[bet on 0-0, 1-0 and over 1.5 goals]

Re: bet on 0-0, 1-0 and over 1.5 goals

Informative answer - thanks... So why would anyone bother with that strategy when they could simply lay the 0-1? Is it because you can get slightly better odds doing the strategy or is it that some people just hate to Lay?

If you are on betfair, then it saves you a reasonable amount of commission to lay the 0-1. With other bookies, it does not matter.

[Quick maths question]

Re: Quick maths question

Erm, it's only 25% if the two events are independent. Usually when one team scores, the other team will react to it. Thus it's a bit too simplistic to put as 0.5 x 0.5. You need to account for the dependence.

That is why I wrote

25%' date=' [b']more or less.

But actually, the effect is not that big. Small sample of 63,114 games. 15,021 times the home side did not score, that's 23.8%. 22,077 times the away team did not score, that's 35.0%. 5,372 games were 0-0, that's 8.5%. In theory: 23.8% x 35.0% = 8.3% As I said, the effect is not that big.

[Quick maths question]

Re: Quick maths question 25%, more or less.

[finding game percentage from successrate?]

Re: finding game percentage from successrate?

then I took (HTws + ATws)/3, who is 33,33 % to find draw To find Homewinpercent I took = HTws / (HTws + ATws + Draw) = 70,69 / (70,69 + 29,31 + 33,33) then I did that also for Drawpercent and Awaywinpercent... Is this something I can do or is this completely wrong? Thanks, Tipser:)

It's neither right nor wrong. :lol 1. If you set Draw = 33,33% then HTws = 70,69 / 133,33, keep it simple. It is always 133.33 then as HTws + ATws = 100, always. 2. Draw = 33,33% brings you to Dws = 33,33 / 133,33 = 25,00%. That's ok for average purposes. But the thing is, tha the draw probability varies very much depending on the strengh balance between home and away. It is mabye 30-35% if both team are of equal strenght. It 15-20% if teams are far from equal strength.

[finding game percentage from successrate?]

Re: finding game percentage from successrate? You could simply calculate some sort of average of both successrates. Home Team weighted successrate = (1 + Home Team successrate - Away Team successrate) / 2 Away Team weighted successrate = 1 - Home Team weighted successrate But that will only leave with some kind of DNB probability. I have no idea how to compute draw odds just from those two parameters.

[Analysing Leagues]

Re: Analysing Leagues

Hi M8, The way i see it, you need to trial your system on odds of approx even money and better. See what your strike rate produces and resultant yield, irrespective of what you bet on. Even money you need 50% s/r and better, if you find you are consistantly proving higher then you have a basis for a system!

Why do people so often look for strategies on odds of around even? Imagine 100 bets for 1 unit @ 2.00 each with BF. You win 50 bets, you lose 50 bets Leaves you with 97.50 units, a loss of 2.50 units due to commission. Then imagine 100 for 1 unit @ 1.25 each with BF. You win 80 bet, you lose 20 bets Leaves you with 99.00 units, a loss of 1.00 units due to commission. (Works also with standard overround bookmaker who offers 1.95 instead of evens, respectively 1.24 instead of 1.25) So whatever strategy you work with, don't go for too long odds, as your strike rate needs to be bigger due to bigger commission payments. (or overround deduction on odds with standard bookmaker)

[Analysing Leagues]

Re: Analysing Leagues

I'am really amazed how such huge number of quite resonable people fail to perceive one simple fact that all this historical data is already in the odds. Everything statistical is already priced in. One could manipulate the data in a million of variations and methods' date=' end up with fascinating R2 close to 1, and got a straight line on the graph. It is very useful in terms of statistical science, but it is worthless in terms of money making.[/quote'] That is why you cannot derive your formula for the selection parameters by analysing the full dataset, and then backtest those parameters for every game in the dataset. That's like kind of a self-fulfilling prophecy as your parameters then predict the outgame of a game, with such outcome influencing the prediction. You have to find a formula working the full dataset that derives the parameters for let's say week 7 from games in weeks 1 to 6, then the parameters for week 8 from games in week 2 to 7 and so on.

