4 Systems are better than 1 ?? (EXTINCT)

[relf]

Hehe! Certainly not losing any sleep as the sun is rising up over here in Singapore, time for me to head to school! Let's hope I will be able to stay awake during the lessons. :zzz

[Grex]

Re: 3 Systems are better than 1 ??

Hehe! Certainly not losing any sleep as the sun is rising up over here in Singapore' date=' time for me to head to school! Let's hope I will be able to stay awake during the lessons. :zzz[/quote'] School :loon?????? How old are you ?.... remember gambling is illegal for under-18s :rollin

[relf]

Re: 3 Systems are better than 1 ?? Oh, I will be turning 23 this year. Currently an undergraduate majoring in, eh, statistics. LOL!

[Grex]

Re: 3 Systems are better than 1 ??

Gratz on the good start Grex. Im curious to know how your shot-system works. I guess its based on the assumption that a team that creates more chances is more likely to win the game, right? How many past games do you take into account for the calculation of a teams 'shots-on-target-value', and how do weight the relative opponents strength in these games? Or am i completely wrong and your system works in a different way?

Hi BeXXS.... first of all, congratulations on your HT system :clap... excellent results today (by the way, I thought you'd be sleeping at this time after being kept awake last night by your son..... try giving him a shot of whiskey before bed-time ;)).

As you suggested, the “Shots” system is a weighted system….. you compare the home teams average shots-on-target over the previous 6 games and the away team’s conceded shots-on-target, and vis-a-versa and come up with a game rating by comparing the two. You then go through the usual cr*p of analysing historical data to produce 1X2 probability equations and establish “value” of the odds offered. The “value points” system is similar. The “risk” system I explained in a previous post.

I think it is good to run a compilation of various systems concurrently….. any down-turn in one system will hopefully be compensated for by an up-turn in another system…. shall we see :hope:hope:hope. Now I'm going to bed :zzz

[Grex]

Re: 3 Systems are better than 1 ??

Oh' date=' I will be turning 23 this year. Currently an undergraduate majoring in, eh, [i']statistics. LOL!

"Statistics"..... excellent..... perhaps I can use your skills to sort out some confusing statistical analysis theories.... I'm sure you're brilliant on applying obscure Excel formulae?

[BeXXs]

Re: 3 Systems are better than 1 ??

(by the way, I thought you'd be sleeping at this time after being kept awake last night by your son..... try giving him a shot of whiskey before bed-time ;)).

I wish i was, but my wife needs the sleep more than i do, so its my turn to look after the little man again.

[relf]

Re: 3 Systems are better than 1 ??

"Statistics"..... excellent..... perhaps I can use your skills to sort out some confusing statistical analysis theories.... I'm sure you're brilliant on applying obscure Excel formulae?

I'm not exactly brilliant, but if I could be any help, you could always let me know. :) Still, studying statistics merely increases one's awareness of numbers, which probably has little to do with these soccer systems. Years of experience, trial and error, back-testing, etc probably help much more. Otherwise, all the statisticians would be rich! LOL!

[dannkez2407]

Re: 3 Systems are better than 1 ?? hi grex, in answer to your question regarding the fluctuation of odds, it kind of falls in line with team news being announced...as you would have seen the odds on united against city dropped for united considerably just before kick off, when it was announced how strong their line up would be compared to the usual carling cup team. anyway i think this is just one of the reasons..........

[Grex]

Re: 3 Systems are better than 1 ?? OK, Dannkez..... but that would not explain the drop in the Liverpool odds, where the probable non-starters were known days in advance :unsure

[dannkez2407]

Re: 3 Systems are better than 1 ?? was there anyone out the yids line up that was nt expected?

[BeXXs]

Re: 3 Systems are better than 1 ??

