2018 Summer League HDAFU Tables

45 thoughts on “2018 Summer League HDAFU Tables”

1. Having erratic winter season at its best, I was not sure if I will go for 2018 summer season. With its start right on the nose, I decided to give it a try. Today, while doing data filtering I discovered one thing that appears to me a serious flaw of how historical odds are collected for HDAFU. The league in concern is Swedish Allsvenskan, but I’m sure it’s gonna be the same with any other league. I expected odds are collected this way: once the 6th season is over, data of the oldest season is deleted and the data of the last season is added to the remaining 4 seasons. Nothing changes in those 4 older seasons that are still relevant for the next season.

   Right Winger, why HDAFU tables are completely redone for all 5 seasons every year with significant changes instead of deleting data of last, no longer relevant season and adding data of last fresh season? How do these changes in odds and number of matches come up? The odds range of 5.29-6.25 dog was profitable in 2012-2016 and fine to bet in 2017, but for 2018 season this strategy “failed” because different data was taken into analysis, not because dog teams failed to win often enough. Relying on 2012-2016 and performance of 2017 season strategy was profitable 6 seasons in a row, relying on 2013-2017 HDAFU just 3 recent seasons were profitable. Quite a difference!

   I sincerely hope I missed something and I’m wrong, but I’m afraid I missed nothing!

   Reply

   1. Jo,

      I read all of your post, and theres definitely a lot of information to digest. It’s a very interesting discovery you have made and hopefully RW can clear this up.

      If I understand you correctly, you are saying that in the 2017 summer league tables, the 5 season data set is comprised of data from seasons 2012-2016. Then in the 2018 tables, the 2012 data is dropped and we now have data from seasons 2013-17. However, looking at say the 2013 data set in both the 2017 and 2018 tables, the data differs, and that applies to the other seasons too, 2014, 2015, 2016, 2017. So, for example, in any given year, the same game is showing up with different odds depending on the table you are looking at, and this is happening multiple times with other games. Am I following you correctly?

      From all the reading I had done on the website articles and the blog I was of the understanding that prior to compiling the latest HDAFU tables, the “most recent” seasons odds were scraped from odds portal, verified and checked for accuracy before being used in said tables. I assumed it would only be done for the latest season as why would it need to be done for prior seasons if the checking and verifying had already been done. It seems to me to be unnecessary to repeat what has already been done. But then how can odds data differ for specific identical matches across different editions of the tables?

      If one seasons odds are dropped and then the latest 5 seasons taken into account, why would all 5 of those seasons matches and odds need to be re-scraped, checked and verified again? Why not just add the latest seasons odds to the other 4 seasons that are already in existence. Surely this would be the most consistent way?

      So if what Jo has highlighted is correct, as I can’t check this myself as I don’t have last seasons summer league tables to compare to, what would be the reason for obtaining a whole new 5 season data set? Also, as JO is suggesting, the same matches in one table are showing different odds than in another, why are they different? Does odds portal change the historical odds? I recall at one point there being mention of odds portal dropping bookmakers from the panel as time goes on, so comparing odds obtained a few years ago with odds obtained now for the same matches, may indeed show different highest odds. This would of course affect whether a strategy looks viable now as compared to back then.

      So I’m confused as to the validity of JO’s discovery and whether this is going to negatively affect us going forward.

      I would really appreciate this issue being cleared up.

      Thanks.

      Reply

      1. Hi Simon,

         regarding odds, yes, you follow me correctly. But that does not repeat with every match, in my example I provided with 2013 season there are 4 identical matches (same dates, same teams) in both tables that have slightly different odds.
         Also, slightly different odds for the same match can mean match from 12-16 table, lets say, 2014 season, which is still relevant for 13-17 table, dissappears in 13-17 table in same specific odds range. And another match of 2014 season that is present in 13-17 table may not be present in 12-16 table in that same odds range, because its odds are a bit different, it “moved” out to another odds range. This kind of match “traveling” can and will affect results from little to serious.

