-
Comment -
Justin,
Do you assume that public money doesn't move lines?Comment -
simple question: suppose a small time bettor w/ 10k br using the same exact model that lead him to bet 500$ on NFL ml +170 and 500$ on Aussie Football ml +170. the closing line for both sports was +100/-120 (for simplicity sake we'll assume they are both using a 20c spread) Is his decision to hedge(partially or not) the same in these two matching scenarios?Comment -
If the market moves against your model, your model was wrong on that play. If this happens with regularity, your model is crap.
The market has almost nothing to do with public money. The market is mostly defined by winning bettors, that will bet as much as they can at a price they think is profitable. When they bet, they are right and the price given to them is wrong. They have the capacity to bet this because their models have destroyed markets for a long time, and grown their br's exponentially until limits hampered growth.
I apologize if I sound frustrated... but I think our premises are very different, and most of the conclusions we draw will vary differently based on our premises. My starting point is that the market is efficient, and this efficiency is driven by smart, big players.Comment -
That is one faulty theory, Justin. First off, just because a subset of games are unbiased, does not mean that individual contests are not. It can, however, mean that the biases in some of those individual games cancel each other out over time. Also, to assume that the market doesn't move against even the very best models is asinine. There are runners here that lose their jobs for not repopping numbers after they move against them. If your theory was correct, every single line would move linearly. There would be no return movement. I don't know if there's a SINGLE game where that's the case. I'm not sure where you get this stuff from, but you really need to rethink it.Comment -
Huh? In a subset of games, I destroy the market. That subset, as a whole agrees with the model. The model of all the winning players destroys markets. On average, they crush the closing lines. Are you suggesting that these long-term winners are just getting lucky on a sequence of individual games, or am I just not following you?
If you ignore manipulation and differing goals for volume, you are correct. Syndicates manipulate. Syndicates have different betting thresholds. Do you really want me to go here to explain non-linear movements?Comment -
pls don't take this the wrong way justin, but your position that market all must be efficient etc etc really reminds me of the time when we met w/ the underwriters of CDS in AIG. they were all so totally convinced that their model can never be wrong, that the CDS market was highly efficient, so they were surprised when we told them we only buy protection never sell protection, one guy coined that we were throwing away free money.. really, just correlation exists between market efficiency and cl does NOT mean it's always 100%.Comment -
How about this: in major "efficient" markets the hedger can get down at no-vig line and therefore must buy back about 80% of his position while at small "inefficient" markets he must buy back 100% because he does not know what side is +EV. Heads, TriMoDa wins, tails Justin7 loses.
Comment -
fatalityComment -
I'm talking about market efficiency. You're telling me that every game is efficient because the whole is. I'm saying that's not necessarily the case.
So movement against the grain is simply manipulation? What does that tell you when THE MARKET IS MOVING AGAINST YOUR MODEL??? Is game day movement in the NFL ever market manipulation? Is it always linear? And does it always move to a point where every line taken shows value according to the closing line?If you ignore manipulation and differing goals for volume, you are correct. Syndicates manipulate. Syndicates have different betting thresholds. Do you really want me to go here to explain non-linear movements?
You're taking some ludicrous shots in the dark here.Comment -
As I said before: interesting.
My view: the market is almost totally ruled by big players, i.e. those who the market perceives as winners and wager enough to make a difference, and the steamers. These folks (big players) can change from year to year and sport to sport. Public money can slosh around but is (almost) totally ignored. A large portion of the big players are or use originators who are based on modelers (who, BTW, generally are the most consistent winners).
If your selections are then in the same general direction as those dominant modelers or originators, you will generally BTCL (usually because of the steamers), if not, you won't. Now, and this happened to me; if you start winning, and gain recognition, you then start BTCL because you start driving the market.
If you are not BTCL, but are winning, you are not necessarily wrong, you simply have not been recognized by the market. If you are big enough and win enough, you will gain recognition and push the line with your release only to see the steam players recognize your plays resulting in the line closing in your direction.Comment -
Have to agree with monkey here. Yes, at least in major sports, the CL is efficient ON AVERAGE, but it takes a huge leap of faith to go from there to Justin's contention that EVERY CLOSING LINE is efficient. If that's the case, what's the point in modeling anything???Comment -
Very good discussions and they are civil as well.
