Pick Performance Update: September 17, 2015

The business of making sports predictions is a wild one indeed. Consider the dynamics of some of the types of picks we currently make here on TeamRankings.com:

  • Sports betting picks. With sports betting, being wrong almost half the time can qualify as a big success, while being wrong exactly half the time doesn’t cut the mustard. What does that mean in practice? Go 38-32 on 70 recommended picks in a given NFL season, and you’re hitting at a highly respectable 54% clip. But go 36-34, just two games worse (which could easily be the result of a couple unlucky bounces, missed calls, or badly timed penalties or turnovers) and all of a sudden you’ve had a losing year.
  • Pool picks. If you’re competing against dozens of people in a typical football pick’em or survivor pool, you’re never expected to win in any given year. However, shortening the number of years between expected prize wins in pools can result in some pretty amazing long term profits, since winning just once can produce a huge windfall. In addition, coming in first place once but dead last the next four times, is almost always better than doing pretty good (but not winning anything) all five times.
  • Daily fantasy. Finally, the new kid on the block, daily fantasy sports (“DFS”), presents players with a range of risk vs. reward options. You can be more conservative and only enter contests that give you close to a 50% chance to double your entry fee, or you can be aggressive and try to strike gold in huge tournaments that only pay out to the very top scoring entrants. Succeeding at one vs. the other requires very different strategies and expectations.

With all that as background, the point of this update, which we intend to do periodically, is about both transparency and education. We’ll review how our picks are doing from the 10,000 foot level, highlight performance over both the short as well as longer term, and provide some insights around what matters and what doesn’t.

Why The Broader Context Matters

As our site has grown over the years, we now make picks for lots of different things, yet many of our customers only get a very “micro” level view of what we do. Some customers only care about a very narrow subset of the predictions we make, and short term outliers (e.g. especially good or bad weeks or days) naturally get the most notice.

When it comes to evaluating results, though, it’s important to understand the broader context. Most importantly, the nature of making the types of picks we make means that even long term success will almost certainly be marked by a lot of short term variance in results. Put another way, it’s very unlikely that any specific week’s performance will exactly match what ends up being our final, end of season success rate.

Perhaps most important to understand is that the more “micro” you get in your evaluation — say, only looking at how a specific type of pick performed over a small sample size of games — the bigger the influence that luck can have on our results. For example, last year our 2- and 3-star rated NFL point spread picks went a very successful 66-54-0 (55%), but we had weeks when they went 2-4, 3-5, 3-6, etc.

In the short term, pretty much anything can happen. That’s to be expected, and judging any system on a week’s worth of results isn’t a good idea.

Now let’s get to the review of the past couple weeks of picks.

Note: With our Football Pick’em Picks product, our customers report their current place in their pool’s standings on a weekly basis, when they come to the site to get the current week’s picks. As a result, we’ll always be a week behind in reporting how our customers are doing in football pick’em pools.

Picks That Did Well

  • Seth’s NFL Week 1 FanDuel Lineup. Our runaway success from NFL Week 1 was Seth’s FanDuel lineup, which he entered into five contests. The lineup finished tied for 3rd place out of 333 entries in the TeamRankings Football Championship, winning triple the $10 entry fee plus a chance to compete for at least $500 in bonus cash. In addition, the lineup took 5th place in a 501-entry tournament, 5th place in a 857-entry tournament, 6th place in a 1,130-entry tournament, and 254th place (99.8th percentile) in the 143,678-entry Sunday NFL Snap. The final results: 5 entries, $32 in total entry fees, $295 won (an 822% ROI), and 4 bonus prize tournament seats won!
  • Week 1 College Football Pick’ems. Entering Week 2 of college football (see note in the intro about why we report pick’em results one week late), nearly 1 out of every 4 customers (23%) who used our pick’em picks exactly as recommended reported being in first place in their season-long college pools, and 36% of them were in the money. This actually isn’t as good as our picks have done in the past, but obviously 23% in first place is still a lot better than typical expectations. Lots of season left.
  • NFL Week 1 Survivor Picks. We got lucky with Dallas as our top pick of the week, but when the dust settled, our top 5 final recommended survivor picks for large pools (Dallas, San Diego, Denver, New England and the New York Jets) all won. That also means customers following our advice for playing multiple entries would have had all picks survive, and saved Green Bay for later. In addition, we recommended against picking the mildly popular Seahawks, who lost.

Not Particularly Great Or Particularly Bad

  • NFL Game Winner Picks. Week 1 NFL winner picks went 11-5, equaling the ESPN pick’em consensus performance and going 9-4 against the ESPN experts.
  • NFL Over/Under Picks. Our final, playable-rated totals picks for NFL went 3-3 in Week 1, but also scored a win in our only 3-star pick of the week, Baltimore/Denver Under 47.
  • David’s FanDuel Lineups. David entered a selection of 50/50 and head-to-head contests, as well as some bigger guaranteed prize pool tournaments, with his Week 1 FanDuel lineups. Since he’s entering some really big tournaments, we expect his week-to-week results to be fairly boom-or-bust — mostly losses with hopefully some big wins every now and then. This week produced some winnings in some contests, but the net result was a very slight loss on total entry fees. The final tally: 80 entries, $225.00 total entry fees, $212.60 won, $13.40 net loss.

Picks That Didn’t Do Well

  • NFL Spread and Money Line Picks. It was an abominable Week 1 for our NFL spread picks, with our closing playable picks going 0-4, and all spread picks going 3-12-1. That’s the worst week we’ve had in years (and of course, it had to come during our Week 1 free trial). To top things off, our sole 2-star rated money line pick, Washington +161 over Miami, was looking like a nice upset call until it was derailed by a punt return TD in the fourth quarter.
  • College Football Betting Picks. Across the board, college football betting picks came down to earth in Week 2 after a ridiculously hot Week 1, with spread, total, and money line picks all in the red. On the season so far, closing playable picks for college football are 19-15-1 ATS and 19-16-0 on totals, while money line picks are down -4.9 units.
  • MLB Betting Picks. Coming off a smoking hot August, MLB money line picks and run line picks were down -14.1 units through the first half of September, taking a chunk out of a still-profitable 2015 season to date. For the year, all money line and run line picks are still up over +20 units largely thanks to the run lines. It’s been a similar situation for playable MLB totals picks (2-star rated), which had gone 1-4 so far this month, but 82-55-8 (60%) on the season as of our 9/16 prediction accuracy report.