         You said: “I assumed it would only be done for the latest season as why would it need to be done for prior seasons if the checking and verifying had already been done. It seems to me to be unnecessary to repeat what has already been done. But then how can odds data differ for specific identical matches across different editions of the tables?”

         That’s what I’m asking Right Winger’s, why to waste your time to scrape odds from 0 for all 5 seasons, check them and etc. instead of deleting the oldest season and adding the newest one? The source of problem lies here in my opinion: since the odds and number of matches differ in both tables, it could be something wrong with oddsportal, the odds maybe change in its database over time. Or, the problem is with checking and verifying odds – their decisions and opinion what are correct odds for particular match may be slightly different after 2, 3 or 5 years. And that “slightly” different can make a huge difference as it is with Allsvenskan underdog in my case. Or maybe issue lies in both of them?

         After this discovery I finished filtering Allsvenskan odds I was interested in and moved on Veikkausliiga. And guess what? You’re right – I made identical discovery. This time discrepancies did not make strategy to fail, although there are significant differences in strategies performance, but both discoveries have same issues in common.

         This flaw should affect our results, in the first case it ruined valid odds range, valid system, in the second case, it didn’t. I spent several hours to recheck Allsvenskan discovery in disbelief and to explain in my comment everything. So, just short example for Veikkausliiga: profit in 2012-2016 hdafu of 2016 season was 1139. Profit of 2016 season in 2013-2017 hdafu is 2736 units. This is massive. Overall profit of 5 seasons is 5451 units in 12-16 table and 5518 units in 13-17 table.

         Such differences will definitely affect us. Tables worked for me last summer season with this flaw, but in the long run we will probably win less with it than without it. If we win at all (knock knock)!

         I don’t promise anything, since I can’t wait to see how Right Winger will explain this, but I will probably email relevant excel sheets to anyone interested so you can check on your own, since this is really serious in my opinion. It will either be amended HDAFU copies left with relevant data only and unusable so no one gets usable table for free. In case amendments will corrupt relevant data I will compile fresh excel sheets with data needed. I hope Right Winger is fine with that?

         I’m pretty sure I will find something like this in other leagues too.

         Reply

   2. Hi Jo and Simon,

      Yes, you will find that I have been through every set of odds on Oddsportal for every single game in each league offered for sale and corrected everything manually for all five seasons.

      Previously, we were indentifying definite errors by looking at overround figures and adjusting where we could see was necessary.

      There are so many problems with Oddsportal now that I have had no choice but to wipe the slate clean and look at every game individually.

      For example, Asian Odds, a newcomer to the Oddsportal compulsory list has whole tranches of odds the wrong way round, which throws out wildly Oddsportal’s statements of what are the highest and what are the average odds. The away odds were shown as the draw odds and vice versa. An absolute total mess and a hugely time consuming job to correct everything according to time stamps.

      The problem with Oddsportal is that they are the only free resource offering time-stamped odds, but they are only interested in affiliate sign-ups to their bookmaker portfolio, rather than providing any worthwhile resource for punters.

      However, I suspect that the majority of punters using Oddsportal for historical odds do not even realise there are problems, let alone possess the experience to identify the different types of problems before correcting them.

      Once the extent of the job in hand became apparent I was able to correct at the rate of 60-85 matches per hour, so you can work out for yourselves just how long this took for the nine leagues on offer.

      The 2018 tables, including the standard deviation adjustments, should therefore be the most accurate we’ve ever produced. When I saw each season’s final overround figures so much closer together, season after season, I knew that it was a good job well done, and very worthwhile.

      Personally, I’m excited to see what happens with the summer leagues this season. It was good last year, so let’s cross our fingers and hope for an even better performance this year.

      Sorry you’ve spent so much time writing these comments. I still owe a couple of articles and an upgrade to the User Guide to incorporate the new features in the tables, but this should be done before the end of the week.

      And sorry Jo, I’ve had to cut some of your comment above as it is giving too much away to other people for free. I hope you will understand. (I have copied and saved the rest of your script if there is anything you wish to discuss for reinclusion – don’t worry, I haven’t deleted it totally!).