I think it is good to run a compilation of various systems concurrently….. any down-turn in one system will hopefully be compensated for by an up-turn in another system…. shall we see :hope:hope:hope. Now I'm going to bed :zzz

I do not agree with you here. I think you cannot bet with independent systems on the same market without forfeiting a part of the profit. For example, lets consider an upcoming match with odds of 2 for a home win. Lets say systems 1 and 2 compute probabilities of 60% and 40% for a home win, respectively. So according to system 1 a bet on a home win is a value bet, but according to system 2 its a bad bet. Which one do you trust? To put it another way: the goal in sports betting is to compute the probability of an outcome as close as possible and identify value bets according to it. How do you accomplish this if all your systems give you different probabilities for this outcome? If they are all profitable on their own, you maximize your profit by combining these systems. You could, for example, compute an overall probability by weighting the individual probabilities. As another way you could only place bets that represent value in all systems. But you definitely need to combine your systems to maximize your profit.

[relf]

Re: 3 Systems are better than 1 ?? I get your point, BeXXs. Just offering another perspective: - Match 1 (odds 2 for home win), System 1 says 60% and System 2 says 40% for home win. You bet 1 unit because System 1 says so. - Match 2 (odds 2 for home win), System 1 says 40% and System 2 says 60% for home win. You bet 1 unit because System 2 says so. A possible scenario is that you can profit from both 'contradicting' situations. Well, of course there is also a possibility of losing in both. My point is that, as long as 2 systems don't run head-on directly, both systems can produce positive yield in the long run. Not sure if my logic is right - feel free to comment.

[BeXXs]

Re: 3 Systems are better than 1 ??

I get your point, BeXXs. Just offering another perspective: - Match 1 (odds 2 for home win), System 1 says 60% and System 2 says 40% for home win. You bet 1 unit because System 1 says so. - Match 2 (odds 2 for home win), System 1 says 40% and System 2 says 60% for home win. You bet 1 unit because System 2 says so. A possible scenario is that you can profit from both 'contradicting' situations. Well, of course there is also a possibility of losing in both. My point is that, as long as 2 systems don't run head-on directly, both systems can produce positive yield in the long run. Not sure if my logic is right - feel free to comment.

I do not disagree with your point. There can be quite a few systems for the same market that produce a positive yield on their own. But to maximize yor yield you need one grand system. Lets look at it in abother way. What do we define as profit? We make profit whenever we place a bet on an outcome for which the given odds are greater than the fair odds according to the probability of that outcome (mathematically speaking, the odds needs to be greater than the reciprocal of the probability). The long run profit is absolutely independent of the result of that bet. Even if we lose 50 bets in a row with this, we will still be making profit in the long run. But this works in the other way as well: whenever we place a bet on an outcome for which the given odds are lower than the fair odds, we lose profit in the long run, even if we can achieve a streak of 50 won bets in a row. The crucial part of succesful betting therefore is to compute the probability of an outcome as close as possible. Without this probability, we are not able to identify tha 'value bets' that we make profit with. But, coming back to your example, with 2 systems giving probabilities of 40% and 60%, respectively, for an outcome, you have no clue what the real probability looks like. Sure, you can say that it is somewhere between 40% and 60%, but that does not suffice to identify a value bet in most situations. In other words: you want to go with all 'profit-making-bets' and stay away from all 'profit-losing-bets'. But what do you do when your systems contradict each other in whether a bet represents a 'profit-making-' or 'profit-losing-bet'?

[relf]

Re: 3 Systems are better than 1 ?? Wow, I feel that my IQ has increased after reading your post. :ok Well, personally, I only discovered this forum November last year so I'm pretty new to these kinds of betting systems. Prior to that, I used to bet once in a while for fun when I feel like it, so the term 'in the long run' wouldn't apply to me. Hehe. Back to the topic, my thinking is that the true probability of a team winning is not known and cannot be calculated. We can only speculate. Take a coin for example, if you want to test if it is a fair coin, you can toss in 100 or 1000 times and see if the heads-to-tails ratio is around 1-to-1. But for a soccer match, if you want to test whether your system/probability is more accurate than the bookies, you can't play the match 100 or 1000 times to see if the % home wins, % draws, % away wins coincide more with your system or the bookies' odds. This probably means the single match is highly subjected to random and unexpected events. So the truth is, we never know the actual probability. We could be having a positive yield even if our system isn't good (due to luck), or a negative yield even if our system is good (due to bad luck). Well, one could say that in the long run, everything will back to normal. Then my question would be, how many matches are considered in the long run? Meaning, if we have a positive yield, how many matches do we need to show that our positive yield is due to the system and not mere luck. Likewise, how many matches do we need to show that our negative yield is due the system and not pure bad luck. And finally, does the punter have the bank to survive in the long run and outlast 'extreme bad luck'? But perhaps I'm not that serious so I don't really dwell too much into these things. Like what one of my friends said, no matter what the system is, it's good as long as it produces profit. Oh, as for your final question, I'm unsure what I would do - probably avoid the match?

