Tuesday, 26 August 2014

Trading Days - US Open - First goal complete

As you may have noticed, my post titles are pretty unimaginative. I've never been great for writing to be honest!

Anyhow, I've completed my first goal to double my bank, the bank now stands at £324.57

The 2nd goal now is to hit £500, then £1000 then £2000 eventually. After that I have no idea.
I guess my goals after that will be to learn more and more and become consistently profitable for a whole season.


I won't get ahead of myself though, once I hit each of the goals I will talk a little more about what adjustments I have needed to, or need to carry out.

I had a pretty much nightmare start yesterday, I tried to trade far too many games at once which ended in some pretty bad results to start off with (see screenshot on bencic vs wickmayer game, ouch!).

I've realised that I am able to trade 2-3 games at once for the time being, that's about my current limit and all I really want to do, once I hit higher stakes, whenever that is, I'm most likely going to cut that down to just 2 tables, as I expect I will need to do more work on the ladders with regards to placement of my bets.

So anyway, after learning that valuable lesson, I'm now sticking to the 2-3 games at once rule.
If it wasn't for the US Open and there being so many matches on at once, I don't think I would of got carried away. It was a nice warning, not to get carried away where grandslams are concerned.

I also learned (once again, sigh) I hate laying from SP as I've told myself time and time again.

This was however the first time I did it since I started again, so it's something I won't be doing (never seems to work for me).

Everything was going swimmingly until later on in the evening, where my trading started to dissolve and was getting near the boundaries of gambling then trading. At this point I shut up shop for the evening and called it quits. You'll see the 2 losses in the last 3 matches (two of these losses were bigger than they needed to be).

I did make a video, it was 7 hours long and I got an error from camtasia once I tried to encode which is annoying.

I'll be making videos this week (it's good to review what you did wrong). This time I'll just be doing 2 hour periods then saving them, hopefully quick enough to record the next sections.

Last night was a LONG hard slog, I can now get a glimpse of how hard this would be to do full time.

Anyway screenshots below.

You should be able to see the matches, where some losses are too big.




Friday, 22 August 2014

Trading Days

So, other than how my straight betting testing went and the database, I haven't really talked about my trading recently. This was partly due to the fact I was up down up down (mainly down).

I could not seem to break my bad discipline habits of old. Fortunately, taking a break and just purely concentrating on doing straight betting, it helped alleviate all the tension I had from wanting (having to) make money. That was at least my mindset before 'I have to make a green'.

Doing the straight betting helped me realise, I don't need winners, I don't need to make a green.
I need to find value and I need to do a lot of things when I've done this, I wasn't doing this before.

Let me go into a bit more in-depth. As I stated numerous times before, the actual finding of value whilst trading has never alluded me, my mentality and discipline however would eventually become my downfall.

After taking a break for a while, I have on and off (a month or so again, used a £20 bank), started looking again at my 'strategies'. Well, I don't need strategies. All I need is to find value, also after reading and listening to Sammy's blog and the podcast he did (link to his blog on the right hand side), he also says the same. He doesn't really have any set strategies, he just looks for value.

All my 'strategies' are, are area's in a match that I look for when the market has overreacted, they don't always happen of course. Before, I used to get involved even if I wasn't completely confident (in my self and own analysis) that the price was a good entry price.

"Market's gone down a lot, this looks a good entry"

What I completely failed to calculate or do, were the following:

1) What is my upside compared to my downside, what is my risk/reward ratio?
2) What is my exit point before I enter.

Now these may sound basic, fundamental issues, however on point number 2, I would sometimes enter and not be quite sure where I would exit if a trade started going my way.

On number 1, I never really had a range that I really 100% knew I wanted to get for risk/reward.

You know, I expect to lose 4ticks max if it goes wrong, if the trade is positive I expect to gain 30 ticks etc.

I've taken all of this into account whilst I've just dabbled with a small bank/s (I was not willing to risk anymore than £20).

It would seem anyway, the lightbulb moment has clicked.

After a few wobbly moments, 1 x £20 bank completely destroyed and numerous other times of going back over 'what the hell am I doing wrong?' and 'How can I fix these issues?', I am starting to make consistent trades.

I am feel I am able to consistently make money, I started with a £100 bank and I am compounding this as my profits go. My first goal is to £200 (double my bank), my second is £500.

I use 5% of my bank for trades I enter.

Before I would say 'If I reach this target' not anymore. Now for me it's 'When'.

I now know, to a close range, what my risk/reward will be before I enter a trade,

I now think beforehand, what will my exit/s be when I enter?

I now calculate see how many ticks I will win/lose, rather than purely focusing on the financial figure of old.

After all, majority of the time I am WAITING for an opportunity, I shouldn't be there just sitting gormless, I should be thinking about these things in advance.

This is something, I was not doing before, at least not time and time again.

I have made a video from my second trading session, I will be making this available (It's encoding as I write this) it's 3 hours long.

The trading sessions was from yesterday evening, if you wish to see it let me know and I will happily link you to the file. I'm sure no one will really care, until I at least hit my first goal, however it does cross reference my results, so people can actually see I've traded not just recklessly gambled!