[Analysing Leagues]

Re: Analysing Leagues

One idea I had is I need to be more specific about the data I analyse. Rather than taking every league into my analysis' date=' I think I should do one at at time (qualitative over quantative analysis) and tailor any system to a specific league. Data from the Scottish leagues should have no bearing on the Spanish leagues right?[/quote'] Right? I don't think so. As the laws of the game are the same, both in Scotland and in Spain, there are of course many similarity inbetween the datasets. You just have to find the right adjustments among the league structures. I.e. Barca and Real are not that dominant as are Celtic and Rangers. Spanish top two play each opponent twice, not four times a year. And so on. But still, the same rules apply.

[Exit Strategy for high scoring draw needed]

Re: Exit Strategy for high scoring draw needed ...

[Exit Strategy for high scoring draw needed]

Re: Exit Strategy for high scoring draw needed Tis better to reply late than to never have replied at all. Regarding your exit strategy regarding high scoring draws: There is none. Games ending 2-2 or 3-3 happen to have goals during the match in a completely random order. There are 20 possible scoring scenarios how a 3-3 can occur. I.e. 1-0 -> 1-1 -> 1-2 -> 2-2 -> 3-2 -> 3-3. And although people often end up spamming the football forums with threads like "3-0 upfront ended in a draw, has this ever happened before?", every of those 20 scenarios has the same probability to occur, that is 5% each.

[Profit in Matches that do not end in draws.]

Re: Profit in Matches that do not end in draws.

It doesnt matter if the home or away team wins' date=' the 10-20% is guaranteed as long as the match doesnt end in draw.[/quote'] You may want to rephrase that. Otherwise I cannot see how you will end up profitable. From my understand of that quote, you make 10-20% on either side winning the match. As about 1 in 4 matches ends in a draw, the system can only be profitable as long as a draw costs you at maximum 30-60% of your stake.

[Exit Strategy for high scoring draw needed]

Re: Exit Strategy for high scoring draw needed There ain't no such thing as a free lunch. So, no there ain't no such exit strategy.

[Calculating team's total goals in footbal]

Re: Calculating team's total goals in footbal

Using only poisson distribution for soccer doesn't work well because 'pure' poisson underestimate the chance of no goals. That's why Skellam distribution is used instead. It only works when goals scored and allowed are independent' date=' in football they are linear independent so modeling using Skellam distribution is OK. I know for sure that live betting models in Bet365 are based on Skellam distribution.[/quote'] Two things: - The difference in predicting even a nil-nil between Poisson and Skellam is not that big, compared to differences resulting from good or bad pitch conditions, key players starting or missing, even floodlight or daylight. - Poisson does not underestimate the chances of no goals. Poisson needs each possible goal to be independent from each other possible goal. But goal time statistics prove, that is not true. We all know, that the likelihood of a goal in any given minute increases during the game. Goals in the 80th minute are far more likely than in 60th or 25th minute. Ok, that's only half the message. It increases only, once there was a first goal. Games still nil-nil merely have the very same likelihood of a goal in every given minute, except for the first 5 to 10. If 0-0, the probablity for a goal in the 80th minute does not really differ from 60th or 25th minute. So the probability of a second possible goal is not independent from the first goal. Therefore poisson does not work, at least regarding 0-0.

[Calculating team's total goals in footbal]

Re: Calculating team's total goals in footbal

Using Bessel functions and Skellam distribution with this geometrical averages' date=' it works surprisingly well for NBA and NHL games. Especially with NBA, the differences are very slight.[/quote'] That's because of far more "goals" in NBA than in soccer. 2.5 goals per soccer match is actually a litte low for having an accurate poisson model. Some 80 "goals" in a NBA fit far better.

[More Goals In Second Half]

Re: More Goals In Second Half

The 4000 latest PL matches: 1st half most goals: 29.3% (meaning odds of 3.41) 2nd half most goals: 43.03% (meaning odds of 2.32) Draw: 27.66% (meaning odds of 3.61)

And including 10 years worth of some other top leagues: 1st half: 28.7% 2nd half: 42.7% draw: 28.6% So basically the same.

