THE USE OF DECISION TREES IN MEDIATION

Attorneys have long used probabilities for success as a rough tool for estimating potential damages in litigation.  A simple calculation would go something like: “we have approximately $100,000 in damages but only a seventy-five percent chance of winning the case.  Accordingly, the value of the case is $75,000.”  Such a simple calculation, however, does not fully and accurately take into account all of the possible outcomes of a case.  As a result, attorneys are increasingly turning to “decision trees” to evaluate cases by using weighted averages of all possible outcomes and their associated costs/values.

It should be made clear at the outset that a decision tree analysis can still be highly speculative.  Depending on the case, outcomes and damages can be difficult to predict.  The resolution of a single factual dispute may radically change the outcome of a trial, as may the make-up of a jury or a change in judges.  Decision trees are also dependent on the lawyer creating them including realistic expectations as to possible outcomes.  As the saying goes: “garbage in, garbage out.”  Properly developed, however, a decision tree analysis may assist in developing a basis for settlement goals either in advance of or during a mediation.

Daniel Klein’ article “Decision Trees and the Arboretum” states that there are four steps in creating a decision tree:

1. Listing the possible outcomes of litigation;
2. Considering the costs and/or gains associated with each possibility;
3. Discounting each possibility by an estimated probability of its occurrence; and
4. Multiplying each possibility by its probability to get an overall estimated case valuation.

Let’s start with listing the possible outcomes of litigation in a case which the only possible outcomes (other than settlement) are that it will be resolved at summary judgment or it will be decided at trial. Let’s further assume that we are creating the decision tree for the defendant. In addition, we will not consider the possibility of an appeal. Thus, the possible outcomes are:

1. Prevail with summary judgment motion;
2. Lose summary judgment motion;
a. Win at trial
b. Lose at trial

Note that the possible outcomes of win at trial and lose at trial come into play only as a consequence of losing the summary judgment motion.

Once the possible outcomes have been laid out, we must estimate the costs and/or gains associated with each possibility. For the purposes of this exercise, we will assume that attorney’s fees are not available to the defendant and that recoverable costs are negligible. In creating our decision tree, we will estimate that it will cost the client $35,000 in order to prevail on a summary judgment motion. It will cost $100,000 to win at trial. If we lose at trial the high end of the range of damages is $200,000. Adding the $100,000 in attorney’s fees results in a total cost of $300,000. The low end of possible damages is $30,000 resulting in a total cost of $130,000. The mid range of potential damages is $100,000 which results in a total cost of $200,000.

Perhaps the most critical element to creating a decision tree is estimating the probability of each possible outcome. My experience is that litigation attorneys are typically overconfident about their cases. If you ask two opposing attorneys what each believes is their likelihood of success, both may say seventy percent which means that, combined, they are overestimating their chances by a total of forty percent. Accordingly, in working with decision trees, attorneys must try to be brutally honest in their predictions. Alternatively, they might make an estimate and then reduce it by ten or fifteen percent to assure a reasonable outcome of the decision tree.

For the purposes of our exercise, let’s assume that the defendant has a 60% chance of prevailing on summary judgment. If it does not prevail on summary judgment, however, it only has a 50% chance of winning at trial. Finally, if it loses at trial, there is a 50% chance that the plaintiff will be awarded the medium level of damages, a 30% chance of receiving the low end of damages and a 20% chance of receiving the high end of damages.

There are several decision tree software programs on the market for purchase. A free resource for creating simple decision trees is “Occam’s Tree” which may be found on the internet. Our resulting decision tree using Occam’s Tree may be viewed at the end of this article by clicking on the word “tree”.

Based on our best predictions, our decision tree suggests a defense value for settlement purposes of our case is $80,800. Remember that the same computations may not work for the plaintiff if, for example, attorney’s fees are recoverable for the plaintiff or the plaintiff’s attorney is working under a contingency fee arrangement.

As noted above, the trial process is fraught with uncertainties and it is difficult to predict whether one will prevail much less the amount of a judgment. In addition, factors such as a client’s or the opponent’s level of aversion to risk, level of emotion or desire to stand on principle affect the amount for which a case may settle. The use of a decision tree, however, provides at least a platform for making outcome predictions and, in doing so, for estimating the settlement value of the case. Ask your mediator for assistance with a decision tree when preparing for your next mediation.

http://decisiontree.kleinmediation

tree

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s