Signing off now. Thanks to everyone for the questions, and sorry I just didn't have a chance to answer everything!

Yeah, I do. I think Jay is a straight shooter and was giving a good answer. The way he thinks about it makes sense to me. If everybody is betting the Vikings in Week 1 and 2 this year, he'll see that and factor that into his Week 3 line. What he doesn't need to do is hire a bunch of researchers and quantify why the…

That was the #1 thing right there. I started betting right when the McCracken article got out and the opportunity persisted for a few years. But it was clear there was less of it each year. I would guess it's almost all gone now, but don't know for sure.

Prediction Machine, incidentally, is another one of those services that does a lot of media, has picks in multiple sports, yet somehow doesn't seem to be a disruptive force. Hmm.

I think Billy Walters is the same way. Not quant, but obviously very quant friendly. If it's a high profile game, and lines are gradually shifting towards the favorite, that's more likely to be the public. When it one of those "screen goes black" moments where all of a sudden Middle Tennessee St price just gap moves…

Pretty lame answer, but I truly don't think I've ever gone higher than about 2%. I think slightly higher than 2% can be justified in some cases though, if you really know what you're doing. (And way way higher than that can be justified if "bankroll" is something you don't mind if you lose)

Good thing you specified that, I was tempted :).

And the other, more general, piece of advice I'd give is ... try to predict line movements. Couple of reasons for this ... first of all if you only look at your own won/loss results, it will take forever to figure out if you're any good or not because you're trying to figure if you can win at a 54% rate instead of…

It's actually quite weird that this article even exists, but in general there is not much out there that explains how good systems work, but then there is this one article that is written by a guy who made tens of millions a year and who writes very clearly. The article is in the book "Efficiency of Racetrack Betting…

If you want to bet for edge and not put a ton of effort into it, I'd say the biggest thing to do is find a niche sport (like CFL or WNBA) and concentrate on that. But if you want to do more than that, and if you are already a quantitative person and want to figure out where to get started with sports analytics…

In rough terms, yeah. I would say Dr Bob and RAS were (are?) successful, but now that everybody knows about their success, the amount of money you can get down on their opinions becomes limited.

Probably worth mentioning that my single least successful gambling weekend of all time was last year's March Madness. But with an insane number of D-1 teams, there's edge in truly quantifying how good everybody is. In broad terms, I'd say you want a giant regression model that relates all the teams to each other based…

The plus of having two entries is getting to play your "real" picks on whichever entry has the lead. As it happened, my backup picks had a better first week ... ie DFL2 had a better week 1 than DFL1, so I switched my "best" picks to that entry. Had I stuck with my top 5 picks on a single entry I would have finished in…

Also will add that Dr Bob was probably quite good at one point, but markets get tougher, and I'm not 100% convinced he's still +EV.

So if a guy's been around and is "famous" and he's not moving lines ... I think the obvious question is "why not?"

Here's the thing ... say you can pick 56% winners, and you can identify 1-2 games per day (throughout the year, across sports) that hit this threshold. Even with a reasonably conservative betting strategy of 2% of your bankroll per game, you'll double your bankroll every year. It's just the basic math that 56% winners…

I think highly of Right Angle Sports. Not because I've monitored him myself but because there are so many people whose opinions I respect who say "wait, this guy actually is good". Plus he is about the only one who really passes the smell test ... the market really does move right after he announces a pick.

They're certainly incentivized very different from the way gamblers are. They're compensated for getting people to think they have opinions that are worth something which is a very different thing from being compensated for predicting.

Like when Cliff Lee had his big breakout season several years back, he had a couple really high strikeout games early on. So you should have thought there is a decent chance he really is a completely different pitcher.

When I was betting, I didn't have any Pitch F/X data (nor did anyone else, I believe). So it was just a question of interpreting stats. You couldn't do that much better than chalking up everything to variance, but you could do a little bit better. Recent strikeout performance was (and I assume still is) somewhat…