How volatile does a stock need to be?

..before trading it with a stop-loss strategy will generate higher returns than a buy/hold strategy? For instance, say you have a basic strategy of planning to take a position everytime a certain equity gains 1 dollar, and sell once it loses 1 dollar. Once sold, you will not purchase again until the equity gains 1 dollar again. This strategy allows you to hold as the stock gains value, and be out for long downswings.

Two problems: the first is that volatility needs to be at a certain level. Obviously if the equity you buy only swings back and forth by a dollar, you will get killed. You will buy the highs and sell the lows. However if it swings by 5 dollars, you will make significantly more than buy/and hold.

The next problem is transaction costs. Buy/hold costs basically nothing..one purchase, one sale. The stop/loss strategy will have greater transaction costs that depend on the frequency of swings within the stoploss. Gains from variance have to make up for this.

...To that end, using an excel spreadsheet, I've shown that given a transaction cost of 10 dollars a trade, and graphing the NASDAQ over the last 3 years. A stoploss strategy can beat the market consistantly by about 5%.

Where are the flaws in this that I'm missing? It seems too simple to work.

Thanks..Matt

[Dariel86]

Beating the market for 5% isn't really that great...

There has been talk recently about how most actively managed mutual funds cannot beat the index. Beating the index consistantly does not seem easy to do. Remember that this technique, if used with a computer program, does not require research or attention.

But other than the small return, you can't find other problems with this?

One problem I see is that volatility will not stay the same year to year...so the correct value of the "stoploss" changes as well. However, there may be a range you can stay in with a reasonable degree of confidence.

[Officer Farva]

Welcome to the beauty of backtesting: you can make any strategy work, especially parameterized ones. There are infinite strategies that "beat the market" looking back simply becuase they are built to obtain historical performance without regards for the spontaneity of the future market. If your stop-loss strategy happens to work for you, more power to you, but please do not confuse risk manipulation with true alpha.

Thanks for posting! I appreciate what you're saying, in that it is easy to find a strategy that extracts alpha in hindsight. However, I believe past volatility may be an indicator, to a point, of future volatility in an index. Obviously I don't know which way it will swing, but if I know it will swing AT LEAST a certain amount in either direction, I can produce alpha.

For instance, if this technique only needed the NASDAQ index to swing more than 5 dollars to produce alpha, I think we'd both agree that it will work in 2006. Since this scheme needs a larger swing than 5 dollars, my question is whether a large enough volatility can be predicted with a high degree of confience to make this type of "smoothing" strategy work. Any thoughts?

[Sniper]

[ QUOTE ]
Where are the flaws in this that I'm missing? It seems too simple to work.

[/ QUOTE ]

There are many simple strategies that work for an individual investor, that simply cannot be done by a mutual fund manager that must remain fully invested.

[Dan Mezick]

Stop-losses are proven to dampen absolute return in winning systems. The reality is, few traders are willing to trade anything approaching optimal bet size without a stop loss. This is usually because the optimal size (without a stop) can produce a catastrophic loss. Stops combine with bet size to define risk and provide some insurance against unpredictable shock events.

You might want to look into the following related subjects:
Average True Range
..ATR provides a way to size a bet based on volatility. One application is to bet more when ATR is small, less when ATR is big.


Optimal f
...Optimal f addresses optimal bet sizing. Few traders have the sack to bet the amount Ralph Vince's Optimal-f says to bet.

You have to be a real Kamikaze to do that.

Thanks for all the responses. It sounds as though this is a technique which may work, though only for smaller individual accounts. Aside from the very real possibility that this is only doable with "backtesting", I keep wondering if there is any way to approach a MINIMUM prediction for the volatility of an equity or index over the next year. Say with 95% confidence.

To smooth that volatility successfully, the transaction costs of the trading strategy have to be low enough not to cancel out the gains from using a stoploss over buy/hold. The higher the expected volatility, the more profit, and fewer trades.

Do traders take a strategy like this one? I've heard of the technique using options to take advantage of expected volatility. Buy a call and a put, and assume the loss from one purchase will be offset by gains from the opposite one. The problem there is timing. If the option expires before the expected volatility occurs, the trader is out of luck.

The stoploss strategy does not depend on timing, only expected volatility. However, the stoploss strategy does not profit from downswings in the equity, it merely seeks to minimize them.

So all in all, is there a way to formulate the level of volatility needed to make the stoploss technique that I outlined viable? Something like, for X expected volatility and X cost per transaction, what, if any, is the expected gain over and above the buy/hold strategy?

Put a more interesting way, assuming you need a 10% return over and above the index, and you have transaction costs of 10 dollars a trade, how much expected volatility is necessary to succeed with the strategy I've outlined?

I define volatility here as the absolute value of the movements of a stock over a given time period divided by their frequency. There is probably a better way

Any thoughts?

[Dan Mezick]

If all you want to do is buy and sell volatility, learn the Options game and learn it well. That's the place to most easily accomplish what you are setting out to do.