On The Definition of Logical Alternative

This is the full transcript of a letter I sent to The Bridge World quite some time ago. After the letter, I'll add some thoughts that have occurred to me recently. --Jeff

To The Editor,

Law 16A states that in cases of unauthorized information (UI) obtained from partner, one may not choose from "among logical alternative actions one that could reasonably have been suggested over another by the extraneous information." What exactly is a logical alternative has been the subject of debate for many years; perhaps the most illuminating discussion of this topic appeared in the "Appeals Committee" articles in The Bridge World years ago.

I have been given to understand that appeals committees have recently been instructed to interpret "logical alternative" to mean "any action a number of the player's peers would seriously consider." I have several objections to this interpretation. I feel strongly that, while the previous interpretation (the "25% rule") was inadequate, this interpretation is far worse.

"A number of ... peers" is vague. How many? What portion? Is one enough? How about zero? ("Zero" is a number.) Two or three? Most?

From a player's standpoint, this rule is hard to use. It is hard enough to evaluate what other players in my shoes will actually do with my cards. To try to guess what they might "seriously consider" is pretty much impossible.

What exactly is to "seriously consider?" Clearly, it is not necessary for any of one's peers actually to choose the action in question. In a Masters Solvers Club context, a logical alternative need not get any votes at all. Is a possible alternative one that gets a mention from a panelist? From a few? One that gets a 20 score? 50? Certainly, one that gets "some number" of votes is a logical alternative, but what about actions that get none?

The standard is too harsh. To be required to take an action that none of one's peers would take is overkill. Since an ethical player will generally choose to err on the side of caution, and since it is nearly impossible to determine if one's peers will seriously consider alternatives, the de facto standard becomes that nearly any action that might be suggested by unauthorized information cannot be chosen.

I am convinced that this new guideline (it is not a law) is not functional in defensive situations. Imagine a case in which partner signals slowly. If I were to work out that obeying partner's slow signal is a 100% play---no other option has any hope at all---I still cannot choose that action because some other players in my shoes might not think through the situation carefully. A signal played as to be obviously a signal must be treated as a possibly careless card; I cannot treat it as a command ethically if other players might seriously consider the alternative. Moreover, even in my 100% case, I must make a knowing error because the unauthorized information clearly suggests the winning action over another. This seems to be contravening Law 16A; the losing action is not "logical." Furthermore, if my partnership style or personal style is to treat all cards, fast or slow, as signals, and is to follow them unless the alternative is 100%, I am still barred from taking the winning action. Why? In my hypothetical style, taking the losing action is not an "alternative." This can be made more poignant by a bidding example. With no one vulnerable in third seat, I might be dealt K432 84 QJ32 1083. When partner passes slowly in first seat and RHO passes as well, am I obliged to psych an opening bid? I am sure that some of my peers would seriously consider psyching. I never do; to psych in this position is not a logical alternative for me. It is not even an alternative. But passing is an alternative suggested by the UI (partner's having values means that the opponents are less likely to have a game, so the upside potential of a psych is reduced, thus making it a less desirable alternative) so, by the "seriously consider" criterion, I cannot choose to pass; I must psych! That is absurd.

In Flight B and other events not of the top quality, the "seriously consider" guideline is next to useless. A primary distinguishing feature between a good player and a not-so-good player is that the weaker player considers more losing alternatives. In a Flight C field, it is likely that a player's peers would seriously consider many clear-cut errors. In the presense of UI, are Flight C players obliged to err? How can a committee rule in these cases? It is hard enough to determine what a consensus of Flight C players will do; determining what they might consider is nearly impossible.

On a purely subjective level, I feel that the "seriously consider" guideline is too strict. I agree that the "25% rule" guideline was too lax; perhaps some figure closer to 10% would be better, but the "seriously consider" rule is much much stricter than that. It is stricter than a 0% rule.

Finally, I object to the guideline on grammatical grounds. The infinitive form of "seriously consider" is "to `seriously consider,'" which is hard to distinguish from the split infinitive "to seriously consider."

All in all, I have enormous reservations about this new guideline. Yes, the problem that it was intended to solve exists, but the solution implemented by the change is not reasonable. At the very least, it must contain some mention of the acting player's decision process. To be required to take the worst ("worst" is too strong, but the cases are complicated, so it will do) action from among all those that any of my peers would seriously consider is extreme. Even to have to try to come up with all the possibilities that my peers might consider is too much effort to be reasonable given the time limits at the table. Perhaps "obvious logical possibilities that many peers would seriously consider" comes closer to the mark. It is still too harsh a guideline, but at least it does not contain so many absurd examples.

Thank you for your consideration.

Jeff Goldsmith

More recent stuff:

Many players very strongly object to the "seriously consider" guideline, sometimes called "LOLA" or the Law of Logical Alternatives. Many err by blurring the distinction between the interpretation of what exactly is a Logical Alernative and the prohibition against choosing one. The latter is necessary and reasonable. The former is much more difficult to judge; indeed, I have stated how strongly I feel that the current definition is inadequate. I haven't made any good suggestions so far, so I think it's incumbant upon me to try.

An action is a Logical Alernative if:

  1. A moderate number (>= about 10%) of the player's peers would seriously consider taking that action, and
  2. At least a very small number of them would actually do it, and
  3. The action is a reasonable action, not a wild, swinging action, and
  4. It cannot be demonstrably proven that the action is an error, and
  5. There is no other action that is "normal and obvious" that nearly all players would take.
This is a pretty complicated definition, but maybe it will cover the bases better than what has gone before. I hope someone can simplify it a bit.
Jeff Goldsmith, jeff@tintin.jpl.nasa.gov, Jan. 17, 1997