calculate min/max/mean/stdev/probability of positive return/probability of negative return. With a decent stat package, you can even do a chi-square analysis to compare the ports ... and probably find that most really aren't statistically distinguishable given the high standard deviation that is inherent in the stock market. | ||
^ Top | |||
Jim Dean![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() | Thanks for the suggestion Steve but no, that's not what I'm referring to. I would like a simple "Max Buying Power" input in Account settings. It should be able to be set to any value >= the initial account balance. I want to be able to run comparisons between strats and portfolios where the available funds (ie total dollars active in the account for trading - including existing positions) are the same (or less) every day. This normalizes the comparisons very effectively and very simply. All gains will remain in the account, so the equity curve can rise above that level, without limit. But the money available to trade needs to remain at or below the buying power limit. | ||
^ Top | |||
Steve Mayo![]() Posts: 414 Joined: 10/11/2012 Location: Austin, TX ![]() |
Sorry, Jim, I should have clarified I was referring to Steve2's posting about comparing chunked metrics. | ||
^ Top | |||
John W![]() Posts: 654 Joined: 10/11/2012 Location: Sydney, NSW, Australia ![]() | Originally written by 210921 on 4/25/2014 6:50 AM I would like a simple "Max Buying Power" input in Account settings. It should be able to be set to any value >= the initial account balance. QUOTE] I think this is a good suggestion, worthy of consideration. | ||
^ Top | |||
Jim Dean![]() Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA ![]() | if anyone wants to "vote" on this buying power cap thing, please post in this thread instead ... http://www.omnitrader.com/currentclients/omnivestforum/thread-view.asp?threadid=6103 That topic does not really fit the title or purpose of this thread. I'm sorry for getting things sidetracked. | ||
^ Top | |||
HClemmons![]() Posts: 9 Joined: 10/11/2012 ![]() | I have done extensive work in calculating the "expected value" of the trades within a strategy based on a grouping of the most recently completed trades. I have used this to rank the strategies and select the strategy with the highest expected trade return. Since the average trade return plays a very important role in determining the overall portfolio performance(along with trade allocation percent and number of trades), the process has worked very well. When the same symbol is traded in multiple strategies, I look to the strategy with the highest expected return to determine whether the trade should be taken. Care must be taken with certain strategies, as the expected return may look good, but the variance can be quite high and result in poor trade results. | ||
^ Top | |||
Mark Holstius![]() Posts: 744 Joined: 10/11/2012 Location: Sleepy Hollow, IL ![]() | Most interesting (HClemmons) - do you use the recent trades to find a "target / average exit" or something similar for the expected return? Mark | ||
^ Top | |||
HClemmons![]() Posts: 9 Joined: 10/11/2012 ![]() | Operating off maximizing slope of the line between beginning and ending MV. Maximizing m, is the same as maximizing: number of trades * average allocation percent * average trade return, where trade return is calculated using (trade income) / (beginning MV * average allocation percent). I am looking at the expected returns of the strategies recent trades, and picking the "best" expected return(not always the highest because of variance and number of recent trades to pick from). Use the trades from the selected strategy and let OV tell me when to exit. Setting a minimum expected value keeps you out of the strategy when you get a series of negative trades. For most strategies, I can get a similar return over a period of time with fewer trades (which is a result of higher return per trade). Works similar to the equity curve analysis, since you are avoiding the strategy (portfolio) when the recent results are not favorable. | ||
^ Top | |||
Steve Mayo![]() Posts: 414 Joined: 10/11/2012 Location: Austin, TX ![]() | That's interesting, Harold. Can you clarify a bit. It sounds like you are evaluating STRATEGIES and then selecting a PORTFOLIO containing that selected strategy? Are you trading only one single-strategy portfolio at any one time? | ||
^ Top | |||
Steve2![]() Posts: 750 Joined: 10/11/2012 Location: Annapolis, MD ![]() | Steve, While "rolling return" would also be an interesting measure, I prefer to look at fixed, non-overlapping periods. It's a simple way to understand how quickly draw downs are recovered and how consistent positive profits are achieved. I typically use quarterly periods as that aligns well with my risk tolerance. Steve | ||
^ Top | |||
HClemmons![]() Posts: 9 Joined: 10/11/2012 ![]() | You are correct that I am evaluating strategies. Based on available amounts to invest, I will start with the strategy with the highest expected return, and select trades from that strategy. My allocation % is high enough, that the process normally only needs one strategy per trading day. However, if the top strategy only has one eligible trade, and I have funds for three trades, it will go to the next highest strategy (if expected return is high enough), and get trades from the next strategy. If not, I stop at the one trade selected and wait for the next trading day. Ever changing PORTFOLIO based on expected return of the STRATEGIES. | ||
^ Top | |||
Mark Holstius![]() Posts: 744 Joined: 10/11/2012 Location: Sleepy Hollow, IL ![]() | Thanks Harold - do you have any statistics on your returns doing this? Mark | ||
^ Top | |||
Eric Severance![]() Posts: 23 Joined: 10/11/2012 Location: Incline Viallge, NV ![]() | I too have been asking (since V1) for a non compounding method to measure returns.... An Annual Rate of Return (non compounding). So one can better compare strats, ports, etc. Hedge funds are often measured this way. The compounding "hockey stick" effect is throwing things off and I fear will really trip up novices...IMHO Thx, Eric |