People often walk up to me at club or in a tournament and ask something like: "Would you challenge ARMPITLIKE?"
Rather than guessing at the word's validity outright, I usually end up leading them down a long list of followup questions like: "Who played it?"; "What was the score?" etc. etc. Eventually we both just wish I had answered the question with "Yes" or "No".
The point is that there is so much more behind making a decision to challenge/accept a play than just word knowledge or hunches. I think one of the reasons we don't often see in-depth discussion of proper challenging technique is that you can always trump challenging strategies with word knowledge. The "optimal" decision is to simply know every word and correctly challenge off the phony ones and accept the valid ones. However, I don't believe that anyone has perfect word knowledge. Everyone is unsure of something, especially when the words exceed 8 letters and multiple dictionaries are clouding the mind.
Besides, it's not like there is anything impressive with having read up on the long -LIKE words before and correctly challenging a word that didn't appear on that list. Knowledge of tens is nice, but it's usually quite easy to apply in a challenge situation. I'd probably be more impressed with someone who honestly had no clue, but correctly weighed all the variables and determined that accepting ARMPITLIKE* was the best course of action.
So I think it's safe to say that "Would you challenge ARMPITLIKE?" is a grossly oversimplified question. As soon as your certainty of a word's (in)validity drops below 100%, there are instantly a dozen other variables you need to consider before challenging/accepting. And it's not like I'm introducing a new concept here. The fact that challenging/accepting questionable words is often based on context rather than word knowledge is nothing new. You often hear people say: "Well, I can't win if it's good so I'll challenge." before challenging what they feel is a plausible word. However, I do believe that this branch of Scrabble strategy is maybe a bit unappreciated. It doesn't seem to get discussed in positions or annotateds much. I think it's surprising, considering that free turns often equate to large swings in game outcomes.
Now, if you're shaking your head at this because you'd rather just learn all the words 100% and make all the correct challenges that way, then fine. But read on, because a lot of the heuristics discussed in this post can also be applied to the art of phony-playing. And even the most prolific players with near-perfect word knowledge can benefit from playing the occasional phony word.
So what are the factors to consider? There are lots of them, I think, and many of them are far more subtle than "I can't win if it's good."
Here's my attempt at a list:In-game factors
1. Game state
This is the big one, obviously. It encompasses the score, the tiles on your rack, the tiles unseen, and the board situation. This is where you need to spend your time, because you need to consider three scenarios:
a) accept the play
b) challenge and win
c) challenge and lose
For each scenario, you have to ask questions like:What play do I make if this happens?What are my chances of winning?
Because you can always choose to pursue situation A, the decision to challenge often comes down to weighing situation A against a coin flip (argument's sake) of hitting situation B or C.
For instance, if you determine your win chances are: A 90% B 100% C 30%, then you probably have a pretty easy decision to accept the play. Same with A 5% B 30% C 0%, being virtually forced to challenge. It gets dicey when you're looking at something like A 50% B 75% C 25%. And of course, you've never got these percentages sitting in front of you. It is pretty difficult to approximate your win chances for each possible outcome of a challenge/accepted play.
Even though this factor is just one in a list, I should emphasize that in-game context (scenario vs. win %) is by far the most important factor, and the one you should always consider first (assuming you aren't 100% certain of the word's validity). Basically, you're going to spend your time here, and begin to consider the other factors when the decision is not obvious based on your analysis of the game state.
Lumped in with this factor are the fun situations where accepting an opponent's phony will give you a comeback bingo. Also, the tricky situations where you might want to weigh lopsided matchups to fudge the win % numbers.
2. Opponent's game state
It's also worth it to consider the game state your opponent faced when he played the word.Was he desperate?Could he have made a better or equally strong play with these tiles?
Of course, that second question is often a trap. Be careful assuming that your opponent wouldn't make a horrendous word knowledge error or miss an obvious bingo. Just because he could have played (H)OUSEtOp, it doesn't mean his play of mOUSE(F)Oo(D) must be acceptable.
The first question is probably a more valuable one to consider. Was my opponent in a situation where he would feel compelled to attempt a word he was unsure of? Out of desperation? Or simply because the game was already decided, or too early to call? Again, it's still always a gamble to assume your opponent thinks like you do. And it's also a gamble to assume that a desperate situation and a fortuitous obscure bingo can't coincide. But it's always worth considering.Meta-game factors
3. Knowledge of the opponent
If I've played against the person a fair number of times (or at least seen them play in some annotateds), then I hope to be able to answer the following questions with a decent level of certainty:What is his word knowledge like?Is he known to make mistakes on words?If he known to intentionally play phonies?
If I've never played against the person and know very little about their game, then I can probably at least answer the following:What is his approximate rating?How long as he been playing?Does he play other dictionaries?
All of these questions will help me to determine how likely it is that my opponent knows an obscure word, is making a mistake with his word knowledge or guessing on a word, or is intentionally trying to phony me.
Can you add more weight to knowledge gained more recently? If an opponent won a challenge earlier in that same game, are you less likely to challenge his next obscure-looking play? You probably shouldn't be, but it's going to affect your decision, likely.
4. BehaviourCan I glean anything from his body language? Can it be trusted?Did he play the word quickly? Does that matter?
Obviously relying on this factor is dangerous, but I don't think it can be ignored. It helps to have experience against your opponent here. I won't list any rules of thumb, since they obviously become less useful the more people I share them with. Non-game factors
5. PressureWho is watching? Do I want them to see me allow this ridiculous word?What is at stake? Could I forgive myself for blowing the tournament on a word like this?
Ideally, these should be nonfactors, but you might find your decisions being affected by them nonetheless.
7. Linguistic knowledge / HunchesHow plausible is this word based on my knowledge of other words?
Duh, kind of an obvious one. But it's one that seems to get overlooked the deeper you get into the game. (Admittedly, the dictionaries do betray logic often.) This is where "word knowledge" is separated from "language knowledge".
Hey, you might not have studied enough to know for certain whether ACARUSES is valid or not, but your knowledge of Greek pluralization might help you there. Maybe your knowledge of chemistry can help you sort out which -ATE, -ITE, -IDE words should logically be valid.