Question for Justin7. If your model is wrong on a play, do you know why it is wrong every time? If you don't know why and thus can't tweak the model, do you still bet all of its plays as long as it generates more btcl plays than not?Last edited by wrongturn; 10-13-10, 11:57 AM.Comment -
Lots of reasons. If I run the model on Wednesday, and new information comes out Thursday (weather, injuries), my original play looks cold look bad. I don't ID every bad line move... but if you BTCL on 70% of your plays, you're going to win.Very good discussions and they are civil as well.
Question for Justin7. If your model is wrong on a play, do you know why it is wrong every time? If you don't know why and thus can't tweak the model, do you still bet all of its plays as long as it generates more btcl plays than not?Comment -
J7 has won the thread already and its not imo. They are just plain wrong and there is nothing really to argue.Comment -
OK, thanks, I now understand the logic of the argument, though I don't agree with it -- since it essentially says that the only way to be successful is to bet (on the "correct" side) on events with enough line movement to counter the vig.
I think people who believe this are fooling themselves -- don't their models ever make a prediction that is signficantly at odds with the line, regardless of how much it moves? In those cases, do they trust their model (in which case their model-projected EV won't change all that much from line moves)? Or do they conveniently conclude that their model just happens to be wrong on that play? If the latter, it's a pretty odd methodology -- assuming your model is correct if it's "somewhat different" from the current line (therefore allowing you to bet into that line), but that it's wrong if it's "too much different".Comment -
Don't understand what "CL is efficient ON AVERAGE" means. Some CLs are definitely more efficient than others. But none can be more than 100% efficient. If you mean the CLs are accurate to true odds on average, I agree. But we need to treat each CL as efficient as it can be when judging plays in major American sports, unless the game is Super-bowl, NCAAB Final etc, which huge public money can force books to move line out of place in order to reduce large exposure in a single event.Comment -
Yes wrongturn, it is essentially impossible to pick and choose games saying the Seahawks closing line this week is efficient and the Patriots closing line this week is inefficient. If you actually knew this you would not be posting on sbr you would be getting down the 70k pinnacle allows you to bet on the correct side of the inefficient line.Comment -
I disagree. A model that projects a line on an individual game basis is essentially a projection of whether the line for that game is "efficient" or not. Whether you then choose to believe your model or the closing line is another question. The problem is that you never KNOW FOR SURE whether it's your line or the posted line that's "correct" for that particular game (or whether it's somewhere in between).
What is so hard about understanding that a market may be efficient on average (ie, over a large number of trials), but not necessarily for each individual trial?
And no, you would not necessarily be putting down the pinny max even if you did "know" that your line was correct. Kelly anyone?Comment -
If the cl outperforms the opener 55% of the time that still leaves 45% where the opener ended up closer to reality than the cl did. Tough to chalk 100% of that up to noise... sometimes the market simply moves the line in the wrong direction.
Is it more important then to be able to forecast which direction the market will move the line or whether or not the line is accurate? Doing both together is ideal so you can always get the best line on the correct side whether that be an opener or a closer. Most important is to be on the correct side though.Comment -
I make my living in parimutuel pools and I would have to believe that Justin is correct. Smart money in my game is very smart and actually drives prices in both directions. Where there is big money, there is manipulation and it would make sense that lines are being pushed in both directions by the smartest and largest money. good luckComment -
If in one period, a model generates 100 plays that beat CL 70 times but only wins 45 times, it is still a winning system, according to J7, which I agree. No one knows for sure that a line reflects true odds. I agree. But it does not mean treating each CL efficient is wrong, because the final evaluation is on a large sample, not a small one.Comment -
I don't disagree with you.
It's just two different ways of looking at it:
One says that you treat each CL individually as efficient, in which case you're fixated on beating the CL by picking off inefficient lines before they close toward efficiency.
The other says that some individual lines may never close to efficiency, in which case you're fixated on finding those you believe to be the most inefficient, line movement be damned.Comment -
Comment -
Yes go back to PT please MF.Comment -
The earlier you bet, the greater the chance that new information will come out. Ergo, it is almost impossible to BTCL 100% of the time if you are betting a week in advance.
On the other hand, the earlier you bet, the lower the limits, which means fewer pros have bet into the line. If I bet as early as possible, my average margin over the CL goes up, even counting games that I "failed".
There might be some small amount of value at close, but not nearly as much (from a percentage perspective) as hitting openers, even if you fail to BTCL 30% of the time.
I'm not quite sure how you expect to BTCL 100% of the time. If you did, I doubt you'd talk about it here.Comment
