My sons and I decided that we wanted to run a Premier League prediction competition this year. We decided on a system where you got 2 points for putting a team in the right place in the table, and 1 point if your prediction differed by one place in either direction. The advantage of doing this with the Premier League is that, at least until the latter stages of the FA Cup, the teams play rounds of matches, at the end of which every team has played the same number of games. (We are all Reading fans and a postponed match has caused our team to have a spuriously low position in the table for several weeks, now remedied.)

A parent with coding skills can put something like this online in a short time. I have written a script that grabs the current table off a Web page and parses out the team names in their current order, together with the number of matches played. Sorting this out probably took most of my time. This order is then compared against four lots of predictions (my husband has joined in, with a random order of teams). The current table, number of matches (so one can see whether the current round is complete) and all the predictions are displayed as columns in a table. Teams in exactly the right place are in a box with a red background; those one place off are in a pink one. Below the table the current score for each predicted table is given, with a breakdown of which predictions have contributed to it.

After each round of matches I note the scores and winner of that round in a spreadsheet. I could have automated that process too, but I decided not to put in all the extra effort. Perhaps at a later stage. A pattern is beginning to emerge. The random predictions tend to score about 3. The non-random ones have gradually improved as teams have gravitated to their natural place in the league, except that the most accurate prediction of all so far was mine after the first round of matches: 13. And predictions can fluctuate a lot between rounds. After round 14 I was the winner with an accuracy of 8; but one round later I scored only 4.

The non-random predictions have some agreement among themselves. We all thought Everton would be 7th; this seemed most unlikely for a while, but they’ve now risen to 10th. We all put West Ham two places below them, which is way off at the moment. And Leicester in 11th which is rather better as things stand.

I attended the following demonstration:

‘ … an opportunity to view the potential ‘research information system’ for [Bristol] University…. The Danish company Atira will be demonstrating their product ‘Pure’.’

Because of flaky wi-fi at the venue, we weren’t able to have the full demo, and in particular the Research Excellence Framework couldn’t be demonstrated.  This system is intended to replace IRIS as a place for collecting details of research outputs; it can also handle personal profiles, details of projects and funding, and include a repository of full-text articles.

It seemed to work nicely, and the interface for merging duplicate records from standard databases (allowing you to choose the most detailed version of each field) was impressive.  But I couldn’t help wondering how well our current practices would transfer to this system.  For example, how can salary details within projects be kept confidential? Who can see personal details in your profile?

The most interesting aspect to me was the potential replacement of the University’s ROSE repository (based on DSpace) though the demo only touched on this in passing.  This looked rather like an add-on to the bibliographic storage system and it wasn’t clear how much control there is over metadata, or whether a licence can be stored alongside the repository items, as with DSpace.  On the other hand, there is some advantage in integrating a repository with the database of information about research outputs, as it is rather opaque who is behind ROSE.  (Someone at another institution recently thought it might be me, on the grounds that I’d posted to a DSpace forum!)