When betting on sports or playing poker, we are trying to think probabilistically in how much we expect to win for every dollar that we bet. For example, if I have a flush draw, I have roughly 36% equity to win the hand, and so if the pot is $100, betting half the pot ($50), my pot odds would be 33%, and I would be making a +EV bet, since 36%>33%. We can apply this logic in the draft as well.
Most draft models that I have seen focus too much on where in the draft a player was taken, rather than the signing bonus the player received. As more nuanced fans of the draft know, teams like to draft a player a bit higher than they normally would if that player agrees to take a signing bonus lower than the slot value of the pick. This is a very important concept of the draft, and one that is overlooked by common fans (which isn’t their fault, it isn’t even explained by the analysts that are commentating during the draft).
To see if we are making +EV plays in the draft, we have to figure out what percentage of the time we think a player will make it to the majors, and what his WAR (or similar valuation) will be over the course of his career. These would be very difficult to calculate, but if we had some templates already built in, that would make it a lot better. For instance, if we had historical data on players that were rated with 60 power, 50 hit, 60 speed, 60 defense, and 40 arm, we could have a rough idea of how those players are doing in the majors and what their WAR is.
However, lets say that we are in a fantasy land have perfect information about a potential draft pick. This player has been determined to have a 10% chance of making the majors, and that he would generate an average of 6 WAR for his first 6 seasons in the majors, if he makes it. For this, let’s assign $3 million per 1 WAR, which is less than the “going rate”(I think the $8 million WAR is pretty useless) but more accurate given the renewal and arbitration. I might be off, but it doesn’t really affect the formulas. Our pot here is $18 million (18 WAR * 3). To make the signing bonus +EV, you would have to offer around $2 to be a push, since 2/20=10%, and less than $2 million for this to be +EV.
To make an accurate model, it would require a large database of not only signing bonus and WAR data, but also previous scout grade data as well, which obviously changes from team to team and is not publicly available. Also, hopefully I am not rehashing work that someone has done already. I haven’t seen this take before, but obviously I could be missing things. Anyways, thank you for reading and hopefully you got something out of this article.