Creation of a "systems calculator" tool for Punters Lounge

[Datapunter]

My fellow system punters, there's an idea i've had for a long time and i think the time has come to make it happen.

What ?
I'm talking about a tool for Punters Lounge, a "Systems Calculator", let's call it "Sniffy" for now.
The most important function "Sniffy" will have is to look at a systems thread and automatically calculate the result of that posted system.

But it will go beyond that. Not only will it show the result like strike rate, yield, profit/loss , it will do so using multiple staking variations such as level, percentile, fixed profits, etc... A graphical display can be included. The poster may even be able to get a copy in Excel format. And there's more...

There's a hint in the name "Sniffy": it will periodically do its rounds on Punters Lounge , picking up systems in whatever forum they are posted and keeping all results up to date. This will allow us, as a community, to display overview pages about systems. A lot of work if done manually, but with "Sniffy" keeping the info up to date it becomes a piece of cake.

How will this all work ?
From a posters point of view things will be pretty much the same as they are now. You simply run a thread on a system, adding bets as you go along, as well as comments, reasoning, suggestions, feedback, smilies, images, no change there. The only new thing would be that bets need to be posted in a way that "Sniffy" can recognise them as a bet. We'll need to give that one some thought but shouldn't be too hard.


Ok, nice idea, now what ?
I'm looking to form a task-force of about 4 or 5 people.
This task-force will stick their heads together and come up with a detailled description of what we want the tool to do, and to some extent how we want it to work. No technical expertise required ! The task-force is about functionality. When we are clear about what we want we will look for a developer to create the tool. The task-force will then also take care of some testing and maybe some documentation.

Obviously everyone is welcome to add their thought, ideas, suggestions in this thread. The task-force is just about practicality.

[ZuluWarrior]

I'm in, DP and I have already chatted and I have some ideas on how to incorperate Sniffy into the PL and other forums...

I'm a programmer so I expect thats where I'll get involved.

Are you thinking of hosting the overviews an advanced analytics on another site DP?

If you are, then tips will be going into a database, so we could have more analytics than you could shake a bet slip at!

also, that would facilitate Email subscriptions, RSS feeds, ect.

That would be cool, your system as an RSS feed automatically...

[ZuluWarrior]

Another thought, we don't want to go too far as otherwise we'd be creating a glorified free proofing tool. What options would we offer for privacy, sports, ect....

Then there's multiple bets, ect....

[Datapunter]

Are you thinking of hosting the overviews an advanced analytics on another site DP?

Nah, Punters Lounge itself is a website, we can generate as much webpages as we want. Got a MySQl server running the forum, don't see why we can't have a database for the tool. Should be possible to do the whole thing in-house.

If you mean use the tool for advertising Punters Lounge on the net, sure why not, we can look at that.

What options would we offer for privacy, sports, ect....
Then there's multiple bets, ect....

That's what the task-force will look into. Pretty much open at the moment.

[EulogioJ]

Hi,
I think it is a great idea . About two months ago I was looking for a similar tool for testing an idea I had. Let me explain the idea and think about it:

I would like to look at several soccer leagues: Premier, Spanish... and take the last 5 years results. Perhaps if we examine the number of goals in each match (over or under) we could find a profitable system. For example, betting on all the football matches of the Premier for over result. In the teams that it is over, continue with progressive staking until, for example, 5 matches. If we succeed in "x" progressive stakings during the season, perhaps it is possible to obtain 10-15 % profit using the system. Of course it would be profitable only using a lot of money, but for me it would be ok. The problem is that looking for such a system (over, under, even goals...) would require to take the number of goals of all the matches in several leagues and in the last 5 seasons and run a program in order to look the required parameters. All of you know that every year there is a team with at least 5-7 consecutive matches over. Perhaps it is possible to create a program to check several strategies taking into account only statistics...

I hope I explained myself, it is easier to think about it than to explain the idea

I am sorry I can not help with the tool, I have no idea about programming

[mcgin]

I'd be happy to help out DP

How will this all work ?
From a posters point of view things will be pretty much the same as they are now. You simply run a thread on a system, adding bets as you go along, as well as comments, reasoning, suggestions, feedback, smilies, images, no change there. The only new thing would be that bets need to be posted in a way that "Sniffy" can recognise them as a bet. We'll need to give that one some thought but shouldn't be too hard.

I see 2 approaches to this one (Which effectively equate to the same thing but the UI is different). Basically boils down to adding a new tag that's recognised by the board similar to the QUOTE/IMG tag etc.

The 2 approaches:
1) Stick a button next to the Post Reply button, say Post Selection which comes up with an interface for entering a tip which auto formats the post as a valid selection recognised by the tool. The tool then only needs to look at each post rather then having to search within a post. And you'll only have one selection per post
2) Stick a button on the regular reply page to enter a selection which pops up with a window similar to the image or link buttons allowing you to enter the details which are auto-formatted and placed within a post. This allows multiple tips per post.


Of course thats the easy bit, need to decide how to format a selection then! Really depends on how much info is needed to run a system.

It'd be pretty decent if you could just set the parameters of the system in an initial post (Initial bank, staking strategy etc.) and then only need to post the selections in the body without having to worry about stakes etc.

[Datapunter]

EulogioJ,
What you describe is certainly possible but i think it's a bit too much for this project We're looking at a more general type of tool.

Zulu and mcgin,
Get a room you 2.



Everyone else,
Guy's, I really intend this to be something everyone can use.

So i'd appreciate some feedback on the idea from some of the less technical members. You like it, don't like it, can't wait, don't get it, need to see it first, you do or don't see you using it, whatever pops in your head.