      All the best for now.

      Reply

      1. RW,

         Thanks for clearing this issue up. It did seem to me upon reading Jo’s comments that the most likely reason for his discovery was that all the odds for 5 seasons had been checked again. But of course due to the time involvements in doing so I thought maybe it could be something else at play. But as it turns out you did indeed re-check all the odds, a very painstaking and laborious task indeed.

         I have to say it’s commendable and very reassuring to me that you go to great lengths to provide such a quality and accurate product here, that improves each year.

         Thanks for all you do to help us out.

         Reply

      2. Hi RW,

         sorry for revealing too much than you can accept, but I honestly thought I have to do it to make things clear. It seems I was lucky with Swedish underdog as it was not so trustworthy to bet on as it appeared.

         What season from did you start to check and verify odds for every match individually both for winter and summer leagues?

         And don’t forget you owe me a response by email on other request, it’s not urgent case, so do it when you have time. Thank you.

         Reply

         1. Hi Jo,

            For the summer leagues, you now know that the five seasons studied in the simulations, 2013-2017, are the first with 100% checked odds.

            Once 2018 is over, I shall add the odds to the list – 2013 will drop out – and the 2019 simulation will comprise the five seasons, 2014-2018.

            The data for 2014-2017 will remain unchanged for the 2019 summer league tables. I hope this is clear.

            The current winter league odds will also see a total overhaul at the end of the season so that they are on a par with how the summer league odds are collected.

            However, here, several leagues already benefit from having had all their odds checked: EPL, Australia, Greece, and a few others – I would have to check and come back to you.

            Regarding your private mail, I’ll reply later this evening.

            Reply

            1. Hi Right Winger,

               can you let me know which leagues of winter 2017-2018 season had their odds collected individually for each match?? Season is about to end, I’m checking results and I think it may be useful to know this. Thanks

            2. Jo,

               Regarding your question of 19 May 2018, sorry, but no time right now to check back and find out what was what last year – GDPR compliance has overwhelmed everything we are doing right now.

               I’ll come back to you next week, but feel free to remind me if I’m later than another seven days.

               Sorry, we have to comply with GDPR or face shutting down the sites temporarily until we are compliant, which is more of an issue with our German sister site, Fussballwitwe. German lawyers are just great at trawling the net to see if they can make a few easy bucks…

2. I had wondered about this myself but I wasn’t a dedicated enough detective to properly look into it. Fair play to Jo for doing so and to RW for the comprehensive response!

   Reply

3. Hi Guys,

   Haven’t heard any feedback from those of you who are involved in a summer league campaign. How are you guys faring?

   My story is much the same as the start of the winter league campaign in which it was week after week after week of losing until the point where I gave up for paper testing – at which point it actually started winning money (just my luck – I always seem to lose when real money is at stake!)

   As it stands my portfolio is running at a loss of 15 units of profit. The hit rate is again the main and sole factor for the losses. It’s a good 10% below the average forecast hit rate. In fact, with 9 systems running this first half, only 2 are in profit, and that profit is about equal to the stake size.

   14 weeks into the campaign and I am seeing time and time again that any time a little forward progress is made, its immediately followed by a regression that sees me slip even further into the hole. 3 weeks ago I was 16 units of profit down and had a great week and half following that which reduced the deficit to just 1 unit of profit. The last week and a half has seen the hit rate completely nose dive and send me right back into to major losses.

   As I say, just cannot get any traction and keep slipping back and back and back and back……

   Reply

   1. Hi Simon,

      my portfolio has no better news than yours. winning strategies are just about the size of the stake. 3 winning weeks out of 12, hit rate is just 28% out of 42% forecast. At this time a year ago I was about to jump from negative balance to positive and had stable climb up since then all the way to the end of the season, this time it’s nothing like that and I’m stably going down from 98th bet. I have placed 232 bets in total. Highest balance was +656 units, now I’m -2300 units. And I have a strategy that is -1700 units down, its longest losing streak over 5 years was 11 bets, at the moment it has 19 and still counting… Its hit rate was pretty stable over 5 last seasons, all of them in profit, 4 out of 5 in 4 figures and this kind of performance deterioration is literally out of the blue. Had I chosen something else, maybe it wouldn’t be so bad, one single strategy ruined my situation really badly. I’m in this kind of situation for a while again: should i stop this strategy? but what if the next bet wins? So, I keep placing bets on it, and what I get is another loss…
      This weekend is almost 500 units loss, and this week is almost certainly a loss.

