SH Goal Filters
-
Filter 1:
Overall Matches Played >= 10
Overall H1 O0.5% >= 70
Overall H2 O0.5% >= 80
Home H2 O1.5% >=55
Away H2 O1.5% >=55
Scope: 5 GamesFilter 2:
Overall Matches Played >= 10
Overall Goals Scored Per Match Ave >= 3.0
Home FT O1.5% >= 80
Average Scored by FT% >= 80
Scope: 10 GamesBoth are restricted to current season only. I have also filtered out a lot of lesser quality leagues. The latest results for both are as follows:
-
@matt-wood Hi matt, Ive recently changed my opinion on leagues a bit so i havent found the exact sweet spot with them yet. Of the 2 filters here, i have actually stopped trading the first one because there was a big drop-off in results and it became very league specific - i was removing more and more leagues and it became a bit skewed. I still collect the data even from leagues i havent traded and i will re-analyse the whole filter in a couple of months with a completely fresh perspective. The second one is the one i concentrate on (thats the one ive just emailed you the info we spoke about).
Anyway, to answer your question - i tend to remove all the smaller leagues to begin with and dont collect data on those (regional leagues, a lot of 2nd divisions from smaller countries, slovenia, slovakia, chile etc) basically anything i think liquidity will be an issue at bigger stakes. This is a subjective but if 10 traders produced this list then they would be 90% similar. Then secondly i look at leagues i have collected data on. Originally i was quite quick to remove leagues that didnt seem to be doing well but i now try to set the filter so its tighter and can include all, or at least the majority of, leagues. Then once a league starts to get 30-40+ (ideally 50) results you can see a real reliable pattern and they can be removed. I do still collect the data on removed leagues though
-
@lee-woodman said in SH Goal Filters:
Not done an update on these filters for a little while but as theyve been collating data for a year and ive just done a bit of tidying up on them i thought id share, in case anyone is interested. Ive filtered out certain leagues for quite a while but ive just gone a bit of a further with it and some leagues just drop-off massively. So the results are below for a goal post 60':
The Danish, English and French leagues are particularly bad for both filters and a few other leagues leave a lot to be desired! If i ever get to a point of wanting something else to trade i will look at some contra trades as these are games which expect goals! I cant say i've always understood why complete leagues are worse than others, perhaps a pro can enlighten me, but there definitely seems to be more than just a correlation
Morning Lee. All good and well I hope.
I know you posted this a little while back but I was just looking at my SHG filter results to see if I was at least somewhere near yours. Do you have a minimum number of results for a league before you include it? Chrs in advance
-
Not done an update on these filters for a little while but as theyve been collating data for a year and ive just done a bit of tidying up on them i thought id share, in case anyone is interested. Ive filtered out certain leagues for quite a while but ive just gone a bit of a further with it and some leagues just drop-off massively. So the results are below for a goal post 60':
The Danish, English and French leagues are particularly bad for both filters and a few other leagues leave a lot to be desired! If i ever get to a point of wanting something else to trade i will look at some contra trades as these are games which expect goals! I cant say i've always understood why complete leagues are worse than others, perhaps a pro can enlighten me, but there definitely seems to be more than just a correlation
-
@lee-woodman said in SH Goal Filters:
@finn-kristensen said in SH Goal Filters:
@lee-woodman Hi Lee
I haven't been trading them so far, but the numbers are so good that I want to start paper trading them now and include a few more parameters to see if they matter (ELO and BTTS springs to mind as the obvious ones to check, but also a few others like losing/winning runs, etc. - some of the new options that Adam introduced last week). I will be running Filter 2 for this.
I can pull the ELO data in (and any other data from the ratings page) but i wasnt really sure what angle i would be looking for
I am generally looking at the ELO difference between the teams. I want to check if there is a correlation between ELO difference and late goals, particularly if the allegedly stronger team (by ELO rating) is not leading by the 70th minute.
-
@finn-kristensen said in SH Goal Filters:
@lee-woodman Hi Lee
I haven't been trading them so far, but the numbers are so good that I want to start paper trading them now and include a few more parameters to see if they matter (ELO and BTTS springs to mind as the obvious ones to check, but also a few others like losing/winning runs, etc. - some of the new options that Adam introduced last week). I will be running Filter 2 for this.
I can pull the ELO data in (and any other data from the ratings page) but i wasnt really sure what angle i would be looking for
-
@chris-osborne Are you referring to the first filter? Its the 1 goal games that i had an issue with aswell (ie 1-0 & 0-1) which is why i looked into it further. Hopefully if you check out your results you will see a similar pattern
-
This is interesting a filter I’ve been tracking since the summer is looking at 1 goal score lines at HT, and the prospect of 2nd half goals,.
It started off doing well but has deteriorated lately, I did a deep dive on my blog a few weeks ago and will revisit at the weekend.
