• Hi Guest Just in case you were not aware I wanted to highlight that you can now get a free 7 day trial of Horseracebase here.
    We have a lot of members who are existing users of Horseracebase so help is always available if needed, as well as dedicated section of the fourm here.
    Best Wishes
    AR

Capturing live "In-Play" data . . .

UK-21

UK-21

Yearling
As in my part of the world it's been pi$$ing with rain for the last hour or so, I've just rustled this up.

InPlayDisplay.jpg

It's some Excel magic that takes data captured from a web page and drops it into the relevant cells. Not as sophisticated as an API feed, but requiring a lot less technical ability and overheads (and it isn't reliant on a third party live feed). It's simply based on copying the data as shown below and hitting a button that runs some VBA code that copies it into a spreadsheet, which then sorts and cleanses the individual data items and drops them in the cells.

In-Play stats grab.jpg

Just keep these elements highlighted and simply pop up the "copy" option and then hit the button to update the table
at anytime.

So fairly rudimentary but it seems to function. Question is...... what to do with the stats presented; how might we make
use of this information to inform an in-play betting decision? Use it to feed into an algorithm that eventually pops up
a "recommended to do this" or "do that" message, use it to apply a points weighting of some sort to each team (possibly
to determine who's going to put the next one in the back of a net) or something else?

Let's discuss.

UK-21
 
Excellent effort UK-21 UK-21

In to excel I am fortunate or mad enough to have had all this coded in Python using an api at a considerable cost. It will give you an edge long term.
 
UK-21 UK-21 ..... a really great effort you have pulled out of the hat there UK 21. Seems to me that with info like that at your disposal, it could provide a useful source for laying the current score when stats like that are shown for one team or another. Most of the stats you have there show Heerenveen well on top and with a 1-1 scoreline which looks set to be bust in due course. Just an idea, hope it helps.:clap:
 
OK, having had a few days to think about it (amongst many other things), I think out of the 10 KPIs that we have data for we need to identify those that are relevant to the consideration in hand (ie who's going to score next?) and ignore those that are just "noise". I'm open to persuasion, and welcome others' views. At this stage though I reckon of the ten, five tell a story. These are:

Possession (x% -v- y%)
Shots on woodwork
Shots on target
Dangerous attacks
Corners

Where one of the teams has clocked up a significant lead in these it would suggest that the action is going on more in one half of the pitch than the other, and that side with the highest count is possibly more likely to put the next one in the back of the net. The other five I'm not so sure are indicators of anything in particular and are just part of the flux of the game?

Goal kicks
Penalties awarded
Shots off target
Free kicks
Throw ins

++++++++++++++++++++++++++++++++++++++++++++++++++

Question is . . . what do we do with the KPIs that we think are indicative of the better performing side? Do we just take a subjective view of the data as a whole or apply some sort of points weighting to the items (ie award a point for each corner or dangerous attack, 2 for each shot on target or on the woodwork etc) in order to create some sort of index for comparison purposes? If so, how would you award points for possession %age? Having calculated the "index count" for each team, what would be a threshold for putting some brass on the line and backing a team to score the next goal, win the match or whatever - a gap of 15 points or more for example? Do we weight these counts in some way with a consideration for the previous past form of a team; ie table position, number of wins over the last five matches etc etc.

Thoughts? 🤔
 
...... what about unforseen things jumping at you when you are at least expecting these? Such as injuries, Red cards etc which I assume are not factored in? My personal thoughts still tend to favour something like lay the current score, as what you have above seems to jump out at you and just about tell you that there is one due to go in the back of the net according to the favourable statistics. I don't think there is any perfect way but it could be the lesser of the two evils. :eek:
 
How about this:

SkyLiveDataGrab.jpg

Which can be captured as this:

SkySS.jpg


This picks up on the Yellow and Red cards, and also the "Attacks" data which was missing from the previous grab. In the above example, at the time of taking the grab, Delhi had scored another two after having someone sent off . . . which raises an interesting question around what impact a yellow card (little if anything IMHO?) or a Red card (dependent on who loses the player and who is leading at the time?) actually have? Also, what is the difference between an "Attack" and "Dangerous Attack"?

There will always be lots of considerations that aren't reported within the KPIs available, and you have to decide whether you want to take all of this into account (in which case you might as well take an overall subjective view), or whether you want to apply some standardised methodology to the KPI data to arrive at something that, longer term, will be more right than wrong?

I'm working on the latter, and am looking at going down the weighting points route at this stage. The data capture, how the numbers are presented and whether they put up flags dependent on what that are is the relatively easy bit - it's the "method" behind the points weighting that's going to be the deciding factor - IMHO. Within my spreadsheets I can make the points weighting user-definable, although I'd like to start with a default baseline.

Nice weekend.
 
You could view multiple matches at once, but it would mean having multiple windows open and would be limited to how quick you can brandish your mouse. Ideally, this is an el-cheapo solution to keeping an eye on a few matches at any one time. I'm also struggling with the fact that the data presentation is inconsistent between grabs.
 
Last edited:
..... when you read some of those comments, it makes one wonder just how reliable match in-play data really is. Is it really down to some guy getting his left correct from his right and so on! As for the football data scout, I wonder if that is what Phil Brown is doing now, as I don't think he is a manager anymore. Maybe he could not recognize his own team!:D
 
..... UK-21 UK-21 Great job there young fella and lovely and clear too. I do love the whole design and what you have tried to do with the data. Hope this works well for you, with an effort like that, you deserve something out of it.:text-goodpost:
 
OK . . . if anyone fancies trying out the BETA you can download a copy from my web site here:

http://www.uk-21.org/data/SkyBet In-Play Data Capture Tool (BETA).xlsm

Filesize is 579K. You'll need a copy of OPERA on your system to use it - this is the only browser I've tested it on, although I suspect it'll work in CHROME too as both share most of the open source code. Please RTFM (or instructions contained within the workbook) before getting stuck in. Although all of the worksheets are locked off and protected, no passwords are in place - so you're all free to do with it as you please. The same goes for the VBA code if anyone feels the urge to tidy it up or add something in.

It should be fairly intuitive to use, but if anyone has any problems with capturing the data from the SkyBet web pages post the details here. On my screen, the area highlighted doesn't show up (although the scores do highlight) although this in itself isn't an issue - just hit the copy option to capture the data displayed to the windows system clipboard.

I'll be very surprised if nobody finds any bugs in it. If you do please let me know and I'll try my best to replicate the issue and fix them.

I think that's about it for now. Enjoy!

UK-21
 
You must log in or register to reply here.
Back
Top