Jump to content

Lunatism's Super Experimental Series Of Interesting Bets (HTFT Lays [19/-5%])


lunatism

Recommended Posts

Hmmm, you folks all know me, so there's no need for me to go into all *that* again, but for anyone who's new, just do a search in the forum for my LSS thread and read *that* intro :tongue2 For me, this is something completely different, hence its very experimental nature, but what it hinges on at the monent, is correct score prediction - my basic idea being this: if you can get yourself in a decent position in terms of predicting the correct score in a match, everything else will fall into place, by everything, i would obivously be referring to the following markets:

  • Correct Score (well duh!)
  • Asian Handicap
  • HTFT
  • FT result
  • HT result
  • HT lays, etc

There are lots more as I'm sure you will guess! Now sure, each one could have little "extra" variables taken into account, but the ultimate goal here, is to generate the basic selections from the correct score prediction. So how do I get the correct scores? Hmmm.....grab a bit of data from Football-Data (yep, i'm a fan, and YOU should be too :ok), most importantly the number of shots and the number of goals a team is scoring, work a bit of wizardry based on these...and hey presto, you have a 1st half score prediction and a second half score prediction, then all you need to do is add the two together and you end up with a final score prediction.... My BIG deal right now, is to ignore any matches that have a hint of rivalry (for rivalry rating, see a previous post of mine that happens to be somewhere on here....:\), it's safer that way, and i'm also about to embark on some more distance related work (i'll explain it later if I use it), to determine how distance affects a match. So what i've decided to do today, is post all my correct score predictions for the next few days, and I guess if they (and their associated markets) continue to perform well, i'll keep updating this thread with *fun* stuff and my particular brand of craziness :tongue2 Obviously over time, i will be able to really see whether this is a system that i want to stick with, but since i haven't done it for a while, i thought i'd start a thread ans share my testing process with you lot (plus, i cant be bothered to do any backtesting on this on, it's more than a little complicated and i've always had an unholy fascination with anybody who has consistently made their bank swell from correct score betting....) .....So onward to selections......;)

Link to comment
Share on other sites

Re: Lunatism's Series Of Interesting Bets (Super Experimental At This Stage!!!) Welcome back to the "anoraks'" forum, Loon :ok...... interesting concept. One of my systems in the "4 Systems are better than 1 ??" thread is based on "shots". I don't try to predict correct scores.... only WINS. Anyway, good luck.... I will follow this thread with interest..... your way of expressing yourself always makes for interesting reading :loon PS your next post will make it 1,000 (I think you get an extra rotating ball for that) :D

Link to comment
Share on other sites

Re: Lunatism's Series Of Interesting Bets (Super Experimental At This Stage!!!)

Welcome back to the "anoraks'" forum' date= Loon :ok...... interesting concept. One of my systems in the "4 Systems are better than 1 ??" thread is based on "shots". I don't try to predict correct scores.... only WINS. Anyway, good luck.... I will follow this thread with interest..... your way of expressing yourself always makes for interesting reading :loon PS your next post will make it 1,000 (I think you get an extra rotating ball for that) :D
:nana glad to be back with something concrete rather than just the occasional comment on somebody elses thread! damn...i was just about to finish my post, but then i read the rest of your comment, and i realise this is somewhat of a milestone for me....erm, but what to say, to make it that extra bit special...in hindsight, i should have used my 1000th to start off the thread, and then i'd have the "Punter's Lounge Magical Thread Counts Variable" behind me too (warrants some research, huh? Have you noticed how people just hitting round numbers as they post new threads always start off so well?...i better get on the phone to JaneFromSpain :rollin) Initially, i was only looking at H/D/A prediction with this, but then i realised that if it could be accurate enough at predicting the right direction of a match, there was a chance it could do a decent job on the correct score front too. Of course, the *thing* with correct score prediction, is that you don't need a really high strike rate, and I might aswell add here, that the kind of strike rate i'm hoping (note: hope is just hope and nothing else...even if i do have a good feeling about it :tongue2) to achieve would be 15%. Now at this stage, the logistics of a 15% strike rate mean that it would be bringing in a winner every now and then, while not setting the world alight, would be doing something useful for someone....plus - hitting that minimum strike rate means that all the other markets I listed in the first post would follow suit (obviously there is no guarantee that a 15% correct score strike rate would be followed with success in the other markets - BUT, my limited trialling with this concept so far, has suggested that it might. ***Non Thread Related - But Surely Worth A Read*** One extra point to add - I put together a little something to measure the average distance travelled for any team in a league (using "As the crow flies rather than road distance, since i save the road bit for fixtures) (since i'll only be looking at league games), and then I compared "As the crow flies" actual match difference to the average (for every result that i currently have in my newly rebuilt data scraping device - almost sounds like a window cleaner, but anyways....). As i'm sure you'll all know, i generate my own fair odds, so using my own fair odds as a basis to decide what *should* happen in a match, i discovered that in matches where the away team had travelled more than 50% above the average, the strike rate for home wins drops by 17.5%....at this stage, not a huge amount of results in the database to look at, but it's an interesting point to note. PLUS, as soon as I get enough data to call this more than just a theory, i'll be applying it to all/any of my systems...applying this to the trial of this system boosted yield quite a bit ;) Fun fun fun :nana There we go then...the 1000th, i feel....not at all different, maybe i need some wine to kick off the party mood
Link to comment
Share on other sites