[PAULM03]

If we were to go down the route of auto-formatting the posts why stop at systems? We could actually follow every single bet on the PL with this kind of formatting.

You could automate all the competitions at a stroke, have daily, weekly and monthly leaderboards of tipsters. You could show profit or loss for the top 10 punterslounge tipsters, break it down by sport, bet type, event type or anything else you wanted to look at.

[Matthew]

Originally Posted by Datapunter

Everyone else,
Guy's, I really intend this to be something everyone can use.

So i'd appreciate some feedback on the idea from some of the less technical members. You like it, don't like it, can't wait, don't get it, need to see it first, you do or don't see you using it, whatever pops in your head.

I probably fall into this group.

I think it is a great idea, I would defo use it for my tips on my EPL thread.

So........when will it be ready?

[Matthew]

Has this gone cold?

Or is there conversations going on in the background?

I know Rome wasn't built in a day..........I just think it is a good idea.

[Datapunter]

Not dead at all, i guess i just need to get the ball rolling.

But i do have to say i'm a little disappointed by the number of replies,
could just be a coincidence but basically the number of non-mods and non-technical replies is..... 1

Should be fine once it's up and rolling and people can actually see something, but until then it's just a low priority. ( for me anyhow )

[Wumpking]

As a non-mod and non-techie I thought I should respond. This something I would certainly like to see Datapunter, as sounds as if it would provide a fantastic opportunity to easily benchmark totally different systems and aproaches as well as see what's hot (or not). Overall, sounds a great idea .

[Sgt.Sunshine]

Only just noticed this thread - it's a great idea. I particularly like PaulMO3's idea, that if you want to be included just use a set format for the bets in your post regardless of which area of PL it appears.

[muppet77]

i think it's a great idea fellas. how difficult would it be to set up initially?
get it going!

[slapshot]

I think this is really good idea

[Rushian]

I think it's a really good idea on a number of levels:

1. Comparison between the systems would be much easier in terms of key indicators (e.g. yield, strike rate etc) which could be produced both numerically and/or graphically for ease of use.

2. It would help to highlight the systems that have positive/negative results due to luck/chance from those that are indeed producing profits/losses consistently (based on some appropriate statistical criterion that I am sure people must be using). This could be calculated directly for any system submitted through "Sniffy".

3. It would also be helpful for the creators of systems to explore the performance of their systems and seek improvements. This is especially true if "Sniffy" could be dynamic in the sense that one could query the performance of each system based on some pre-determined parameters (e.g. how did a system perform for away picks with odds under 2.2 etc) This in effect would help highlight problematic/value bets of each system, and together with point 2, would suggest whether these bets were down to luck (variance) or consistently problematic/valuable.

Just my thoughts.

[phantomDude]

It's a good idea. An example of a similar concept is www.blogabet.com. A feature I like about blogabet is that it requires users to enter odds for their picks from some website (oddsoddsodds.com or something similar). Users can't enter inflated odds for their picks and sneakily exaggerate the returns from their systems. It may be worth pointing out that a clunky, inconvenient mechanism for entering bets in a strategy imposed upon posters may actually deter people from posting their systems.

As for making Sniffy a "dynamic" tool that allows picks to be filtered by some predicate (eg how did a system perform for away picks with odds under 2.2), that's probably going to increase the level of complexity of the tool by quite a bit. The more flexibility such a tool offers in terms of the queries that users can pose, the more difficult it's going to be to implement. Not impossible (or anything near it), but storing data entered by a user, and displaying that static content on a summary web-page is a lot simpler than storing the data, and creating a "dynamic" web-page that can show the returns on systems when only bets with odds less than 2.2/2/1.8/etc. are shown. If you want this to happen soon, it's probably best to put the "dynamic" tools on the long finger, and try get the static version working first.

You could even argue that making sniffy a "dynamic" tool mightn't even be a particularly good idea. Betting systems need to be evaluated over a large number of bets in order to gauge their success. Filtering the bets posted by a poster as part of their system might return a very small set of bets. What if only 5/6 bets are returned by such a filter? If a system is filtered so that only bets where the odds are less than 1.5 are counted, and 6 bets are returned, 5 of which are winners, could it be argued that this is proof that the system is profitable for odds under 1.5? A lot of people would argue no. Since people aren't going to be posting hundreds and hundreds of picks (most wouldn't even be posting dozens and dozens) the results returned by using such filters would usually be statistically insignificant.

[EDIT]

[marron]

Great idea

Good Luck with sorting it

[muppet77]

great idea.

how would we start it with systems already running?
would there be a 'quick start' method to enable the 'sweeper' to pick up the current stats, rather than typing in all of the past bets in the correct format?

[Rushian]

Originally Posted by phantomDude

You could even argue that making sniffy a "dynamic" tool mightn't even be a particularly good idea. Betting systems need to be evaluated over a large number of bets in order to gauge their success. Filtering the bets posted by a poster as part of their system might return a very small set of bets. What if only 5/6 bets are returned by such a filter? If a system is filtered so that only bets where the odds are less than 1.5 are counted, and 6 bets are returned, 5 of which are winners, could it be argued that this is proof that the system is profitable for odds under 1.5? A lot of people would argue no. Since people aren't going to be posting hundreds and hundreds of picks (most wouldn't even be posting dozens and dozens) the results returned by using such filters would usually be statistically insignificant.

You are right phantomDude regarding the results of filtering. However, if combined with a statistical significance measure of those filtered picks (as I mentioned in my second point of my previous post), then the filtered "system" will be evaluated in terms of its (statistical) profitability.

I accept, though, your point in it being too complicated/difficult to implement. I am definitely not a technical expert, so I would leave it to those of you who know best!!