      Reply

      1. Jo,

         Sorry to hear your campaign isn’t faring too well either.

         I know with my just concluded winter campaign that in hindsight the decision to stop betting and instead paper test was the wrong decision – because even after 11 straight losing weeks and a maximum deficit of 35 units (£3500), it bounced back to end up with a profit of £5800. So I know that I left a LOT of money on the table and for that I’m kicking myself. The point being if one can just shut off emotion and not be affected by seeing money disappear, in theory they can ride out losing streaks and come out the other side with a profit.

         Here, with the summer league, with that experience in hand, I’m seeing a similar scenario unfold, in which I’m steadily sliding further and further back into losses. The last 2 weeks have been so bad and at this point my losses are 19 units/£1900. This is after 36% of the forecasted number of bets have been placed. The hit rate is actually 12% lower than the forecast. 2 of the nine 1st half systems have already concluded and are losers. The remaining 1st half systems that are losing have almost run their course too, so I’m potentially facing having almost all my 1st half systems end up losing. That leaves me needing all the 2nd half systems of which their are 9 also, to win to claw back the losses.

         I remember RW pointing out that for a system to record its worst result in its 6th season, the percentage chance is 16%. Now on the basis that most of my systems were 5 out of 5 seasons with profit, if all my 9 systems do end up losing, then they will record their worst season in the last 6…we are talking lottery winning percentages here. Surely all 9 of my 1st systems will not lose. If they do, I’m one unlucky guy!

         I’m in the quandary again of saying “Well, this is heading south real quick with no glimmer of hope, do I just pull the plug?”

         Or, “Do I hang in there, write off the 1st half of the season and hope that the 2nd half salvages some losses?”

         Logic tells me….I paper tested through to the end of a 1300 bet portfolio and it won £5800. Yes it was volatile, and yes it was a rollercoaster ride, but it did settle down, and it did end up making a profit. I guess I just need to continue placing the bets and try to put aside the worry and negativity that stems from long time losing streaks.

         Reply

         1. Simon,

            pretty much same story regarding my winter campaign. Instead of making 2700 profit, I ended with 2000 deficit and can only blame myself. Volatile performance broke my nerves and I started doing mistakes that were very costly. it appears this summer will be no easy going, but I think I will run everything as I’m supposed to no matter what, it’s hard to accept that I failed winter portfolio, but not portfolio failed me and I don’t want to repeat that again. Last winter looked very bad, I had no hopes for any profit, but it made it. Thats a reason not to worry about this summer too much, it must be fine in the end.

            Reply

            1. Jo

               Update on my progress.

               Hit rate has plummeted over the last 3 round; of the last 49 bets, just 11 have won. 1 bet out of the last 15 has won – that pretty disastrous.

               Prior to the 49 bet stretch I had got back to breaking even, but this brutal stretch has left me running at a loss of £2700 based on £100 per bet.

               Suffice to say I’m pretty despondent about the situation and very frustrated that yet again – as per my winter campaign – the hit rates across the board is much much lower than the forecasted amount. The hit rate is 13% lower than forecast.

               I’ve looked at everything to see if there are places where I’m making mistakes. I can’t see any. I’m not missing any betting opportunities, and am placing bets within 5 to 10 minutes prior to kick off, so I’m not getting bets included that shouldn’t be. I’m using pinnacle and marathon bet for my bets, so I’m getting the best odds on the market in most cases.