Re: Lunatism's Series Of Interesting Bets (Super Experimental At This Stage!!!)

Good luck Lunatism' date=' interesting :ok[/quote'] Thanks for all the good luck's lads :clap hopefully this will be something worth following with interest - and as long as it hits the minimum % i stated, we can look forward to having a bit of fun with this I never mentioned it before, but as usual with everything, i'm always looking for ways to improve my systems/ideas...so if you have any ideas for improvement or just suggestions of other ideas to try...go for it, i'll try to take on board everything that is said :ok
Link to comment
Share on other sites

12/02/2010 (results)

DATETIMELEAGUEFIXTURESELECTIONODDSRESULTRETURN
12/02/1019:45English League TwoNorthampton Town v Aldershot TownBack 1 - 16.800-1

Well there you have it, first selection out of the way ;) By the BBC look of things (http://news.bbc.co.uk/sport2/hi/football/eng_div_3/8493662.stm), it was a one-sided result but given that the home side managed to get 8 shots off tells me they were a little unlucky to only get 1 on target Selections: 1 Strike Rate: 0% Total Staked: 1 Total Returned: 0 Yield: -100%

Link to comment
Share on other sites

Re: Lunatism's Series Of Interesting Bets (Super Experimental At This Stage!!!) In much the same way that i've calculated the score predictions, i'm going to calculate the number of cards in a match, and then decide on a "safe" number of cards that can be received. Sure this idea has been prompted by the fact that possibly a red card affected the result last night, but i don't think i'm the only person who's come to the conclusion that if you reduced the number of "risk" matches (based on whether there would be lots of yellows, and maybe even some reds), then you would be left with the more "fair" game, the game that is less likely to have anything apart from your specially put aside variables affecting it And at this stage (as with the distance thing), it's just a theory, and even if my stats support it, i don't really have enough data on cards to be able to say for sure if it is affecting my matches, so it'll be something to sit on for a while once i've established the initial theory...but i'm sure that if i was to start generating a list of "safe" games each week and posting them up here (or something, who knows what i'd do with them), then somebody would find use for them...maybe lay the red (or maybe back the red for all the excluded matches?).... Anyways...this isn't me on variable overload (ok - it kind of is), it's very much along the same lines as removing any game where there is a hint of rivalry about it - i just want only the "pure" matches the remain - the ones that are less likely to be screwed up by yellows, reds, etc, and i hope you all understand that (once...if i ever do...put this filter into play!)