-
@lee-woodman This is excellent
-
@lee-woodman hi mate I’ve started tracking and trading Filter 2 as of last weekend. Will update you at the end of the month. It’s good as although we have a lot of the same selections (comparing to my filter) there are also quite a few different ones.
-
@lee-woodman Hi Lee
I haven't been trading them so far, but the numbers are so good that I want to start paper trading them now and include a few more parameters to see if they matter (ELO and BTTS springs to mind as the obvious ones to check, but also a few others like losing/winning runs, etc. - some of the new options that Adam introduced last week). I will be running Filter 2 for this.
-
Ive done a bit more delving into the scores after pulling all the 60' and 70' scores into my results. Below are the figures for both filters:
I stated previously that 1-0/0-1 have been an issue for me so i broke it down into whether the goal came before HT or after, clearly if it came after then its a very good selection but if it came before HT then its one to swerve. Im try not to complicate things too much as i do prefer simplicity but obviously something as glaring as this makes a big enough difference to matter. If i make changes or spot patterns in data i like it to be backed up with some sort of logical reason
Im going to update the original post with the up-to-date criteria i am using, its not changed for filter 1 but its tightened up a bit for filter 2
Clearly filter 1 is better for a goal after 60' and filter 2 is better after 70'. I was hoping theyd be the same to make things easier!
I am running the same filters simultaneously to see whether the scope makes a difference and im also collecting additional criteria to see if that affects anything. Id be interested to know if anyone is tracking or even trading these
Any questions let me know!
-
@darri said in SH Goal Filters:
@richard-latimer yep its all relative to the filter. You look at lees and its not setup for just 1 goal (ill try explain). So his filter is more likely to have better results with higher scoring games based on the criteria he sets out. But on the low scoring not have such great results. You see every stat on this filter is targetted towards a higher scoring game than a game that is currently tight. For me the 0-1/1-0 are my best scorelines for the past 2 years. With late goals (72min onwards) operating around 68% strike rate for just one more goal. Remember tho 0-1 scorleines tend to be worse odds so not totally impressive. But thats indicative of the fact my filter is setup around just one more in the SH. If i get a goal between 45-60 mins i tend to have bad results because my filter isnt setup for more than 1 in the SH. So it could be a big difference for me and you his will have different angles and stats.
If you look at the 0-1 scoreline and look at the filter they dont totally compliment each other. Reason : there isnt a criteria thats encouraging the home team scoring. While the home FT ov1.5 sounds like it does its not actually meaning the home team will score score its just a macro stat that allow the game to have more than 1.5 goals. To optimise the 0-1 id setup a criteria on a seperate filter and aim for Home scored in H2 to = 80. This would directly target games that are 0-1.
Obviously its not helping the current filter but to stop the rot on the 0-1/1-0 you could implement this criteria to get a better s/r on the 0-1 as right now there isnt a direct focus on that, and why your getting good results when the game is already a high scoring one.
I missed this initially. Could explain what's happening to my own. as when 1 or 2 goals scored prior to 60 mins then further goals 70/75mins+ seem to be quite likely. The 1-0 started off good but is coming a cropper and the 0-1 started off poor but has picked up a bit.
It's an interesting train of thought. I'm going to keep paper trading for now and see where they go.
-
This post is deleted!
-
@lee-woodman said in SH Goal Filters:
I know home by FT is broader but the Ave SH goals at 1.7 also gears towards teams favouring SH
yes but if a team is already 0-1 up and is the fav away from home they may just be able to sit back and soak that up. Directing a set criteria towards the home team in this case for the 2nd half targets thats gap. Same can be said when its a strong home team already 1-0 up. Sometimes broader is good when looking at big data but targeting 0-1/1-0 2nd half based around who is more liekly to score could be a improvement. Hypothetical to say right now but at least that adds motivation/home form to your side in terms of data for the 0-1/1-0. A mix of both favs being 0-1 down and a criteria totarget 2nd half home team goals might plug the gap. Ill set that up on my own sheet using your filter and see if those two things have an impact, id rather forget just my opinion and actually get some data behind it for an actual answer.
-
@matt-wood I did that in a spreadsheet once ive exported everything. I had it set to 80 and 80 but it was quite specific so using the average is quite tight criteria still but allows for a few more options, 80/80, 90/70, 100/60
-
@lee-woodman Afternoon Lee, Sorry to be a distraction here and ask a really stupid question but I dont understand the last line of the filter 2.
I understand what you are doing with it but cannot see how you got it on the filter software, or is it something you have done in your spreadsheet post download?
Cheers in advance, cracking conversation going on here that is teaching me loads
-
@darri It may well do and im more than happy to be wrong. I prove myself wrong all the time and im happy if something comes to light that improves things for me, even if it was against what i originally thought. It could probably add in a % or few % but itd need to improve SR by nearly 10% to become a viable trade. I will add it in as an open criteria to my filter and then i can check in a month or two and see how it affects things. I know home by FT is broader but the Ave SH goals at 1.7 also gears towards teams favouring SH
