"The Good Day" LTD system
-
Hi, I've been trialing a LTD system for a few months now. Its based on ELO and works off the assumption that when two fairly evenly match teams play each other invariably one of them has a "good day" and beats the other. The current sample is still relatively small at circa 300 games but the results are encouraging and I thought it was time to turn it over to the community. Please see attached spreadsheet and feel free to ask any questions.
Disclaimer - please dont follow this blindly! As this thread seems to be gathering pace and more and more people are following it I wanted to just add a few words into the opening post. Please be cautious with this, we're still testing it! it's likely there will be significant variance at some point too so use sensible stakes if your getting involved. keep liabilities to between 1-3% of the bank (that's liabilities not stake!).
Check the team news for the games selected, don't just use the selections, team news can dramatically alter the likelihood of the outcome. Try to think of this as the starting point for match selection and not a magic silver bullet. Data driven systems can sometimes miss out key variables like weather on the day or the particular referee for the match. Don't risk more than is sensible and feel free to trade out for a profit or a loss if the game isn't going to plan. Protect your bank at all times, there will always be more trades!
If your relatively new to trading and unsure about anything then please ask! either me or one of the pro's/experienced traders will answer as soon as possible (please remember were not online 24hrs a day though!).
Link below for results so far.
https://www.dropbox.com/s/orl9kn3xsujibhw/Lay Strategies 2019.xls?dl=0
-
@denis-caunce yeah Im in
-
@chris-watts Chris, I did a spot check a few weeks ago and Good Day seemed to be bobbing along. I can't remember the actual figures but, at the time, my system was out performing Good Day but I haven't checked them lately. Might do that over the weekend if I get a chance. A shocking start to August has hit my profit so it may be that Good Day is doing better now (although the overall number of draws in August is higher than the historical average so Good Day may have been hit as well).
True that VAR isn't accounted for. Obviously it won't affect all leagues but my thinking is that it removes one element of variability in a game so, in theory, ought to help us. Time will tell on that.
-
@richard-latimer I tend to queue up lay bids in Smarkets pre match providing the odds are less than about 5. Above that I will try to enter in play once the lay odds drop assuming that there is no early goal for the favourite. Often I will wait until HT to enter. I do sometimes trade out at 85mins but the stats say that overall it is better to let them run. Gives better overall profit.
Overall my stats are based on entering pre-match and letting it run. I believe there is a profit in that but so much the better if we can limit losses by entering late or exiting early. -
@denis-caunce said in "The Good Day" LTD system:
@richard-latimer Well it is more that I have developed a system inspired by the Good Day system rather than a variation of it. I did a lot of back testing on the Good Day and that showed that it had done very very well over a couple of months but not so good long term, but it got me to thinking.
So I analysed all the factors that I could find that could impact the draw %. I now have 9 different factors that I can show affect the number of draws. I have then looked at how the factors can be combined to give the best overall profit if we lay the draw. So 30,000+ games, 40+ leagues, over 110,000 different permutations and a lot of man hours later I have a written a computer system that takes in game data (including some from the game ratings) and spews out two levels of qualifiers which I am calling 'top tips' and 'value lays'. I have a 3rd level (which I call drawmageddon lol) which should generate a long term profit but has underperformed in the live test so I am sitting on that at the moment.
Obviously the system would have made a profit over the period hat I based the analysis on (it would have generated 434x stake profit over that time 338x stake profit in the period up to end May 2019. ) so that doesn't give any real confidence that it will do well in the future.
At the start of June I moved into live testing and up to the end of last month it had generated 47x stake profit. The start of this month has been shocking with a run of 5 draws out of 6 so I have given back 28x stake but still well up.
Anyway, sorry for the ramble but I'm happy to post up tips if anyone is interested. No guarantees of course but a trip to Vegas next year depends on me making a good profit!
The system ignores the first 3 games in a season so only about 10 tips/week at the moment but they will pick up significantly soon. I'm expecting between 1300 and 1400 tips/year. One downside of the system is that the draw odds tend to be a little on the high side so variance is high when we get the inevitable run of bad results. As with any lay system it is good to have a bankroll!
Have you any data for trading them? Entering ar ht? These days I'm interested in waiting for an opportunity which gives way better odds rather than taking what I can get.
-
@denis-caunce hi Dennis, great stuff would be interested to see what you’ve got. I’ve not been trading the good day system since we last spoke but I have been monitoring the selections.
Surprisingly it’s still showing a very healthy profit! I just can’t bring myself to back it since the analysis showed that it wouldn’t be profitable in the long term. Might revisit the idea once the data sample grows significantly.
It’ll be interesting to see how data driven systems perform in the upcoming seasons because VAR’s impact won’t have been accounted for.
-
@richard-latimer Well it is more that I have developed a system inspired by the Good Day system rather than a variation of it. I did a lot of back testing on the Good Day and that showed that it had done very very well over a couple of months but not so good long term, but it got me to thinking.
So I analysed all the factors that I could find that could impact the draw %. I now have 9 different factors that I can show affect the number of draws. I have then looked at how the factors can be combined to give the best overall profit if we lay the draw. So 30,000+ games, 40+ leagues, over 110,000 different permutations and a lot of man hours later I have a written a computer system that takes in game data (including some from the game ratings) and spews out two levels of qualifiers which I am calling 'top tips' and 'value lays'. I have a 3rd level (which I call drawmageddon lol) which should generate a long term profit but has underperformed in the live test so I am sitting on that at the moment.
Obviously the system would have made a profit over the period hat I based the analysis on (it would have generated 434x stake profit over that time 338x stake profit in the period up to end May 2019. ) so that doesn't give any real confidence that it will do well in the future.