Link to comment
Share on other sites

Re: Lunatism's Series Of Interesting Bets (Super Experimental At This Stage!!!) Marvellous thread so far, lunatic. Some questions/doubts/ideas though. 1. Distance Are your findings based on Premier League only? Or do they work for Bundesliga, Serie A etc too? Because I gave it a thought too some time ago and analysed the match tables from football data. My finding was that the home/away win ratio for the major league were somewhat the same, all between 57% to 59%, and for at least all the 2000s. Thing is, the distance away teams have to travel is much bigger in Spain and France than it is in Italy and Germany with the latter still being significantly bigger than in England. If distance would have an effect, I'd say that the quota of home wins must be far bigger in all those other major leagues than it is in Premier League. Especially as in Premier League you have a contration of more than a dozen teams in and around London, where there is practically no travelling. Many fans and probably have several away games with a shorter travelling distances than in home games. That's something you won't find in any other of the big leagues, where there are usually not more than 2 or 3 teams in a big city like Madrid, Barcelona, Paris, Berlin, Munich, Miland, Rome, etc. within a 25 mile radius. But as football data proves, the home/away quote in England does not differ significantly from France, Spain, Italy and Germany. 2. Selection process About two thirds of your selections were 1-1. As I cannot image why it should be the 1-1 being the most valued correct score odds, there must be something in your formulas that overstates 1-1 score line. I remember, I've put up a correct score sheet in Excel a while ago based on recent from data, wins, shot and stuff from that football data websi. Did so by having some kind of usual suspect poisson formula. And most predictations, more than half, were to back the 1-1. I then did some double-checking by deriving the correct score odds from over/under odds and match odds. Findings were same. So there must be something within the poisson mechanics that does not match what happens on the pitch when a game leads towards 1-1. So if you also use poisson, you definitely have a problem there regarding the 1-1 ...

Link to comment
Share on other sites

Re: Lunatism's Series Of Interesting Bets (Super Experimental At This Stage!!!)

1. Distance
That was just an idea - not really something I'm going to incorporate because I don't feel like I have enough data behind it (sure, i could backtest the theory in the same way that I did with the card idea, BUT that would also not be with a satisfactory number of games that I had generated the odds for (if that makes sense)
Are your findings based on Premier League only? Or do they work for Bundesliga, Serie A etc too? Because I gave it a thought too some time ago and analysed the match tables from football data. My finding was that the home/away win ratio for the major league were somewhat the same, all between 57% to 59%, and for at least all the 2000s. Thing is, the distance away teams have to travel is much bigger in Spain and France than it is in Italy and Germany with the latter still being significantly bigger than in England. If distance would have an effect, I'd say that the quota of home wins must be far bigger in all those other major leagues than it is in Premier League. Especially as in Premier League you have a contration of more than a dozen teams in and around London, where there is practically no travelling. Many fans and probably have several away games with a shorter travelling distances than in home games. That's something you won't find in any other of the big leagues, where there are usually not more than 2 or 3 teams in a big city like Madrid, Barcelona, Paris, Berlin, Munich, Miland, Rome, etc. within a 25 mile radius. But as football data proves, the home/away quote in England does not differ significantly from France, Spain, Italy and Germany.
My data analysis was carried out it a basic way, I never looked at it in a league by league basis, nor did I consider the fact that some leagues would have a much larger distance to travel. What I was trying to establish, was whether distance had any relation to result or not (which could prompt further studies at a later date)...and so it didn't matter the actual distance that teams were travelling - rather the percentage difference between the actual match distance and the average league distance, just whether the distance travelled was above 50%. SO, it was not looking at actual distance in terms of actual numbers, but the relative distances. And with regards to the leagues that were considered - I just looked at the the leagues that I use for my selections (top european leagues)...but for this, instead of doing it the Cards way (where I used FD data for a season), I just used my own data (and therefore my own fair odds calculations). Maybe a bit later i'll run it over the FD data too, because the findings WILL be a little different, plus, there can potentially be a greater sample size.
2. Selection process About two thirds of your selections were 1-1. As I cannot image why it should be the 1-1 being the most valued correct score odds, there must be something in your formulas that overstates 1-1 score line. I remember, I've put up a correct score sheet in Excel a while ago based on recent from data, wins, shot and stuff from that football data websi. Did so by having some kind of usual suspect poisson formula. And most predictations, more than half, were to back the 1-1. I then did some double-checking by deriving the correct score odds from over/under odds and match odds. Findings were same. So there must be something within the poisson mechanics that does not match what happens on the pitch when a game leads towards 1-1. So if you also use poisson, you definitely have a problem there regarding the 1-1 ...
Would a simple "i'm not using poisson" suffice as an answer to this? :tongue2 I've used poisson before, but the "new idea" that came to mind, which is much more basic and over my trial period, fairly successful. In fact, I was able to use a similar method of calculation for figuring out potential numbers of yellows and reds in matches (as it follows similar principles) Thanks for your comment!!
Link to comment
Share on other sites

