Aside from a computer on the desk, my local betting shop has a traditional look, complete with newspaper racing pages sellotaped to the wooden walls, stubby pencils and drawn blinds. As I entered, a man with a lived-in face and an unlit roll-up cigarette protruding from the corner of his mouth was exchanging a slip of paper for some ten pound notes.
Most betting here is on the horses, but I was looking to place a bet on politics. More specifically, I wanted to place a bet against the psephologists who are predicting that the SNP could take every seat in Scotland.
Ipsos Mori; Panelbase; Lord Ashcroft - the pollsters after getting it right on the referendum could be riding for a fall. Scotland's merry band of psephologists led by the giant-brained Professor John Curtice, who appears to have instant recall of almost every British general election you could mention, have been running a very long race. Throughout the run-up to the independence referendum there was a massive global interest in Scottish political polls, all the different varieties,the analysis, the polls of polls, the swings and surges, the ups and downs.
At times the psephologists seemed worried. They were concerned that voting behaviour could prove hard to measure and predict. One got the sense that they were nervous about calling it wrong. The polls shifted this way but the eventual result was closely foreshadowed by the eve-of-vote polling and the psephologists were riding high. They were extremely proud of the accuracy of their efforts.
But with barely time for a breather, the race was on again in Scotland with the first polling evidence that Labour could be out to pasture while the SNP broke all the records with a massive win.
But could this be the pollsters Becher's Brook? Are they in danger of coming off at the fence?
I think they may be. The reason for this is the very large number of undecided voters. About 30 percent of people who say they are certain to vote are undecided, and that rises to 40 percent in Glasgow.
It is standard practice to omit the undecided from the poll summary. That makes sense if the people really are undecided - in which case they are likely to break the same way as the decided people.
But there is another possibility - that they could be shy about their voting intentions. I think this could be the case here. I wrote in the run up to the referendum about the way that 'No' voters seemed reluctant to pin their colours to the mast. Reading the runes, it seems that many of those who voted 'No' in the referendum may vote tactically against the SNP candidates.
Specifically, there is concern in Edinburgh that the city may be in danger of slipping into a long, slow economic decline similar to that experienced by Montreal after the Quebec referendums. Financial firms are considering their options. The property market, especially for larger houses appears sluggish and won't be helped by Holyrood's new tax on homes over £350,000 which came in on April 1.
For sale signs that have become fixtures on big houses across town tell their own story. Edinburgh voted 'No' in the referendum and many of those voters are keen to avoid any prospect of further uncertainty or another referendum.
It seems to me that the odds on voting against the psephologists are too good to miss. After getting the odds from the man at the counter in my local shop, I chose the gambler's bets, eschewing boring evens or 13 to 8 or anything like that and returned from the bookies with a couple of slips: six to one on Sheila Gilmore holding on in Edinburgh East and five to one on the SNP getting from 31 to 40 seats - still a huge increase on six, but less than the predicted 50 plus or 'all the seats'. At the very least, it will add to the excitement of election night.