All my sessions next week for the US Open I will also be recording and uploading, again they will be available if people want to see them, they do take a day or two for me to do however so anyone who wants to see them needs to be patient for me to encode them during that period.

I will be posting my results up on the blog for each session where possible.

I won't deny, my first going 'live' session on the 20th (I also did a video test run), was a bit shaky, I think this was down to nerves also of recording my trading and speaking, however it needed to be done so I got used to doing such a thing properly.

Below are my results below, yesterday's trading was the most successful, just because I had several trades that didn't work out after I had made some money, however I never wavered, not once.

The only bad side was that I was busy talking at one stage and missed an entry.

This video does not have sound, however my next sessions will, although I won't be really discussing where I think value is, I will be discussing other aspects about errors I have learned from and the two points I discussed above.

1) risk/reward
2) entry and exits, knowing them

At least with the sessions, people can see I am trading and not 'just getting lucky'.

Anyway, my results below (Click to enlarge).



Hopefully this also breaks up the slightly tedious database post's I've been having to write - and people having to read recently.

Sunday, 17 August 2014

Database Progress

Since my last post I have managed to have a chat and discussion with the developer.

We now have a main 'goal' in place for the start of the new year.

By the start of the new year, we want to have version 1.0 in place.

This will mean that the database will be fully operational, all the known bugs for entering, modifying and deleting trades will have been fixed. All of the front-end of the database will be working without any issues, this includes graphs and the sorting of strategies and sport categories (This was the new part of the database introduced in the latest version).

We also expect to have a logo and basic color schemes in-place for version 1.0, as well as an office name for the program. Version 1.0 should also come with an instruction manual.

Basically, everything that is in it at the moment, should work 100% without any major issues, this is the goal for 1.0

After 1.0 is out, I will get the developer to begin working on ways that we can start to analyse the data within the database.

The new 'sports and strategies' part of the database, is perhaps the buggiest of them all at the moment, it works in basic functionality, however it is also broken in parts and I didn't quite understand it to begin with.

After having a chat with the developer, I myself, now fully understand it and have asked him to fix quite a few issues with it for the next version.

Thanks for the feedback anyone has left so far, we have taken it on board and you can expect to see the improvements in for 1.0

There will be quite a lot of fixes in the next version (should not be too far away), below is what will definitely be included:


Sorting of trades in the overviews will become possible by clicking on the relevant headers in the table. That way you can more easily search and compare your trades. 

It will also become possible to hide (and unhide) specific columns for as long as you want. You might want to permanently hide them for example, because they contain data that you do not want to work with in your own trades database. But you could also hide them temporarily if you do not need them at a specific moment.



The Sports & Strategies dialog will receive some major improvements. A big improvement will be the possibility to move strategies around from one sport to the other. A bug will be fixed that caused the background graphics not to display properly.



Some more bugs that were reported will be fixed in the next version. The strategies do not show up well in the trades overview for example, this will be corrected in the next version. 
The numbering of the axis in the graphs will be improved because more testing by our users showed that there were some cases where they were not perfect yet. The full-screen graphs will also be functional in the next version.



Some other suggestions for improvements are still being discussed. They may or not make it into the next version of the program, if they do I will post about them here. But rest assured that every suggestion our testers make will be noticed and talked about. So keep those suggestions coming!


Oh, we also will hopefully be adding support to import data from Betfair profit & loss/results spreadsheets. This will take a little time to work out how to do this though.

Thanks to the horseman for suggesting this!

We should also have the ability to add data into the database, in the same format that it is currently stored.

Both of the above will take time though, we have to make sure 100% that all of the column data in the database is there, to support importing data into the database. Secondly, the sports and strategies part of the database has to be 100% working correctly before this is possible (Currently it isn't, however this is our main priority going forward).






Thursday, 7 August 2014

Update on Further Progress

In my last post I talked about the graphs and my current straight betting results.

However after talking to someone who I know, getting him to look at the results, he's made me realise that the accuracy of the model may be out.

I talked about how I need to factor in the surface more, I am not going to record, concurrently two set's of data. One with my current settings and another with another set of settings.

I should note that, after tweaking the settings, it doesn't seem to make a whole lot of difference in certain areas of probability %. However at the top end 60%+ it is making a substantial difference to how the % is calculated to each player. I have delved deeper into my results, with the help from an online friend, have realised something needed to changed.

I wouldn't say I'm really confident these changes are the exact thing that needs to be done to give me the 'perfect' model, however I do believe they will have a positive effect on winnings.

I'm hoping the new predictions will make my model a lot more accurate and give me the kind of figures I want, so that I can use this to place my own bets in the new season, rather than follow other tipsters.

Tuesday, 5 August 2014

Further Progress?

Didn't realise it has been so long since my last post, how time flies.

Hopefully people are getting to grips with the new database.

I have arrange a meeting tomorrow with the developer, so I will be going through the database with him via a screen sharing utility, explaining what needs fixing, tidying up, as well as some other things.