Re: Lunatism's Series Of Interesting Bets (Super Experimental At This Stage!!!) Nice Ideas in this thread. I actually think that about a half of your selctions from the 13th were spot on. It doesn't matter that there was no hit. The hits will come in. It's all about the odds. But I bet you know that ;) It should be possible to gain out of every market high value odds. Maybe the bigger the market odds the more room for value(if you value them with a good system/nose behind you). Sure, a good strike/value ratio related to the market is the one thing that matters. As for the correct score market. Once you have the edge you are able to print money. Today it was already a +30 points day(sure you dont have them but would gain them if this day would be repeated with some very little variables in your favour). Nice that you share your picks. I will watch it as I think there is a good system working behind you. Your approach in sorting out the games with higher possibility of unexpected disasters is top of the world. I already included/excluded the variable of derby games into my calculations. When you extract them out you even can use them in a new calc. to make out of these variables profit when you find a market that fits to these variables. As for the cards I think there is much more in it as your tables show till now. there should be one or, more possible, some factors which give these fluctations. question is which :D Could be the referee which is totally underrated in most calculations. Sometimes it is hard to include such findings into the calculations. I often get strange decimals out of calculations, which dont seem to be useable in percantage probabilitys, but as soon as I can set them in relation to odds and outcome of a game they make in the greater picture sense. hope you stay some weeks with this project! Oh and please can you explain what "poisson" you guys mean? a virus in the calculations?

Link to comment
Share on other sites

14/02/10 (selections) Through no fault of their own, today's selections have whittled down to just two:

DATE

TIME

LEAGUE

FIXTURE

SELECTION

ODDS

14/02/10

14:30

Scottish Premier League

Rangers v Hibernian

Back 2 - 1

8.80

14/02/10

15:30

German Bundesliga

Schalke 04 v Köln

Back 1 - 0

7.00

Link to comment
Share on other sites

Re: Lunatism's Series Of Interesting Bets (Super Experimental At This Stage!!!)

Nice Ideas in this thread
haha, crazy ideas leak out all the time, just are just the mildly sane ones
Nice that you share your picks. I will watch it as I think there is a good system working behind you.
I appreciate your support, and hopefully the system will reciprocate!
When you extract them out you even can use them in a new calc. to make out of these variables profit when you find a market that fits to these variables.
That's a good point, recently i've just been spotting any potential banana skin games and sweeping them under the rug, perhaps I should be paying more attention to them :ok
As for the cards I think there is much more in it as your tables show till now. there should be one or, more possible, some factors which give these fluctations. question is which :D
Any suggestions? My *problem*, is that the more variables i look to include, the more i'm having to cut down on the leagues that I'm taking selections from, and this means that i'm reducing the sample size and that any kind of solid theories take a LOT longer to come to...which is a little irritating :lol I was thinking about looking at possession, but i haven't yet found a have-everything site that covers possession in addition to all the other stuff i'm looking at.
Could be the referee which is totally underrated in most calculations.
I DO think the referee plays a big role (but don't yet have the stats to support that theory...have read a few articles online about how they do though, so you can't be alone in that line of thought), but again, there's the issue of not having referee data at hand. At present, i collect referee data on a match by match basis, but the referee isn't always listed, and so in those cases there's really nothing much I can do....
Sometimes it is hard to include such findings into the calculations. I often get strange decimals out of calculations' date=' which dont seem to be useable in percantage probabilitys, but as soon as I can set them in relation to odds and outcome of a game they make in the greater picture sense.[/quote'] Any particular problems you're having with this? I spend a lot of time trying to convert variables into usable percentages, so i may have some ideas that could be of use... Poisson Distribution was based on the work by a mathematical genius...about probability of distribution....blablablabla (just read this: http://en.wikipedia.org/wiki/Poisson_distribution) Thanks for you input :ok
Link to comment
Share on other sites

Re: 13/02/10 (results)

Selections: 11 Strike Rate: 0% Total Staked: 11 Total Returned: 0 Yield: -1100%
... bad luck indeed.... things can only get better :hope. BTW, I think your yield should read -100%, not -1100%..... if you had put 11 units on a single losing bet, the yield would be -100%,..... the fact that the 11 units was put on 11 separate losing bets makes no difference to the yield.
Link to comment
Share on other sites

Re: Lunatism's Series Of Interesting Bets (Super Experimental At This Stage!!!) Don't be disheartened by the strike rate and yield. Since you are backing individual scorelines, I guess you just need two or three correct scorelines to rebound into the positive profit region. Would it be better to have 2 scorelines instead of 1, to boost your chances? It's just a suggestion though, not absolutely sure how your system works.

