The science of quantitative methods for making business decisions has been around for decades. The benefits reaped in managerial decision analysis stretches across operations research involving transportation problems, waiting lines, inventory, shadow pricing, network modeling, forecasting, behavior simulation, Bayesian sample analysis, and mock experimental staging.
A relevant example to patent analysis and IP strategy of decision-making in action involving a probability analysis of a possible patent infringement compared to paying an $80 million licensing cost was outlined by Fine & Palmer in “Patents on Wall Street” from the book, From Ideas to Assets: Investing Wisely in Intellectual Property. In the example, a decision tree is used to assess the expected cost of infringement based on payoff probabilities to determine the potential cost to a party. Quantitative analysis and data from professionals and industry experts can be used to provide the necessary inputs to frame the IP management decisions.
The hypothetical scenario assumed that at the time when mobile phones were first invented, a fictitious company (let’s call them ABC) obtained a 17-year patent on mobile-phone technology. Another party (XYZ) wants to evaluate whether it should begin producing mobile phones without a license, or purchase an $80M license from ABC to produce the mobile phones. The decision tree (also called “extensive form” analysis) for the outcomes and their associated costs and benefits is shown below.
In a very clear, logical framework one can readily conclude that the most sensible decision is for XYZ to pay the license fee to ABC (assuming that XYZ’s economic projection for net sales of mobile phones is forecasted to remain positive). Obviously, the analysis is highly dependent on the input parameters – as the saying goes garbage in equals garbage out. The authors point out that changes in the probabilities and payoffs can change the preferred decision strategy significantly. For XYZ, if the probability of winning the lawsuit jumps from 10% to 20%, the outcomes for the whole scenario change. Similar decision trees can be constructed to include chances of patent invalidation (or annulment) during litigation and a host of other real-world contingencies relating to IP strategy.
One familiar with the stakes involved with intellectual property would consider playing games with IP to be a careless and risky behavior. The situation is quite different when talking about “paying games” applied to patent analysis. A more advanced form of decision theory involves more than one participant in an interactive decision-making process. The foundations of game theory assume that the players are rational decision makers looking for optimal payoffs with minimal risk in selecting an “act” to perform. An equilibrium act-pair (one for each player for two-player games) corresponds to acts from which there is no incentive for either player to deviate.
A dominant equilibrium is one in which a single act-pair remains in the payoff table after all inadmissible acts have been eliminated. Each player obtaining an outcome that is the “best of the possible worsts.” The more common Nash equilibrium is one in which neither player finds it advantageous to leave, unless the other player deviates. For zero-sum games, the largest payoff value in its column and the smallest in its row is a saddle point. The minimax principle indicates that each player makes a choice so that the maximum possible loss to the opponent is minimized.
A case study discussed at the University of Strasbourg’s Center for International Intellectual Property Studies (CEIPI) examined the pursuit of IP appropriation in which the game has several rounds – known as an iterated elimination. In the IP management case, two companies are trying to protect and market an invention. Company A is considering applying for a patent, or using secrecy as strategies for appropriation. Company A evaluates these options based on the available strategies for their competitor, Company B. Company B has the options of applying for a patent, using secrecy or using lead time as strategies for appropriation.
- If both Company A and B apply for a patent, the payoffs are 1 for Company A and 0 for Company B, since Company A will be faster in this scenario.
- If Company A applies for a patent, while Company B maintains secrecy, the payoff are 1 for Company A and 2 for Company B, since Company B will not incur significant costs for maintaining secrecy.
- If Company A applies for a patent, while Company B uses lead time, the payoff are 0 for Company A and 1 for Company B, since Company B will exploit the market sooner than Company A.
- If Company A maintains secrecy, while Company B applies for a patent, the payoff are 0 for Company A and 3 for Company B, since Company B will be able to prove infringement by Company A.
- If both Company A and B maintain secrecy, the payoffs are 0 for Company A and 1 for Company B, since Company B will have better market access.
- If Company A maintains secrecy, while Company B uses lead time, the payoff are 2 for Company A and 0 for Company B, since Company B will have a higher monopoly rent.
Such an IP appropriation-strategy game can be represented in “normal form” in a payoff table as follows.
For Company A, neither a patenting or secrecy strategy is dominated. That is, for Company A, patenting is better than secrecy, if Company B patents (1>0). Likewise, for Company A, secrecy is better than patenting, if Company B uses lead time (2>0). For Company B, lead time is strictly dominated by secrecy (2>1 and 1>0). Such an act is referred to as inadmissible – it will never be chosen because it is dominated by another act that is, under all circumstances (meaning all other options), at least as good, and, in at least one case, strictly better.
