Help needed with back testing

[Justin7]

Just curios.

When you type this bullshit arguments, do you keep a straight face, or you do blush just a little.

The limits on tomorrow's (20 hours from now) 1st round matches at Family Circle Cup are $3000 on pinny and you telling me that it was just 500 on a men's semifinal of the SonyEricson?

What's next? Whales who live in US can't bet on pinny?
I'm sure, some in your regular audience are stupid enough to swallow even that.

IF you only want 3k, you bet now. If you want more, you look at the EV cost of waiting, versus the EV gain in more liquidity. Which is pretty much what I said earlier... The earlier you bet, the easier to win.. and the greater the EV (as a percentage) of each play.

And to answer your question, no. I don't bet when limits are 3k (unless that is max).

[HuskerExpat]

Just curios.

When you type this bullshit arguments, do you keep a straight face, or you do blush just a little.

The limits on tomorrow's (20 hours from now) 1st round matches at Family Circle Cup are $3000 on pinny and you telling me that it was just 500 on a men's semifinal of the SonyEricson?

What's next? Whales who live in US can't bet on pinny?
I'm sure, some in your regular audience are stupid enough to swallow even that.

You definitely do not understand the market, that is for sure. Betonline is generally the first offshore book that opens the line on college basketball games. They open usually around 1 p.m. central time the day prior to the game. Those games are limited (at least for me) at $250 per game through later that night, or even the following morning. So the whales don't really have much interest in those lines until they open in Vegas, the following morning, or at least until the limits are lifted to the maximum offshore.

Perhaps tennis has a different system. I don't, however, bet tennis, so I have no comment about that.

[hutennis]

Limits on my Djokovic example were at least 10k. And that's on Pinny only. There are other places.
Sizes on betfair bid/ask were in tens of thousands right from the start.

But that's not a point.
The point is that in speculative markets the competition for value never stops and is not anchored to any particular segment or level or time frame.
Same as in nature competition for protein is not only among lions and sharks.
Whatever lion left behind or did not get to first, for whatever the reason, jackals are will be fighting for,
or wild cats, or rats all away down to bacteria level.
Are you saying that jackal has it easy just because he is not fighting lions?
Are not there plenty of jackals to make his life constant struggle?

Rational person will never pass value just because it lesser amount of value than he hoped to get involved with. There is no reason to have the EV cost of waiting. You simply grab the bargain now and, however small it is, it will still do it's job by lowering the cost of your total bet. If you want to pass your "opener"
at -1500 you still can do it. Smallish bet you made when it was -1000 should not hurt that much, should not it? There are plenty of smarts out there on any level and at any given point in time who understand this simple concept to make sure that value is not just laying there waiting for hours to be picked up.

To say that it is easy to win betting early or simply use "easy" and "win" in one sentence, you either don't understand how the world works or you have a reason why you'd want to disseminate this kind of heresy among not the sharpest knifes in a drawer.

[Justin7]

You definitely do not understand the market, that is for sure. Betonline is generally the first offshore book that opens the line on college basketball games. They open usually around 1 p.m. central time the day prior to the game. Those games are limited (at least for me) at $250 per game through later that night, or even the following morning. So the whales don't really have much interest in those lines until they open in Vegas, the following morning, or at least until the limits are lifted to the maximum offshore.

Perhaps tennis has a different system. I don't, however, bet tennis, so I have no comment about that.

Basketball and tennis. They all work the same. If you want to bet more than $250, you likely won't touch Betonline's openers.

[cyberbabble]

hutennis -
If you've caught your breath, perhaps you could troll a little on a specific situation. I'm interested in your opinion.

The topic is the classic correlated parley. College football, large pointspread, relatively small total, parley spread/total. I assume you know about these.

The market did indeed adapt. The bookies refuse to take the bet.

Were these legitimate bets with a positive EV? Were they short run, small sample, lucky streaks? Datamined non-meaningful trends?

Do these represent a niche where the market is inefficient and can be exploited?

Is it possible to construct a synthetic bet - parley - that is inefficient by using a combination of two efficient bets - spread and total?

[hutennis]

If my grandma had a balls she would've been my grandpa.

How can you exploit anything if no one takes another side of the trade?

[durito]

Basketball and tennis. They all work the same. If you want to bet more than $250, you likely won't touch Betonline's openers.

why not, they aren't being copied. when cris opens bol just moves their lines to copy.

[cyberbabble]

hutennis - re correlated parleys

I noted that these are not bettable.

Were these short term, small sample, random luck? Datamining mirages? Bookies posting an exploitable bet that is now gone?

[hutennis]

You lost me there.

I think I understand correlated parlays.
They are great bets nobody will take.

What does this well known fact has to do with your "Were these short term, small sample, random luck? Datamining mirages?"

[pedro803]

I think he is asking you if those parlays are truly correlated, or if it is a case where there was a string of hits due to random luck, and the books banned them when actually they weren't correlated.

So I think he is asking if there is genuine value to be found in bets where the total is low and the pointspread is high, or did the books act to fast in cutting punters off on that purportedly correlated parlay

[hutennis]

I don't see where luck comes in play here.
It's simply logic.
if winning one leg makes likelihood of winning another to go up it's correlation and there is a genuine value
in making such a bet. Unfortunately, it will not be taken.

[pedro803]

ok, so that is the question, are leg one and leg two actually correlated in this case? (or are the books mistaken about it being correlated)

[hutennis]

Books are salivating about parlays. If they reject their favorite thing you can rest assured they have a pretty damn good reason for it.

[pedro803]

ok, thanks for the answer, that is the question I took him to be asking

[cyberbabble]

OP politely asked a beginners question. He has "done the work", although probably not a lot of work yet.
He should be encouraged. It's possible to give him useful help, without giving away any insider secrets. He might come up with something interesting to post here. Serious posters seem to have disappeared lately.

Menahem -

You can Google around to find historical data files that can be downloaded for Excel. You can copy and paste from a web page to a spreadsheet. It depends on what data you are looking for. In general, the older the data is, the less relevant it is. The game dynamics change due to rule changes or shortened, compressed seasons. The betting market changes because you are up against a lot of sharp people in addition to the bookies.

Probably a few years back is enough.

The NBA season is about over so you can't do much this year. Spend the off season doing backtesting of a few years of data. Read old posts here and you can find out a lot.

Good luck. I am looking forward to some good NBA posts from you. I have one and a half losing seasons for the NBA and have given up on it for now. I won't be one of the people competing against you for all the easy NBA money.

[Sportsguy_USA]

Really the future of modeling and parsing countless data points to enhance your prediction outcome leans towards machine learning. To many variables, to many changes, etc. Google is using machine learning now to predict and tailor your search results. Move away from excel (its limiting), move into a SQL dbase of sorts, hire a good overseas developer and build your app, and maybe explore the use of Googles prediction API to help with building some machine learning algo's to help.

Myself and a partner have spent the past year or so working on a sophisticated solution like this...comparing it against other expert opinions and prediction. In the end, you don't need to get the highest win % prediction from the data, just the avg. profitable wins to build the coffers up over time with proper management.