[lunatism]

Re: 3 Systems are better than 1 ??

Wow, I feel that my IQ has increased after reading your post. :ok Well, personally, I only discovered this forum November last year so I'm pretty new to these kinds of betting systems. Prior to that, I used to bet once in a while for fun when I feel like it, so the term 'in the long run' wouldn't apply to me. Hehe. Back to the topic, my thinking is that the true probability of a team winning is not known and cannot be calculated. We can only speculate. Take a coin for example, if you want to test if it is a fair coin, you can toss in 100 or 1000 times and see if the heads-to-tails ratio is around 1-to-1. But for a soccer match, if you want to test whether your system/probability is more accurate than the bookies, you can't play the match 100 or 1000 times to see if the % home wins, % draws, % away wins coincide more with your system or the bookies' odds. This probably means the single match is highly subjected to random and unexpected events. So the truth is, we never know the actual probability. We could be having a positive yield even if our system isn't good (due to luck), or a negative yield even if our system is good (due to bad luck). Well, one could say that in the long run, everything will back to normal. Then my question would be, how many matches are considered in the long run? Meaning, if we have a positive yield, how many matches do we need to show that our positive yield is due to the system and not mere luck. Likewise, how many matches do we need to show that our negative yield is due the system and not pure bad luck. And finally, does the punter have the bank to survive in the long run and outlast 'extreme bad luck'? But perhaps I'm not that serious so I don't really dwell too much into these things. Like what one of my friends said, no matter what the system is, it's good as long as it produces profit. Oh, as for your final question, I'm unsure what I would do - probably avoid the match?

Being fairly new to betting, i guess you're not familiar with the Chi-Squared Test for testing the robustness of a set of selections? Take a look here, basically you are looking at your set of "system-generated" selections, and trying to formulate whether the results you have achieved were down to chance or a good system. Of course, this doesn't go as far as predicting how well you'll go in the future, but it's as good an indication as any to let you know if you're a lucky bugger or potential millionaire (although millionaires sometimes are lucky buggers, so you never know what will happen :dude)

[lunatism]

Re: 3 Systems are better than 1 ?? ....because you can't replay a match 1000 times to determine what would happen as all the visible/invisible variables come into play, the only solution (for you and any other folks who are looking into making successful selection generating ideas) is to have a huge amount of data to support your claims. The more, the better really. For example - i'm sure most knowledgeable punters would rather take a system with a consistent 10% yield over the last 10 seasons (from thousands of selections), in place of a 50% yield from 500 games this season.... if it's something that you're serious about making cash from (as opposed to just a bit of fun), then these are the long term considerations which need to be addressed :D (fun, huh?!)

[relf]

Re: 3 Systems are better than 1 ?? Thanks for your comment. I actually do know the chi-squared test - learnt it in school. :)

[Kumquat Tree]

Re: 3 Systems are better than 1 ??

I think you cannot bet with independent systems on the same market without forfeiting a part of the profit.

But what about running independent systems on the same market and only placing a bet all if three (or at least two out of three) predict a value bet?

[relf]

Re: 3 Systems are better than 1 ?? Hi Kumquat Tree! :welcome Hmm, I'm not the expert here, but I might think something is fishy if all my systems predict a value bet. The bookies can't be that stupid, can they? :unsure

[Kumquat Tree]

Re: 3 Systems are better than 1 ?? Thanks for the welcome. Good point, relf. But in other words you say: If one of out three systems predict value, you go for it. If all three, you don't? So you bet on something that two out of your three system say, you shouldn't. I am getting I bit confused here, the longer I think about running three systems in parallel?!

