Very good point. Bookies have reams and reams of historical data to base their football/racing/sports betting prices on, as well as what they see on the pitch week in week out. Even for general elections or Scottish elections, they have historical data on turnouts, margins of victory and again what the general mood of the population is in relation to the parties involved.
There is no real data here bar the opinion polls (from smallish sources?) and as you said, from the flow of money.
As a slight side issue, in terms of opinion polls etc I was on FB last week and saw the Nae Danger energy drink had posted the question on their page after last weeks debate, yes or no and asked its "Likers" to post their comments.
Now obviously Nae Danger energy drink probably has a certain demograph that it appeals to and who are likely to "like" it on FB, and that section of people are probably not a fair representation of the Scottish voting public. However they do have 35,000 Likes and, when I last looked, they had over 1,700 people responding to this FB post. I scrolled few quite a lot of the replies, in fact a very lot. From a rough estimate the percentage saying Yes was, IMO, anywhere between 90-95%. There were pages and pages of Yes replies with only the odd No here or there, usually one in every 15-20 replies or less at times. End of the day its still an opinion poll and it had quite a high number of responders. Just shows that if you ask certain people a question you can get widely varying results depending on who you are asking.