What is the "value" of a prediction?

Net worth is calculated as the sum of cash on hand and the Prediction "worth". Cash is cash, so this is all about the latter. As far as I can tell, the Prediction value is the sum of the open questions wagers' cash-in value.

But what is the cash-in value of a wager? Is it the "actual" value of that wager? In one sense, yes. If you have exactly one wager per question, the actual value can be interpreted as how much you'd get if you cashed in. (The idea being that a wager has no intrinsic value other than the money you can get for it NOW.) However once you have more than 1 wager on a given question, that interpretation has problems, because cashing-in the first wager will change the odds, and thus the value of your other wagers on that question. So the practical "a wager is worth what you get for it if you cash it in"approach is not as simple as it seems.

The other way to value a wager, which I feel more accurately reflects both common-sense and the predictive utility of the market, is to multiply the current percent for that wager's prediction choice by the amount that wager gets if it's correct at settlement time.

Why does this matter? Well, for people making wagers at or less say $200 it hardly does. But since people can and will be making wagers at up to $10,000, it makes a huge difference. The cash-in value of a wager is calculated relative to how the market will look AFTER the wager is cashed in. A small wager is a drop in the bucket (unless it was bought at very low odds and is now at 99%). But a larger wager stands a good chance of being a significant portion of the total money wagered on the question. This wager will affect the odds dramatically both when it's made (it sends them up - relative to the chosen answer), and when it's removed through cash-in (it brings them down). So, for a large wager, its cash-in value will be calculated at a significantly lower rate than a smaller wager.

I'm not sure I think this is entirely a good idea. From what I understand, the buy-in mechanism softens the blow on what percent a wager goes in at (somewhere between the going rate and the rate to come as a result of the wager), but the cash-in value takes the full percentage drop. I realize you can argue that that's the penalty you take for cashing in. I think the penalty is excessive on large wagers, but I can basically accept the reasoning.

But basing the Prediction portion of the Net Worth on the cash-in values sum means that these penalties are accrued serially, such that someone who only bets $10,000 per question will be noticeably lower in current net worth than someone who bets identically but instead wagers $100 a hundred times per question, even if they start out with the same amount of money.

I believe that a wager's value should be separate from its cash-in value, based on the market rate. This would make clear that cashing-in costs big-time, and that huge wagers on questions are not a smart strategy. But it would make the net worth much more reflective of the net worth before a series of large wagers, and to some extent of the likely net worth of those wagers after settlement.

jersjusttrolling has lost a huge amount of apparent net worth because of this mechanism, and if their predictions weren't crazy, they will jump up shockingly later (leading to potential unfair suspicions of dishonourable conduct, when it's all the result of, to my mind, an unfortunate mechanism for calculating net worth).