               I wonder if there is some error I’m making in constructing my portfolio that results in this crazy volatility where the hit rate just drops to ridiculously low levels. All the systems I chose made good profits, win at least 4 of 5 seasons, show consistent hit rates, good number of betting opportunities e.t.c.

               Is it just a case of bad luck that all my systems are losing at the same time? Is it just a case of me having to stick to my money management plan and hope that with a properly allocated bankroll that I can recover from this? It always seems that when you get a disaster like this occurring that the whole venture is doomed to fail.

               Ironically as mentioned, while putting real money on the winter campaign games, I lost every week for 11 weeks. As soon as I stopped betting real money, the systems began recovering and started winning again. That experience makes me loathe to drop out again, but obviously I’m not happy with losing a lot of money in a short space of time. Seems I lose when real money is on the line, but paper testing produces amazing results….go figure!

4. Hi Simon,

   for some reason I don’t have “Reply” button and can’t reply directly to your 10 June 2018 at 7:02 pm comment, so I will reply here.

   Yesterday I finished my 15th betting week and I didn’t ever have more disastrous week as this one: 21 bets and just 1 win. My stakes were already reduced quite a lot since portfolio is losing, another big reduction will be made if I don’t stop doing this at all. 3 more weeks like this and I will be “betting” 0 unit stakes. I’m -3900 down, last 3 weeks were very bad, lost 2000 and with already significantly reduced stakes.
   We were all losing with winter portfolios, it seems we are all losing again and I don’t see any mistakes to blame. I placed every bet I had to, and everything is going down. It’s the volatility this system seems to have little resistance against. And compared with winter, this summer is performing significantly worse. Summer’s highest profit (+650) was 5 times lower so far compared to poor winter season (+3600) and to get this low (-3900) “required” 280 bets compared to winters 380 bets.

   Reply

   1. Hi Jo and everybody else,

      can you please share the number of systems you are playing at the same time with the expected Hit Rate, expected Yield and number of bets each round.

      Some list like this example:
      (1) expected hit rate: 18%; expected yield: 38%; expected number of bets each week: 5
      (2) expected hit rate: 52%; expected yield: 24%; expected number of bets each week: 3
      and so on…

      I’m currently in the process of preparing an article about the variation of winning and losing streaks when playing systems with too many systems that contain low hit rates (probabilities of winning) and it would be very nice to have a few real life examples to base the article upon. No need to give away the system, or the league.

      It’s all about risk diversification and as I’m also currently substantially upgrading the HDAFU tables to get them ready for sale before the Winter Leagues kick off again, it would be also very useful to know on what I should concentrate on to help you guys to find more reliable systems in the future that are less nerve wracking. Therefore feel free to make suggestions what you would like to see in the upgraded tables.

      Best wishes,
      Soccerwidow

      Reply

      1. Jo,

         My experience is much like yours. Now the world cup is upon us and the summer leagues are for the most part on a break, I think I am going to accept the losses incurred and cease with the summer campaign. I have to be honest the kick off times are pretty hard to work around and I really don’t want to continue with it. My plan is to really focus on taking my experience of the past 2 campaigns and put that knowledge to use in compiling a good 2018-19 winter leagues campaign. Waking up mega early for asian leagues, or staying up little later for MLS is no good, especially when its lose, lose lose. So I live and learn as they say.

         Reply

         1. Hi Simon,

            to get around of having to place the bets to unmanly hours the upgraded HDAFU Tables will introduce the HO/AO quotient to chose bets; this will allow to compile a portfolio by quotient (perceived strength of the teams) instead by highest odds and thus, allow to place bets the evening before the match without having to sit in front of the screen in the hour before kick off. Another option is of course, using a bot but this requires quite a bit of programming and getting your head round bots.

            Reply

      2. Hi Soccerwidow,

         Thanks for getting back to us.

         My summer league campaign was based on using all 9 tables on sale. So through my analysis I chose 1 whole season system, 8 1st half systems, and 8 2nd half systems.

