A couple of nasty pick-flips due to closing lines cost the model a winning week, but that kind of thing usually evens out by the end. Since Week 5, the model has gone 48.3% ATS and 67.2% SU, which is nothing to write home about, but does show improvement when using 2018 data only. Clearly 2017 numbers were no bases on which to make 2018 projections – at least for this projection model, and I will need to make a change for 2019. Either start the model up after Week 4, or – do something else?

A reminder: Some sites don’t allow the editing of posts, or limit the amount of time allowed to make changes. In such a situation, this means that as changes occur during the week, they will need to be posted as “replies”, so you’ll need to scroll down to see the latest projections.

GLTA

I knew my model’s success at O/Us wouldn’t last. While the Vegas line delta has dropped to 9.3 points, the Limper’s delta has climbed to 7.7 points – which is getting close to coin-flip territory. Like I said, “I’ve been to this movie before”, like watching ‘Titanic’, I know that iceberg is coming. If that delta gets above 8.0, I’ll pull the plug before it sinks – and – once again – work on it in the off-season.

Time to put this turkey to bed. Totals are a statistical bear, and although it had a decent start, the delta has ballooned to well over 8 points – and that’s a big balloon. Next season, I know I’ll try again, with a different set of variables and weights – which, I think, is the definition of insanity – but, there you go.

I wouldn't give up on it just yet. Run it for the full season.