This will be the next stage to really fixing anything that's currently wrong with the database. I think after this stage, we will look to do further work, so expect a rather quick update just fixing a few things currently wrong. Then expect a major update, this will take a lot longer obviously, so you will have to be patient! I will keep you updated on what we intend to add for the next major release on the blog, however I am quite busy at the minute, so expect a post every 1-3 weeks.

In other news, I have finally been able to get my data out of the database thanks to the developer adding in the copy trades button.

I've been analyzing some of my tennis betting data in Excel..slowly, very slowly. Slowly because I am not great with formula's in Excel at all and have been having to trial and error.

So far majority of my data is from the Clay court season. I didn't really begin recording until the 31st March this year, So I basically missed the first part of the hard court season for 3 months - annoying.

I then missed Roland Garros pretty much, the grass court season and most of July.

I started recording again last week, so I've already missed the grass court tournaments and a few of the smaller hard court tournaments (which tend to be particularly good for upsets I think).

Now we move on with back to back masters tournament's this week, so I have a feeling my results aren't going to be great, as more of the match ups are a lot more even and it's harder to find value.


As for the analyzing, I have just been 'messing' about today, I have noticed that a probability greater than 40% with my settings seems to be the most profitable area.

Realistically, I should be able to find value across all probabilities, which is why in the future I will need to change my system, for now I just want to get the results in, analyze and see if this system is  actually accurate enough to produce profit somewhere or if it's just random luck so far.


Graphs

Just to point out, the number of trades along the bottom is 'broke' on the  graphs below. This is because the way I have so far, of calculating a running profit total, it returns a 0 figure when using the probability filters.

So for example, when set to 40%, any bets below the 40% probability filter return a 0 as Profit/Loss. The graphs are based off the running total, so they still count the zero's/same running total number.

If the running total was £2000 and the next bet was below 40%, it would produce a 0, the running total for that bet would still be £2000.

Hopefully this does not sound too confusing, basically it's something I need to work on fixing so that the running total does not repeat numbers based on the 0s. Once this is fixed, I will know the exact number of trades for each graph. At the moment only when no probability filter is on, are the trade numbers accurate. Just to clear that up in-case someone pointed it out :).

If anyone's interested in a specific graph, I can always let you know how many bet's there are in that specific one / sample size - by manually counting! Not great I know, however it's something I'm going to fix in the near future.



Below, are a few graphs I've managed to create in Excel today.

This is relatively in-depth, maybe even boring to some of you, however should be interesting to others. So keep reading if interested! If not, hit back on your browser ;)

WTA - All Surfaces - No probability filter
 


Notice the lowest point, just over -£6000, the huge up curve is from the Washington/Stanford Hard court tournaments. Series 1 = profit..if you hadn't already had guessed!

This graph so far has a steady/slight decline overall.

WTA - All Surfaces - 40% or greater probability filter


 

 A rather bumpy graph, this one shows a very slight positive climb, for all surfaces with the filter.


WTA - Clay - No probability filter

 

As you can see, so far a steady downwards decline.

WTA - Clay - 40% or greater probability filter



Improvement from above, relatively steady, however need more results. Still losing money though.

WTA - Hard - No probability filter




Very small sample size still, however positive incline so far.

WTA - Hard - 40% or greater probability filter



Smoother graph, positive incline. Small sample size than above - 66 bets.


So far, I can't presume too much from the graphs as I don't have enough data really - especially for hard courts for women, however I am fairly confident that I need to factor the surface for women on Clay a lot more that I currently am.

Now onto the Men's....

ATP - All surfaces - No filter



A positive graph with shall inclination, although it's a bit uncertain where it is going. It peaked at around £6600 and hit a low after that of around £1800. It would seem that the hard court tournament was responsible for the sharp decline currently at the end. Need more hard court data to interpret better.

ATP - All surfaces - 40% or greater probability filter



 Very similar to above, again, the recent hard court tournaments I have recorded (first hard court's recorded for men), have brought the graph downwards. Positive graph, gradual inclination however more data required.

ATP - Clay - No Probability filter



Positive graph, gradual/steady increase.


ATP - Clay - 40% or greater probability filter

 

Not quite as smooth as above, this will be due to the number of trades, however as we can see. A positive graph. Bumpy increase.

ATP - Hard - No probability filter
















Negative graph, downwards trend. Not enough results so far to get an idea of how accurate it is.


ATP - Hard - 40% probability filter



Even worse than above, however even less bets. Will need to give it a few months to see how the hard court bets perform for men (and women still for that matter).


So that's it. I have a sneaking suspicion it is going to be a role reversal with settings by the end of the year.

Here's what I 'predict'.

I need to factor in Clay surface type more for women's games.

I need to factor in Hard court surface type more for men's games.

I should also note, all this data is based on 3 set matches only. Missed the WTA French open so was not able to record results for this unfortunately.

However based on all of the above, I have high hopes for the new season that I can make something positive out of this going forward.

I guess the positive sign is, I've still made overall small profit/broke even essentially, without filters and without the need to change anything to the system throughout the season. If it was a continuous down-trending loss, it would be a disaster. As it stands it is quite positive to move ahead into the new season!