[More Goals In Second Half]

Re: More Goals In Second Half And just for fun. What about "lay the draw" after let's say 76 minutes are played? :dude 5,368 matches were 0-0 after 76 minutes, of which 1,811 still had a goal, that's 33.7%. 6,470 matches were 1-1, 2-2, 3-3, ... after 76 minutes, of which 2,469 still had a goal, that's 38.2%. 5,368 matches were 0-0 after 76 minutes, of which 3,720 ended in a draw, that's 69.3%. 6,470 matches were 1-1, 2-2, 3-3, ... after 76 minutes, of which 4,242 ended in a draw, that's 65.6%. Make up your mind...

[More Goals In Second Half]

Re: More Goals In Second Half Ok, I took the liberty to check a bit more than the last 5 days. Let's say a small sample of 37,782 matches in several European leagues over the last 10 years. 12,438 matches were 0-0 at HT, of which 8,881 had a 2nd half goal, that's 71.4%. 4,221 matches were 1-1, 2-2, 3-3, ... at HT, of which 3,107 had a 2nd half goal, that's 73.6%. 12,438 matches were 0-0 at HT, of which 5,229 ended in a draw, that's 42.0%. 4,221 matches were 1-1, 2-2, 3-3, ... at HT, of which 1,670 ended in a draw, that's 39.6%. So at first glance, we have HT draws other than 0-0, to produce a little more matches with 2nd goal halfs and therefore ending a littles less in a draw.

[More Goals In Second Half]

Re: More Goals In Second Half Newsflash: Even matches between sides 2,000 km apart across Europe have more goals in 2nd half. Usually 55% are scored after the break.

[Trading- What am I doing wrong?]

[Trading- What am I doing wrong?]

Re: Trading- What am I doing wrong?

Come again. Back 100 @ 1.75 , lay 102.94 @ 1.70 = profit all around +2.94 ( 2.80 after commission ) Back 100 @ 1.75 , lay 97.22 @ 1.80 = loss -2.78 all around At a strike rate of 90% i think i'll make plenty.

Common sense test: Why would someone a) back @ 1.75 in the first place, but not lay @ 1.75 (or 1.76)? b) trade out in your scenario? whilst fair odds are 2.00. If I had a 90% strike rate in identifying 1.75 availlable odds going down to 1.70 whilst my data tells me fair odds are 2.00, I'd go for a 100 units lay bet and hold it till the end. Because that's +12.50 average profit per every trade in the long, not just some "lousy" +2.94 in "only" 90% of trades. So what ever do you there, is does not stand common sense test. And don't have me asking which niche in BF still is that inefficient that in 90% of cases the odds move from 1.75 to 1.70 whilst 2.00 is fair value. ;)

[Trading- What am I doing wrong?]

Re: Trading- What am I doing wrong?

Again you are mixing value from traditional betting with value for trading' date=' they are different and you cannot mix them up like that.[/quote'] It's basically the same. Only more complex in trading, as you need to determine the value in two bet, back AND lay. Doesn't it? You sure? What about a goal, whilst the trade is open, then your back bet runs to the end. As it is terminated immediately. Even if the lay bet is matched, you may practically feel the trade closed. But technically - and more important - mathematically you still run two bets. A back bet with 5.00 profit expectation. A a lay bet with 5.26 loss expectation. No, I don't. As can be read above, I'm talking of "strike rate" as the percentage of trades gone well, no goal whilst lay bet unmatched, that's 95%. Not the somewhat 50% the back bet or lay bet wins. ;) So far, I'm with you. You'll only make a profit in the long run, whilst the fair odds drop from 2.00 to at least 1.94 in your scenario. Otherwise your strategy can be mathematically proven unprofitable. If so, perfect. But you are not profitable, because you correctly identify the drop from 1.75 to 1.70 in 90% of all trades. You are profitable in that scenario, because you correctly identify a time frame, where the fair odds drop faster than the availlable odds. Meaning, more value in the back bet than in the lay bet. Or the other way round, less "undervalue" in the back than in the lay bet. That's what I'm taling about all the time. Nothing else. But just fooling around with a 1.75 to 1.70 drop, even at a 90% strike rate, is no profitable trade itself. Only if the fair value drops more than such 0.05 during the trade. ;)

[Trading- What am I doing wrong?]