         I will list the systems as you described for the current ones in operation.:

         1) Underdog 40.38%, 23.55%, 2.26
         2) Underdog 41.88% 32.26%, 1.88
         3) Underdog 40.91% 98.18%, 1
         4) Away Win: 44.73%, 15.88%, 2.6
         5) Away Win: 48.09%, 24.43%, 4
         6) Home Win: 55.40%, 17.25%, 1.9
         7) Draw: 45.95%, 45.68%, 1
         8) Favourite: 55.14%, 24.70%, 1
         9) Home Win: 46.4%, 17.59%, 4

         As you can see, average hit rate across the 9 leagues is 46.5%. The actual hit rate to date is approx 34%.

         If it helps, I can send you by email the spreadsheet I use to keep a record. It’s basically the spreadsheet you gave out for free in the summer league article a while back. That way you can see exactly what systems I chose, and all the details for each one. Let me know.

         Reply

         1. Hi Simon,

            thanks for sharing your figures. According to the data you provided your system has an average expected hit rate of 46.41% with an expected Yield of 26.71%.

            I emailed to you a spreadsheet that simulates all kind of probable outcomes. In that you’ll see that the sequence changes every time you press the F9 button, sometimes with a very good start but unfortunately also often with a painfully long losing sequence at the start. That should help you hopefully to sit through losing sequences better and not abandon a sound system too early.

            I’m in the process of writing an article about that topic in due course explaining a little bit more random sequences.

            Reply

      3. Hi Soccerwidow,

         1. Expected hit rate: 25.24%; expected yield: 50.07%; expected number of bets each week: 1.61
         2. Expected hit rate: 41.88%; expected yield: 32.26%; expected number of bets each week: 2
         3. Expected hit rate: 30.93%; expected yield: 54.02%; expected number of bets each week: 1.72
         4. Expected hit rate: 26.14%; expected yield: 77.99%; expected number of bets each week: 0.51
         5. Expected hit rate: 55.4%; expected yield: 17.25%; expected number of bets each week: 1.67
         6. Expected hit rate: 42.51%; expected yield: 34.73%; expected number of bets each week: 1.1
         7. Expected hit rate: 31.53%; expected yield: 37.17%; expected number of bets each week: 1.32
         8. Expected hit rate: 45.47%; expected yield: 15.45%; expected number of bets each week: 3.66

         Average number of bets per week adds up to 13.59, but usually it’s around 20 bets per week.

         HDAFU tables needs more accuracy and less volatility. Systems that looked perfect – even number of bets, stable hit rate etc. fail badly out of the blue. Is it possible to foresee that? Based on my current summer and former winter HDAFU experience it’s hard to say that low odds systems are less risky than high odds systems. If they win anything, the amount they win provides very little support for the whole portfolio, and if they perform poorly, they recover very slowly due to low odds range. I doubt if they are worth the risk. Currently, I have no support whatsoever form low odds systems, their combined result is a loss. My current hit rate is 26%, predicted average is 42%. I would break even with hit rate of 31%. Just 5% makes a massive difference. Can you improve HDAFU accuracy by 10%? Because I want to win something…

         Let us know how your summer systems are going?

         Reply

         1. Hi Jo,

            Thanks you for sharing your figures. According to the data you provided your system has an average expected hit rate of 39.61% with an expected Yield of 33.15%.

            If you compare your system with Simon’s figures you’ll see that you have an higher expected yield but a lower expected hit rate. A higher yield means higher risk. Unfortunately, results of bets don’t line up in a nice, easy on the nerves, order and for you that means that you are more likely than Simon to have to sit through longer losing sequences.

            What staking plan are you using? Are you betting with flat stakes or do you distinguish between your most volatile system that has an expected hit rate of 25.24% and your less volatile system that has an expected hit rate of 55.4% within your portfolio of bets?

            Reply

            1. Hi Soccerwidow,

               I use ratchet-stop loss staking plan for the whole portfolio. Currently I have only 1 system in profit and that is not a low risk system. Everything else is a loss and overall result is a really heavy loss. The problem is lower risk systems are performing badly all together like in winter and since I do this for the 3rd season I start doubting if they are really worth the trouble…