The Smarter Money?

[Rudy]

In the elusive hunt for a reliable indicator using consensus statistics from various sources, I am having pretty good success thus far in the College Football season on games where opinions are line up largely on one side, and then comparing the Wagerline and BigGuy consensuses on these games and using the BigGuy players as the smarter money. Have tested now for 3 weeks prospectively after settling on methodology, and the results are encouraging:

Week 1: 11-9 55%
Week 2: 16-8 67%
Week 3: 16-11 59%
YTD: 43-28 61%

Here's the methodology:

The Wagerline consensus on a game (sides only) must reach the 60% level or higher. I then look at the BigGuy consensus on that game, and if the differential between the two consensuses is 3 percentage points or more, then a play is triggered. If the BigGuy consensus is even bigger on one side than the Wagerline consensus, then that team is the play. Conversely, if the BigGuy consensus on the game is more muted than the Wagerline consensus, then the play is against the team heavily favored by the Wagerline players. In either case, the BigGuy consensus is deemed the smarter money.

If I change the initial filter from 60% on one side to 70% on one side, the YTD record is 19-10 (65%), a better percentage than the 61% with the 60% initial filter, but with fewer than half the number of qualifying games.

I recognize that the sample is still small, and unfortunately there is no way to backtest because there is no archive readily available of the BigGuy historical data. But with 3 straight winning weeks, it appears that there may be some merit here. Since the Wagerline consensus is so large, it is a good proxy for the public. The BigGuy consensus has evolved into primarily a monitoring service for professional and aspiring touts, so my assumption is their picks are thought through more thoroughly.

One caveat: The consensuses generally aren't settled until less than one hour before game time, especially BigGuy.

I'm also testing this on the NFL, and the results in the 1st 3 weeks are 7-6, 3-0, and 4-5, for a 14-11 (56%) record YTD. Last year in the NFL, I started testing a system using the Hilton data as the smart money versus the Wagerline data, and it finished 28-24 (54%) after having started like a house afire at 19-10 through 10 weeks. So I gave up on it becasue of the late season fade, but now in the 1st 3 weeks of this season, that system is 13-4 already. Go figure.

I'll try posting the plays in CFB as they show up this weekend in this thread, and we'll see if the trend continues.

 

[JEFF]

Very interesting stuff Rudy. I also use consensus sites as a mjor tool for my handicapping. I have one question about you rmethods though:

The Wagerline consensus on a game (sides only) must reach the 60% level or higher. I then look at the BigGuy consensus on that game, and if the differential between the two consensuses is 3 percentage points or more, then a play is triggered. If the BigGuy consensus is even bigger on one side than the Wagerline consensus, then that team is the play. Conversely, if the BigGuy consensus on the game is more muted than the Wagerline consensus, then the play is against the team heavily favored by the Wagerline players. In either case, the BigGuy consensus is deemed the smarter money.

Can you clarify this? I'm not sure I understand. The thing that is throwing me off is the 3 percent difference thing.Are you saying that if the Wagerline consensus is on Notre Dame at 63% and the Big Guy consensus is Notre Dame 60 or 66%, its a play. Or do you mean 3% in the other way, as in the wagerline consensus is 60% on Nebraska, but the Big Guy is 63% on Iowa St, you take ISU?

Sounds good, I only ask because it interests me ...

 

[UT-Longhorn]

Sounds like a good plan my man, please post the plays.......hopefully the first one you find this week is FSU and the UNDER

 

[Rudy]

Jeff: the 3% thing is to make sure there is enough separation between the consensuses. You had it almost right in your example. Take the Thursday night game this week, the Wagerline consensus on FL ST is presently 65% -- high enough to qualify as a play. Then, the BigGuy consensus must either be higher than 68% on FL ST to qualify as a play on FL ST, or under 62% on FL ST to qualify as a play on Louisville. Right now the BigGuy is 67% on FL ST, so it wouldn't be far enough away from the Wagerline consensus to be a play.

UT-Longhorn: I haven't done any testing on CFB totals because the lines come out late and the # of picks on BigGuy is typically pretty small. Right now, only 11 people have selected total on the FL ST/LOUIS game, which I consider too small of a sample to be meaningful. But if the sample size is big enough, you can probably infer that the data is just as valid for totals.

 

[Skinar]

Very interesting Rudy. Thanks for the work and I'll be looking forward to the selections this weekend.

 

[bassmaster007]

ditto 2 what skinar said can u state which is 60% and which is 70% thanks gl

 

[Rudy]

Looks like no play on the Thur night game, as both consensuses are almost the same; FL ST has 62% on Wagerline and 64% on BigGuy. But things could still move in the last 1/2 hour -- if Wagerline goes to 61% or BigGuy to 65% it would trigger a play on FL ST. I normally don't sit on pins and needles waiting for the data to tick, especially on multiple game days. I would just call it a no play now.

 

[Rudy]

New Mexico is a play tonight.

 

[pt1gard]

thanx rudy

thanx rudy

i love this type of stuff, great write up/explanations

looking fwd to seeing the plays

thanx,
gregg

looked at wagerline

here are the fishy games and teams that have the scales on them are:

temple owls
tulsa
rice *
Nev
utah *
toledo

if you use reverse logic, those are the plays--betting against the public tho lines arent reflective


* games actually moved other way -- double whammy of logic:shrug:

 

[Rudy]

Looks like there will be about 20-25 plays today. Remember this is still experimental. Right now, I'm using this mostly to keep me off the other side rather than as a "bet-on."

I'll asterisk the ones that qualify under the more stringent 70% filter, which is hitting at 65% as opposed to 61% for the 60% filter.

Early games:
CIN
EC*
IND
WMU*
CMU
LSU

 

[Rudy]

Mid-morning games:
ARMY
UCONN*
NM ST
PITT

 

[Valuist]

I think if you are going to use Wagerline, use it only as a go-against; esp. when they are at least 2/3 on one side.

 

[Rudy]

Valuist: I agree it would seem intuitively true just to use Wagerline as a go-against, but while I've been working on this the data is showing that if the BigGuy consensus is even higher, that makes it a go-with. Just looked at the YTD data for such cases, and when the Wagerline consensus is 70+% and the BigGuy consensus is higher, go-with is 5-1. I set the methodology after noticing this effect last year, but I don't have backtested data. Most of the time though, the tendency for the BigGuy consensus to be lower than Wagerline, so we'll usually be anti-Wagerline.

 

[Rudy]

Late morning play:

SDSU

 

[Rudy]

Early afternoon plays:
TULA
ID
NEV

 

[Rudy]

Late afternoon plays:
VAND
RUTG
S FL*

 

[Rudy]

Late games:
RICE
N TX
SYR
UTEP

 

[Rudy]

Saturday Recap: 13-7 with one still pending.
Early games: 3-3
Mid-morning games: 2-2
Late morning game: 0-1
Early afternoon games: 2-1
Late afternoon games: 3-0
Late games: 3-0 (UTEP pending)

I thought I was embarassing myself until all the money was made late.

I'll do it for the NFL tomorrow. The NFL record isn't as good as CFB, but still is 56% YTD so far. Fewer plays in NFL too.

 

[Rudy]

Final Recap:
Saturday: 13-8
Friday: 0-1
Week total: 13-9 59%
YTD: 56-37 60%

The consistency so far is impressive -- not much variance week to week; weekly range 55%-67%. But that will change once I decide to start betting on it.

 

[pt1gard]

thanx rudy

thanx rudy

for all your work, much appreciated

gregg