Link to comment
Share on other sites

Re: 13/02/10 (results)

... bad luck indeed.... things can only get better :hope. BTW' date= I think your yield should read -100%, not -1100%..... if you had put 11 units on a single losing bet, the yield would be -100%,..... the fact that the 11 units was put on 11 separate losing bets makes no difference to the yield.
Rookie mistake huh? ;) You're 1100% right :lol
Link to comment
Share on other sites

Re: Lunatism's Series Of Interesting Bets (Super Experimental At This Stage!!!)

Don't be disheartened by the strike rate and yield. Since you are backing individual scorelines, I guess you just need two or three correct scorelines to rebound into the positive profit region. Would it be better to have 2 scorelines instead of 1, to boost your chances? It's just a suggestion though, not absolutely sure how your system works.
Good idea, i'll take a look at that - i guess it would be easier if i was looking at a series of scores and had assigned a percentage to the likelihood of each, then i could just tag on another...but i'm (kind of - it's quite complicated) looking at each teams score independtly...i say kind of, because it's only kind of - they aren't completely in isolation.
Link to comment
Share on other sites

Re: Lunatism's Series Of Interesting Bets (Super Experimental At This Stage!!!)

haha.....or maybe "a series of unfortunate events"? no just kidding...good luck!
Funny enough - that's exactly what I was thinking when I titled the thread...well, that and the fact that there would be more than one market tackled, at the moment it's correct scores - because that's been the untouchable market for me for some time, there are still plans to "reveal" some of the other affected markets (affected by my score prediction). Even though all the markets are tied to my score prediction method, they manage to do well...basically the precise nature of the correct score prediction is affecting it...but that's the fun! :loon
Link to comment
Share on other sites

15/02/10 (selections) (sorry about the delay!) it was 11, now it's 13, but who's counting :lol Lot of stuff needed doing today, but seems i'll just be able to post my selections before kickoff.

DATE

TIME

LEAGUE

FIXTURE

SELECTION

ODDS

15/02/10

20:45

English League One

Bristol Rovers v Charlton Athletic

Back 1 - 1

7.40

15/02/10

21:00

Spanish Primera División

Tenerife v Mallorca

Back 1 - 1

7.20

I've got two games today. Plus (because i fancy a bit of joy), i'll introduce another of the linked markets - HTFT lays.

DATE

TIME

LEAGUE

FIXTURE

SELECTION

ODDS

LIABILITY

15/02/10

20:45

English League One

Bristol Rovers v Charlton Athletic

Lay AA

4.30

£226.85

Link to comment
Share on other sites

14/02/10 (results)

DATE

TIME

LEAGUE

FIXTURE

SELECTION

ODDS

RESULT

RETURN

14/02/10

14:30

Scottish Premier League

Rangers v Hibernian

Back 2 - 1

8.80

0

-1

14/02/10

15:30

German Bundesliga

Schalke 04 v Köln

Back 1 - 0

7.00

0

-1

Oh so close Schalke, there was no need for a late goal to screw the prediction, but again at least the only market suffering was the correct scores (just to prove this, i've started my HTFT lays as you can see in the last post) Selections: 13 Strike Rate: 0% Total Staked: 13 Total Returned: 0 Yield: -100%
Link to comment
Share on other sites

15/02/10 (results) :rollin can't help being 1 goal away, huh?! I'm enjoying it nevertheless, i should invent a new market "almost Correct Score Prediction"!

DATETIMELEAGUEFIXTURESELECTIONODDSRESULTRETURN
15/02/1020:45English League OneBristol Rovers v Charlton AthleticBack 1 - 17.400-1
15/02/1021:00Spanish Primera DivisiónTenerife v MallorcaBack 1 - 17.200-1

On the brighter side of things (to be expected really, as i've stated numerous times :tongue2, the other markets have all still been doing well), the HTFT lays got off to a good start (well, thread start):

DATETIMELEAGUEFIXTURESELECTIONODDSLIABILITYRESULTRETURN
15/02/1020:45English League OneBristol Rovers v Charlton AthleticLay AA4.30£226.8510.95

CORRECT SCORE Selections: 15 Strike Rate: 0% Total Staked: 15 Total Returned:-15 Yield: -100% HTFT LAY Selections: 1 Strike Rate: 100% Total Staked: 1 Total Returned: 0.95 Yield: 95%

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...