Just a quick qestion: how much do you diversify, 10, 20 or 100 Stocks?
Are there others like me that have 100 Stocks in their Portfolio?
Why do I have 100 small and micro caps in my trading portfolio?
→ very low volatility
→ relativ low correlation with the market
→ Not afraid of trading, since I know, that even if my Modell picks the
wrong stock, it will be Offset, a “mistake” (Even if the stock looses 50%) does
not kill you and there are allways big runners that Offset the big loosers very well.
I also do but this represents all cap sizes and most asset classes. I do it for the same reason that you mention as well as a reduced yearly standard deviation in my potential band of returns, psychologically easier to continue to invest in one of many ports that is fundamentally sound but underperforming due to market conditions, as well as secondary safeguard to minimizing my risk of data mining. This works well for me.
Jim,
I would concentrate more if I had an objective mechanism that was forward predicting that would tell me which system(s) to concentrate into however I have not found evidence for such a system and as a personal preference I like a little less bumpy ride.
I do not want to hijack this thread but am curious regarding this comment “Removing a strategy and adding a new strategy this weekend so a bit of figuring to do.” This can be challenging if you are changing from good/decent to better. What criteria do you use?
Hi Andreas,
I run 5 models with 5 stocks each, so 25 holdings in total. All models are equal weight, two of them defensive large cap models and three more specialised in small caps with low correlation to each other. I only started this strategy in mid February and I’m up already 13% so far.
With the portfolio growing I will allocate more capital to the large caps, but also add one or two small/micro cap models. I doubt I will ever trade more than 40 stocks - just too much work for me…
I see how your large cap and small cap models might have mutually exclusive universes.
What do you do when all 3 (or 2) of your small caps recommend the same stock.?
BTW, I have 31 stocks in my portfolio. I have limited the concentration of stocks by making some of the universes mutually exclusive. But I do not reject any recommendations. I get too concentrated in a few stocks due to repeat recommendations. Even so, if my most concentrated stock went to zero, I would still be beating the SP500 for the year.
I’m paper trading some ports that will give less concentration but the return will certainly be less–but less volatile and more liquid.
I never buy the same stock in 2 or more Ports. If a Port recommends a stock I am holding I reject it and get the next lower ranked stock. If 2 Ports recommend the same stock on the same day I buy it in the better performing Port (That one knows what it is doing better? ), and I get the next best stock in the other Port.
74 stocks across 8 portfolios each with 20 stocks, so some common holdings across portfolios but no holding across the aggregated portfolio of more than 4.5%. Cap weighting is highly diversified/evenly dispersed across total portfolio and no one sector weighting makes up more than 30% . This concentration in certain names has proven to be fruitful as many of the concentrated positions have done very well. Diversification - across cap and sector spectra - is by design to dampen volatility within the portfolio. I seek mid double digit turnover in the overall portfolio so focus on sell rules has been a point of attention for me.
Enjoying the ride so far. Glad I stumbled upon P123
I currently run a book of 4 ports, with 100 holdings. Due to overlap, this translates to about 80 names. While I can see why others reject duplicates, I don’t mind it, because if multiple strategies recommend the same stock, I would prefer to be overweight.
Current allocation is 75% small and micro, with the rest in large and mid cap. Though, both cap and sector weights are mainly a residual of the bottom-up stock selection strategies. Average holding period is between 60-100 days. I used to invest in other R2G strategies, but now only allocate to my own ports.
The total book is long only, comprised only of stocks (no ETFs), and excludes hedging or market timing with 100% invested at all times. My P123 book represents a significant portion of my total retirement portfolio, and I allocate to other asset classes (i.e., international equities, high yield bonds, emerging market debt, total bonds, REITs, commodities) for diversification and risk reduction. I would probably allocate more to P123 strategies, but ran out of room in my Roth IRA, and the rest is in my 401k.
Two ports with some similarities in their models. One holds 30 stocks with a minimum 2MM AvgDailyTot(60) CPI adjusted (will increase the minimum over time) and is 29% large cap, 50% mid cap, and 21% small. The other holds 10 stocks with minimum 5MM ADT and is 50% large cap and 50% mid cap. Currently 6 stocks are common to both due to the model similarities. I don’t hold microcaps or low turnover stocks.
I diversify with separate holdings in an index ETF, two mutual funds that I will probably liquidate, I bonds held for 13.5 years, real estate, and an assortment of stocks held for many years.
~75 stocks in 7 ports. As you can see below (I have one port where I aggregate all for tracking purposes), I have spent a lot of time configuring the ports to give me the diversification in ranking system (all are quite different), stock picks (little overlap), sector (I tried to model on the SP500), and market cap (>50% large cap) that I want (conservative). I am only looking for ~15%/yr total return for the whole thing, but with less std dev and drawdown than the SP500. So far so good.
I noticed that one common theme is to create many ports that are less correlated, and therefore presumably narrowly focused. On the other hand, I prefer diversifying within the ranking system with multiple factors.
I like the multi-factor ranking system approach, because I prefer to purchase companies that are good (but not necessarily great) on multiple dimensions. On the other hand, running many individual portfolios that are mutually exclusive might result in purchasing companies that are great in one area (e.g., value), but not so great in another (e.g., momentum). I would assume that the combined book for this approach would have higher turnover than a multi-factor portfolio. Though the downside to a multi-factor and more diversified portfolio is lower returns (backtested).
Quad - multiple factors within a ranking system doesn’t help with diversification, unless you are using conditional nodes to do the diversification.
Steve
I have one port that has very high turnover. The only downside I can see is that it is particularly sensitive to higher than expected slippage. Fortunately the slippage so far–for my real trades–is less than for the sim. This may not always be the case as I may have more money to put into this port in the future.
I am testing a lower turnover (less sensitive to slippage) and a 10 stock version (actually less market-moving slippage) should this “problem” arise.
@jrnnie: If two models recommend the same stock, i buy both. But this rarely happens. In fact, it only happened once so far, and i sold half the position after 2 weeks to secure a 50% profit.
Wow, thank you for the great Responses! Nice to hear, that pretty big diversification is going on!
I hold 100 Stocks all equal weighted, all Micro to small caps.
No I would not Change to a 50 Stock Modell with higher Performance, since it buys usually with
higher vola (Since August 2013 the mdd was 4,8%!!!, I love that!!!).
(I changed from a three month rebalance to a 1 Week, but with relative low turnover of 600% per year. I did that because the Deviation
of a 3 month System is to high depending on the start date.)
Further more I trade 3 5 Stock modells but only with 3k a stock (e.g. 15k), since only then I seem not to deviate from
the modells due to psychological reasons.
What a fascinating thread! I’m working with a 4 asset book. All the assets are R2Gs with some of the R2Gs hedged with bonds and inverse ETFs. At this point, there are 40 open positions. The book is tilted towards HEALTHCARE and large cap stocks.
While the book is relatively new and needs more time to prove itself, I’m very happy with the results to date.
Attached is a screen capture of the equity curve.
Walter
EDIT: the equity curve is constructed with P123 estimated open/close prices.
Steve, I was referring to the fact that a multi-factor model is more diversified than a single factor model, all else being equal.
One can construct a ranking system with multiple factors, or construct multiple ranking systems of the same single factors. In theory, you arrive at the same exposures, but I was pointing out that the securities selected will be different, and I prefer selecting securities that rank well on multiple dimensions, but not necessarily the best in a single factor.
), and I get the next best stock in the other Port.