[BeXXs]

Re: 3 Systems are better than 1 ??

But what about running independent systems on the same market and only placing a bet all if three (or at least two out of three) predict a value bet?

As i mentioned in the post you quoted, that would be a way to combine different systems.

[BeXXs]

Re: 3 Systems are better than 1 ??

Hi Kumquat Tree! :welcome Hmm, I'm not the expert here, but I might think something is fishy if all my systems predict a value bet. The bookies can't be that stupid, can they? :unsure

Well, actually youre not betting against the bookies but against the other bettors. The bookies adjust their odds in order to make money no matter what the outcome of the specific match will be. Lets look at an imaginative match between Man Utd and Liverpool. Lets say the fair odds for a liverpool win are at 4. The bookie might offer you odds like 3.8 in this case. Now imagine everybody bets on Man Utd and nobody bets on Liverpool. If Man Utd would win the match, the bookie will face a big loss. Therefore he wants the punter to bet on Liverpool, too. To accomplish this, he lowers the odds for Man Utd an increases the odds for Liverpool. And if he has to increase Liverpools odds to over 4, you can place a value bet and the bookie still makes profit. My main point is: its not the bookies that make the odds, its the behaviour of the crowd of punters. (I hope its clear what i want to say) And therefore you have to trust your probabilities, even when it looks like theres something you miss (as long as there is nothing significant you miss. For example, if a team plays with its B-formation, your system is obviously worthless for that match) Oh, and Grex: im not sure if its ok to have this argument in your thread, so feel free to stop us (or at least me).

[Grex]

Re: 3 Systems are better than 1 ??

Oh' date=' and Grex: im not sure if its ok to have this argument in your thread, so feel free to stop us (or at least me).[/quote'] No problem...... my brain is not in gear at the moment, and I'm having difficulty absorbing all the arguments and counter-arguments..... after I've had a few glasses of wine :beer, I'll probably be able to make a constructive contribution ;)

[Jenspm]

Re: 3 Systems are better than 1 ??

But what about running independent systems on the same market and only placing a bet all if three (or at least two out of three) predict a value bet?

..but i one of the systems is worse than the others, and removes your winners, you won't be maximising your profits..

[relf]

Re: 3 Systems are better than 1 ??

Thanks for the welcome. Good point, relf. But in other words you say: If one of out three systems predict value, you go for it. If all three, you don't? So you bet on something that two out of your three system say, you shouldn't. I am getting I bit confused here, the longer I think about running three systems in parallel?!

After spending some time thinking, here's my perspective. It depends on what your system is taking into account. Let's have a hypothetical match: Liverpool vs Man Utd. Let's say your system takes into account recent form and predicts that anything more than 2.2 for Liverpool to win is a value bet. However, in the past few days, there's a list of Liverpool players picking up injuries or suffering from a infectious disease, or perhaps their first few GKs are suspended or injured recently. Thus, Liverpool might be priced at something over 3. So, like what BeXXs have said, there are instances when the system might be worthless (depending on what your system takes in account). So whether one or three systems predict a surprisingly high value, I might want to do a quick check to see there's anything fishy. Hope it's less confusing, and thanks to BeXXs for making me learn something new everyday. :ok

[lunatism]

Re: 3 Systems are better than 1 ??

It depends on what your system is taking into account. Let's have a hypothetical match: Liverpool vs Man Utd. Let's say your system takes into account recent form and predicts that anything more than 2.2 for Liverpool to win is a value bet. However, in the past few days, there's a list of Liverpool players picking up injuries or suffering from a infectious disease, or perhaps their first few GKs are suspended or injured recently. Thus, Liverpool might be priced at something over 3. So, like what BeXXs have said, there are instances when the system might be worthless (depending on what your system takes in account). So whether one or three systems predict a surprisingly high value, I might want to do a quick check to see there's anything fishy

