So overvalued, check. Overbought, check. Overbullish, check. Upward pressure on yields, check. Market internals? – certainly mixed, but not bad – and there's the wild card. Historically, markets featuring a combination of these other risks have been vulnerable even without clear deterioration of internals. What the mixed internals (rather than clearly negative) buy you is variability in the timing of subsequent weakness. Sometimes the market plunges immediately, but often it bounces around for a while and continues to make marginal new highs. More often than not, what the syndrome produces is an abrupt plunge within a window of about 10-12 weeks. Not a forecast, just a regularity. This time might be different, but we wouldn't risk a great deal hoping it will.
It's important to recognize that when I quote probabilities, I am generally using a form of Bayes' Rule. So when I say, for example, that I estimate a probability of about 80% of fresh credit difficulties accompanied by a market plunge over the coming year, that figure is based on various combinations of historical evidence, and what has (and has not) happened afterward, and how often. As a side note, a “market plunge” in this context need not be a “crash.” In the context of a credit-driven crash and rebound (which is what I believe we've observed), a typical post-rebound correction would be about -28%, but even that would take stocks to less than 20% above the March lows.
Probabilities, however, are not certainties. If the probability of a given event is “p”, then the probability of “not that event” is (1-p). This, in my view, is what makes probabilities and average outcomes different from forecasts. When people forecast, they say “this or that is going to happen,” and very often they establish investment positions that will do them a great deal of harm if they are incorrect. What we try to do is say, “on average, these conditions have been associated with this typical outcome, as well as a range of other possible outcomes (risk) that is this wide.” The larger the typical outcome is, compared with the possible range of outcomes, the larger a commitment we are willing to make. But if the average outcome is weak, and the range of outcomes is wide, we'll defend against the risk.