Strategy Design Pattern

1
Strategy Design Pattern
2
Outline

• Example
• Forward Price and Delivery Price
• Forward Price and Spot Price
• C Classes
• Strategy Pattern
• Design
• Analysis
• Strategy and C Policy

3
Forward Price and Delivery Price

• The forward price is the market price that would
  be agreed to today for delivery in the future.
• For example, 1 ounce of gold costs 315 today to
  be delivered in one year.
• 315 becomes the delivery price of 1-year forward
  contract.
• As we move through time the delivery price does
  not change, but the forward price is likely to do
  so.
• In 6 months the 1-year forward price may change
  to 295 if the market conditions change.

4
Forward Price and Spot Price

• We illustrate the reason why spot and forward
  prices are related.
• We consider a forward contract on gold. Assume
  there are no storage costs associated with gold
  and that gold earns no income.
• Suppose the following
• The spot price of gold is 300 per ounce.
• The risk-free interest rate for investment
  lasting one year (e.g., CD in a bank) is 5.
• What is a reasonable 1-year forward price of gold?

5
Strategy 1 Buy

• Assume the forward price is 340.
• Trader's actions
• Borrow 300 at 5 for one year.
• Buy 1 ounce of gold for 300.
• Enter into a short forward contract to sell the
  gold for 340 in one year.
• At the end of one year, the trader needs to repay
  the loan for 315, which is used from 340 he
  gets from selling the gold. Hence the profit

6
Strategy 2 Sell

• Assume the one-year forward price is 300.
• An investor who has gold in portfolio can
• Sell the gold for 300 per ounce
• Invest the proceeds at 5
• Enter into a long forward contract to repurchase
  the gold in one year for 300 per ounce.
• The profit is

7
Arbitrage

• The first strategy is profitable when the
  one-year price forward price of gold is greater
  than 315.
• As more traders attempt to take advantage of this
  strategy, the demand for short forward contracts
  will increase and the one-year forward price will
  fall
• The second strategy is profitable when one-year
  forward price lt 315.
• The demand for long forward contract will push
  the forward price up.
• Assuming that individuals are willing to take
  advantage of arbitrage opportunities, the market
  will push the forward price to be 315.

8
Arbitrage Application

• Assume an application that deals with any
  arbitrage opportunities on market.
• The application is based on market listeners that
  prompt the user on any arbitrage opportunity.
• The listener would we triggered if a particular
  arbitrage strategy can be applied.
• Once the strategy can be applied, the listener
  would inform the user of actions that need to be
  taken, profit, etc.
• At the heart of such application will be an
  Arbitrage class.
• The Arbitrage class keeps the knowledge of a the
  strategy to be used.

9
Arbitrage C Classes
class Arbitrage public
Arbitrage(ArbitrageStrategy) void
run(double maturity) double rate
MarketgetRate(maturity) double spotPrice
// get from market double forwardPrice
// get from market this-gtstrategy-gtsetRate
(rate) this-gtstrategy-gtsetSpotPrice(spotPri
ce) this-gtstrategy-gtsetForwardPrice(forward
Price)
10
Arbitrage Class
double profit this-gtstrategy-gtgetProfit(ma
turity) if (profit gt 0) cout ltlt
"Actions" ltlt this-gtstrategy-gtgetActio
ns() ltlt endl ltlt "Profit" ltlt profit
ltlt endl protected
ArbitrageStrategy strategy string asset //
some asset
11
ArbitrageStrategy Class
class ArbitrageStrategy public virtual
string getActions() 0 virtual double
getProfit (double maturity) 0
virtual void setRate(double r)
this-rate r
12
ArbitrageStrategy (cont.)
virtual void setSpotPrice(double)
virtual void setForwardPrice(double)
protected double rate double
spotPrice double forwardPrice
13
BuyArbitrageStrategy Classes
class BuyArbitrageStrategy public
ArbitrageStrategy public virtual string
getActions() ... virtual double
getProfit(double maturity) return
this-gtforwardPrice this-gtspotPrice
pow(1this-gtrate,this-gtmaturity)
14
SellArbitrageStrategy Class
class SellArbitrageStrategy public
ArbitrageStrategy public ... virtual
double getProfit(double maturity)
return this-gtspotPricepow(1this-gtrate,t
his-gtmaturity) this-gtforwardPrice
// Client code Arbitrage a(new
BuyArbitrageStrategy()) a.run() ...
15
Strategy Pattern UML Diagram
16
Strategy Pattern Analysis

• Intent
• Define a family of algorithms, encapsulate each
  one, and make them interchangeable.
• Applicability
• many related classes differ only in their
  behavior
• different variants of an algorithm needed
• an algorithm uses data that clients should not
  know about.
• Participants
• Strategy (ArbitrageStrategy)
• declares a common interface to all supported
  algorithms.
• ConcreteStrategy (BuyArbitrageStrategy)
• implements the algorithm.
• Context (Arbitrage)
• uses Strategy.

17
Policy and Strategy
template ltclass Tgt class Arbitrage public
Arbitrage() strategy(new T()) void
run(double maturity) ...
this-gtstrategy-gtsetRate(rate)
this-gtstrategy-gtsetSpotPrice(spotPrice)
this-gtstrategy-gtsetForwardPrice(forwardPrice)
18
Policy Based Arbitrage Class
double profit this-gtstrategy-gtgetProfit(ma
turity) if (profit gt 0) cout ltlt
"Actions" ltlt this-gtstrategy-gtgetActio
ns() ltlt endl ltlt "Profit" ltlt profit
ltlt endl protected T
strategy string asset // some asset
19
Policy Implementation
class BuyArbitrageStrategy public
ArbitrageStrategy public virtual string
getActions() ... virtual double
getProfit(double maturity) return
this-gtforwardPrice this-gtspotPrice
pow(1this-gtrate,this-gtmaturity)
class SellArbitrageStrategy public
ArbitrageStrategy ...
20
Policy Implementation Usage
// Policy is the Strategy design pattern with //
the twist that policies are compile-time
bound! ArbitrageltBuyArbitrageStrategygt
a ArbitrageltSellArbitrageStrategygt
b a.run() b.run() ...