Re: Trading- What am I doing wrong? I'm not talking about the strike rate. I'm talking about changes in value between back bet and lay bet. Some simple math: Trading the under 2.5 in-play midway 2nd half whilst the score is 1-1. a) Backing the under for 100 units at 2.10 whilst fair value is 2.00. (Sounds yummy, doesn't it?) 50% of back bets will have a profit of 110 50% of back bets will have a loss of 100 The math: 50% * 110 - 50% * 100 = 5 units profit per back bet in the long run b) Laying the under a couple of minutes later for 100 units at 2.00 whilst fair value is 1.90. 47.37% of lay bets will have a profit of 100 52.63% of lay bets will have a loss of 100 The math: 47.37% * 100 - 52.63% * 100 = 5.26 units loss per lay bet in the long run Grand Total : 0.26 units as a loss expectancy per round-turn. Doom! By the way: The strike rate (no goal during the trade) in the scenario above is 95%. Now tell me, how can it be, that such trader isn't profitable, - although he correctly identifies the price movement (again, no rocket science here), - although he backs a value bet, massive 5% value in there, yummy, - although he has a strike rate of 95(!)%. as the math proves otherwise? And the math proves otherwise because the trader does not manage to grab more value in his back bet than he later sets off in this lay bet. So again: It's not only about identifiying the price movement correctly. It's not only about identifiying the value. It's also about the knowlegde, when the value will be eliminated from the odds. Back before as long as there is value , lay after soon as there is less/no value.

[Trading- What am I doing wrong?]

Re: Trading- What am I doing wrong?

If the draw is available at 2.06 all a trader has to do is predict it will go down' date=' then back with the intention to lay. Wether or not 2.06 is value in terms of the match ending in a draw or not is irrelevant. You do not need to determine this, only the movement.[/quote'] Completely disagree. Let's analyze the example a bit further: If the draw odds are 2.06, and this is no Porto v Arsenal gentlemen's agreement, we are halfway down 2nd half with the current scored tied. No other realistic scenario possible. 99% of BF punters know that the odds will go down, minute by minute, unless there is a next goal, no rocket science in that. Thing is, that if you back the draw with such 3% value, wait a couple of minutes until the price is down a little, but then lay the draw at maybe 1.90 while the fair odds are 1.80, you gou are on the way to the poorhouse. Because you managed to lay the bet whilst the value has risen to 5.5%, giving you a negative total expectancy on the outcome. You do not grab enogh money from the winning trades covering your losses from trades with a goal while having an open trade, bacl bet accepted, lay bet pending. Conclusion: you need to know the value within evey trade. Because in the long run you are only profitable, if the value bought in your back bets is greater than the value sold in your lay bets. So a trader just fooling around without any idea on fair odds, just because of the assumption of an odds movement in either direction, stands no chance. Because he may have a hunch for the right direction more often than for the wrong one. But still he loses money in total, if he makes bad trades regarding the value implied beteween back and lay bet. Your trader will not be profitable in the long run, if both the odds are the fair value is - let's say - 1.40 beforesuch 6th game and 1.38 after the set. He is just changing money and paying commission. He will be only profitable, if- let's say - the fair value is 1.39 before and 1.38 after.

[Trading- What am I doing wrong?]

Re: Trading- What am I doing wrong?

Eh ??????? Trading has nothing to do with backing value. Trading is about predicting the movement of prices, not whether a price is value in the traditional (outright) sense.

Sorry, that's a contradiction within itself. Trading is all about backing value. And laying once the price movements have eliminates the value from the odds. So trading is all about backing value AND movement. Exemple1: Assume in an in-play soccer match a trader backs the draw at 2.06 with 2.00 being fair odds. A couple of minutes later he lays the draw at 1.80 with 1.80 being the fair value. Do the math, in the long run, this trader will a make fortune, as he "earns" himself 0.06 value with said trade. Exemple2: Assume in an in-play soccer match the very same trader backs the draw at fair 2.00 ... lays the draw at fair 1.80. Do the match, no profit in the long run, commission kills the bankroll sooner or later.

[Streak System]

Re: Streak System Two remarks: One. I'd hardly call winning a single game a streak. Two. Too many selections. You cannot beat the bookies by betting at every other game. Conclusion: The idea is definitely worth a try, but maybe by selecting streaks of three or four.