At the start of June I moved into live testing and up to the end of last month it had generated 47x stake profit. The start of this month has been shocking with a run of 5 draws out of 6 so I have given back 28x stake but still well up.
Anyway, sorry for the ramble but I'm happy to post up tips if anyone is interested. No guarantees of course but a trip to Vegas next year depends on me making a good profit!
The system ignores the first 3 games in a season so only about 10 tips/week at the moment but they will pick up significantly soon. I'm expecting between 1300 and 1400 tips/year. One downside of the system is that the draw odds tend to be a little on the high side so variance is high when we get the inevitable run of bad results. As with any lay system it is good to have a bankroll!
-
Ive not been doing it since it was said that it may not be profitable longer term, im interested if you have a variation of it though
-
@denis-caunce as far as I know the ratings are for the current season, so needs a few games to be played first before any are available. If I am right I would be wary of the ratings early in the season as they may take time to settle. @Adam-Williams can confirm this though?
-
@denis-caunce said in "The Good Day" LTD system:
@chris-watts Is anyone still doing the 'good day' system. I have now developed a variation inspired by Good Day. I have tested it for a couple of months and seems to be performing as expected. Obviously still early days and fair to say that the start of this month has been mixed. Happy to post the qualifiers if anyone is interested.
I am taking the football ratings as part of my data source but notice that a number of games are missing from the lists. I have tried raising a query but not got a response. Anyone come across this and can explain? I wonder if it is missing the first game of the season for any promoted or relegated team?
Interested to know what the variation is. All ratings not there for weekend yet last time I looked.
-
@chris-watts Is anyone still doing the 'good day' system. I have now developed a variation inspired by Good Day. I have tested it for a couple of months and seems to be performing as expected. Obviously still early days and fair to say that the start of this month has been mixed. Happy to post the qualifiers if anyone is interested.
I am taking the football ratings as part of my data source but notice that a number of games are missing from the lists. I have tried raising a query but not got a response. Anyone come across this and can explain? I wonder if it is missing the first game of the season for any promoted or relegated team?
-
Any games today?
-
@keith-anderson said in "The Good Day" LTD system:
@chris-watts said in "The Good Day" LTD system:
@lee-woodman said in "The Good Day" LTD system:
@chris-watts Great run.....34 winners out of the last 37!
I’m not going to be happy until it beats @Keith-Anderson wining streak record!
Well, did it?
No!!
-
@chris-watts said in "The Good Day" LTD system:
@lee-woodman said in "The Good Day" LTD system:
@chris-watts Great run.....34 winners out of the last 37!
I’m not going to be happy until it beats @Keith-Anderson wining streak record!
Well, did it?
-
@james-h said in "The Good Day" LTD system:
@chris-watts disappointed to hear that but thanks for sharing. My results had taken a slight turn so I guess it's best to quit while we are ahead.
@Denis-Caunce any chance you would share your analysis of this system? Your posts are a really detailed and interesting read so it would be great to see in more detail how you go about determining the profitability of this strategy.
Im sure @Denis-Caunce will be happy to share the data, he's on holiday currently and we're still looking at it. We may be able to come back with an alternative based on better data. will share asap.
-
@chris-watts said in "The Good Day" LTD system:
Guys, this idea is on hold for a little while. @Denis-Caunce has been doing some phenomenal work on back testing it and has produced a very detailed report on the results of the system over the last 3 seasons. The early indicators are that it wouldnt be profitable over the long-term and that the results achieved over the last few months since we started are way above the mean expectation. We're looking at it further to see if it can be refined or if other ideas can be generated from it. In the meantime tread carefully with it as the data suggests it will revert back to a much lower baseline for predicting positive outcomes in matches.
What about as a means of trading inplay
Do results show a lot of goals still and not many 0-0's? It's a shame you can't track what effect dirty play rankings have longterm. -
@chris-watts disappointed to hear that but thanks for sharing. My results had taken a slight turn so I guess it's best to quit while we are ahead.
@Denis-Caunce any chance you would share your analysis of this system? Your posts are a really detailed and interesting read so it would be great to see in more detail how you go about determining the profitability of this strategy.
-
Guys, this idea is on hold for a little while. @Denis-Caunce has been doing some phenomenal work on back testing it and has produced a very detailed report on the results of the system over the last 3 seasons. The early indicators are that it wouldnt be profitable over the long-term and that the results achieved over the last few months since we started are way above the mean expectation. We're looking at it further to see if it can be refined or if other ideas can be generated from it. In the meantime tread carefully with it as the data suggests it will revert back to a much lower baseline for predicting positive outcomes in matches.
-
@richard-latimer Id still be reluctant to draw conclusions based on 26 outcomes. If you happen to hit 5 in the next 10 then your SR for a SH goal goes from 96% to 83%.....a huge swing.
-
@chris-watts said in "The Good Day" LTD system:
@richard-latimer said in "The Good Day" LTD system:
@chris-watts said in "The Good Day" LTD system:
@richard-latimer I’m not sure off the top of my head, I play them from pre off. I’ll work it out for you tomorrow.
Sorry, I meant as a trade rather than leaving it alll the way until the end so essentially LTD at HT, trade after goal. Also, I left out 2-2 score lines as only 2 of these and it's a 50/50 split. Always the same. If anything 2-2 is a great scoreline to back at HT if you can be bothered.
Richard - if it was 0-0 or 1-1, 2-2 at halftime there were 63 selection and 57 of them had a goal in the second half. so strike rate 90.47% if traded. Interestingly though if you left them to run then the strike rate would have only been 50.79%.
Hi Chris. Confused? I looked at your sheet and it shows 52 wins for 0-0 HT but 11 losses?