but unless you have decided from the off that you WILL take into account the actual "state" of the team, you will have to somehow find a way of recording all that information for every single fixture that has been put forward as a selection. And it makes back testing nigh impossible, given that you would have to trawl through team news for every back selection for both teams for however many matches you have on your records. That's not even taking into account the fact that you would also have to assign value to each player, or at least formulate a basic strategy for determining a players worth to the team. the only "even" playing field you could come up with, that would fairly (well, as fairly as possible) look at each game only on a statistical basis and generate your selection would be if you ignored injury records, suspensions and subjective team strength completely - working on the premise that quality of the team would show through over the course of the season... ...just look at Arsenal over the last few seasons, despite the stupid injury records they like to maintain, somehow Wenger manages to get them into the top four every season, etc (that's just an alternative btw, not the ultimate solution :tongue2) If there are times when your system will be worthless, then your system is always potentially worthless. And a potentially worthless system is worthless :unsure Otherwise you will also need to list up some factors where your system behaves worthlessly, and each game look at the variables to determine if "today will be the day that the system is worthless"...and if you need to do that in the first place? Waste of time If you KNOW when your system will be worthless - simple, add criteria to your system, so that it ignores games where your select "worthless criteria" are all positive...

[BeXXs]

Re: 3 Systems are better than 1 ??

Hope it's less confusing, and thanks to BeXXs for making me learn something new everyday. :ok

Thank you, relf. Im glad that i can help.

[Grex]

Re: 3 Systems are better than 1 ?? OK, I have a few minutes….. so I’ll give some input to the “Great Debate” about the value of using more than one system concurrently. No matter how much hard work you put into a system, obviously you can never be even 80% accurate with your probability predictions (otherwise we would all be millionaires and all the bookies would be bankrupt). All systems are inherently subjective and, therefore, you are bound to get differences between the various systems. The way I do it is to establish a range a value % where there is a marked positive return on your betting investment…. with one system, the range may be 100% to 125%, with another it may be 90% to 115% (I’m mot concerned if the value is less than 100% - if value bets between 90% and 100% theoretical value produce a positive return, then, for me, they are good value).

Take an example….. Spurs v. Fulham…. using System A, the probability vs. odds of a Spurs win falls within the cut-off criteria and, therefore, it is a selection. However, using Systems B the value falls just outside the cut-off criteria and would not be selection for that system. However, even with System B, there is no reason why Spurs should not beat Fulham…. there are obviously many matches outside the selection range that win, just not so many as within the selection range.

Therefore, I consider that the fact that a particular match only satisfies the selection criteria of one system, and not the other two, should not detract from the level of confidence in placing a bet on that selection. I am sure that the statistics “purists” will disagree (for example, I'm sure Loon will tear all my comments to shreds ;)) but I’m going to stick with my conviction….. until everything goes pear-shaped, of course.

In fact, to underline my confidence, I will introduce a fourth system :loon based on “value goals". It’s something Jenspm raised last week…. and I reconstructed some old Excel files I had on the same principle (I shared the files with Jenspm, and he will probably enhance it and produce his own system).

[relf]

Re: 3 Systems are better than 1 ??

...there are obviously many matches outside the selection range that win' date= just not so many as within the selection range.... Therefore, I consider that the fact that a particular match only satisfies the selection criteria of one system, and not the other two, should not detract from the level of confidence in placing a bet on that selection..

I agree very much with this point! Looking forward to your 4th system. :ok

[Kanga]

Re: 3 Systems are better than 1 ??

do not agree with you here. I think you cannot bet with independent systems on the same market without forfeiting a part of the profit.

Imagine you have 2 systems both profitable .... one is a back system one is a lay system .... system 1 says back Liverpool at 2.20 .... system 2 says lay Liverpool at 2.00. If both bets are matched on betfair (or wherever, but I chose betfair for the ease of a visual example) you will see that you have still made a profit even by following 2 totally independent and apparently contradicting systems.

But you definitely need to combine your systems to maximize your profit.

As long as whatever systems you are following have sound reasoning and make longterm gains, then you need to follow them even if they seem to conflict.