Any low-barrier-to-entry crowdsourced valuation model must simultaneously satisfy two fundamental principles:
Otherwise, the system will have a signal-to-noise ratio problem.
Obviously, rewards $p$ must exist to attract valuators to contribute signals — these can even be non-monetary rewards in a broader sense. Clearly, rewards should only occur when valuators are successful (reach consensus), otherwise it would directly lead to an influx of "noise" from people gaming the system.
The necessity of a punishment mechanism can be mathematically proven as follows:
Let's first assume a scenario without a punishment mechanism,
A crowdsourced valuation system has n valuators. They are numbered $1...n$.
Where valuator $i$ has an actual success probability of $\alpha_i \in (0,1)$. Due to the non-trivial nature of the valuation problem, $\alpha_i < 1$; because valuations have a reasonable upper limit and finite precision, all possible valuations are finite, so $\alpha_i>0$. Here, $\alpha_0$ represents the success probability of the default random "coin flip" valuation method.
Where valuator $i$ has an observed success probability of $x_i \in [0,1]$. More specifically, when there is only one observation, $x_i \in \{0,1\}$ with probabilities $1-\alpha_i$ and $\alpha_i$ respectively, and when there are $\infty$ observations, $x_i = \alpha_i$. It's worth noting that since valuators can choose to either use the same identity to carry forward their previous prediction data or use a completely new identity (commonly known as alt accounts) each time they enter the market. This means valuators can freely switch between these two modes.
According to "The Market for Quacks" model [1], such a market will inevitably contain many random valuators ("quacks"/noise), and "bad actors will drive out good ones."
Therefore, any crowdsourced valuation system without a punishment mechanism will inevitably lead to insufficient signal-to-noise ratio.
Thus, since a punishment mechanism is reasonable, using penalties as rewards is also reasonable.
[1]R. Spiegler, “The Market for Quacks,” Review of Economic Studies, vol. 73, pp. 1113–1131, Feb. 2006, doi: 10.1111/j.1467-937X.2006.00410.x.