Definition of a 'system'
 

Reading another thread got me to wondering what the actual definition for a system would actually be. Whilst the dictionary offers these interpretations:

1. A group of related things or parts that function together as a whole.
2. An ordered set of doctrines, ideas, or principles.
3. A particular method of procedure, organization, or classification.

For our purposes a racing 'system' seems to be something different again. I know what I think a racing system should be and am wondering what you think?

For starters, here is an off the top of my head definition.

My racing system should be any collection of data that once analysed gives an indication of a horse which represents a better chance of winning a race than the other horses in the same race.

That's a bit wordy but I am sure I could rephrase it if I thought about it, but the sense is clear, it should be able to assist you to select a winning chance. However the purpose of selecting a winner is not always the only consideration. I know you need to make a profit, and sometimes this does not involve selecting the winner per se, but perhaps just a minor placing or an exotic selection etc.

What about gut-feeling, the most powerful weapon yet devised by man for picking winners? Anyone have that little voice inside that screams at you DON'T back that one, PEN this one, NOT with your money, and so on. How do you program that into your system?

Systems are different things to different folks, so what are the commonalities? When speaking in specific computer terms, what are the prime factors, secondary factors and so on? Is a system better if it involves more factors, or does the KISS principle work here as well?

As some of you know I am formulating a system, I have some computer skills and lack in other areas, but don't lack experience on and off the tracks. So I am asking for a set of criteria from each of you, what makes your system tick? What is important to you. Not only will I benefit from your answers in compiling my own system, but the sharing of ideas may force some of you to rethink your own systems and modify them and perhaps improve them in terms of profitability.

I don't think the telling of what defines your system will harm you, the game is to big and there is no such thing as the perfect system, but that should not stop all of us from striving for it anyway :-)

Tailwag

I just wrote a reply to this and realised that what I came up with was a complete contradiction to my opening paragraph.

A system depends on what criteria you look at. If you look at the form, you should get the same selection every time. There will be no "gut feeling". However, I look at the prices, so sometimes there is a need to pick the "most likely" selection.

Many would argue that definition should finish as ....a horse who's odds are better than that horses chances of winning the race.

Gut feelings can work two ways.
One school of thought would be that the mighty computer that is our brain is seeing things our conscious mind fails to see. This gives the nagging sensation I think YOU are referring to tailwag and that is a good thing.
Alternately the data that your brain may be using to give you a good nagging can be false. It remembers isolated incidents and makes too much of them. For example it remembers that time you took a punt on the outside barrier (no sleight intended to outside barrier proponents) and got a big win, then it falsely thinks anything not from the outside barrier is suspect.
It would be interesting for an individual to make a note of any "Gut Feeling" bets they make and see how successful they really are, you may end up hating your own guts.

KV

I've done that many times lately myself. The reason I mentioned gut feeling was more as a handbrake against making a system so all encompassing that it becomes just too complex. I guess I shouldn't have really mentioned it because gut feeling really has no part in a computerised system for obvious reasons.

What I am trying to do is get some discussion going about what individual factors people think are important in a system, things like the obvious weight, latest form and so on, and things perhaps not so obvious like the ratio of winners to losers at certain odds to tote turnover etc.

Tailwag

Tailwag, if you look for obvious things "like the obvious weight, latest form and so on", you are wasting your time. It is all there for you free, reflected by the odds, tipsters polls etc.
To find that elusive angle, you have to look outside the square.
I worked long and hard on my original rating only to find, that I was getting the same horses almost in the same order as everyone else, so I stopped using it.
Now for the last 4 or so years I've been using a different method altogether AND making profit. All I am prepared to disclose is, look at the reason, why some horses are shunned by punters?
Once you selected those, you can still put them through a proper rating to sort them in ranking order.
Good luck!

Tailwag, as well as considering which factors to use you should look closely at how these factors are applied.

Most mechanical systems use benchmark rules like 'horse must have a 40% win strike rate'. What sounds like a good logical rule falls down in practice in that it doesn't properly evaluate a horse's strike rate relative to other horses in this race.
e.g. in race A one horse might have a 40% strike rate and next best might be 10% yet in race B the same 40% horse might be opposed by five other runners with strike rates of 39%.

Of course in this instance you also have the problem that strike rate is not necessarily a good measure of current form or ability.

wunfluova

Oh my God, the wheel is being re-invented here.

OK lets define [re-define?] other commonly well known meanings in racing now that 'system' no longer means system as most of us know it: A method of punting following a strict set of clear rules to determine selection[s].

!. Handicap race ? A race for disabled horses.

2. Dead-heat ? The temp. at a Dead track.

3. Welter ? Race for horses in a certain weight bracket.

Etc. Etc. Etc.

Thank you.


Thank you.


I agree with you, I actually said the same thing.

:)

My racing system should be any collection of data that once analysed gives an indication of a horse which represents a better chance of winning a race than the other horses in the same race.

That is one way of doing it but not the only way, nor does it necessarily acheive your true objective...I guess it depends upon your wish to produce a profit from (a) backing the selctions, or (b) selling the selection method..

For instance, the objective of a racing system, (for many people) is to make an easy and consistent profit over time. This does not have to mean selecting the horse that has a greater chance of beating its rivals in any given race... For example, it may mean selecting the horse (or horses) that, over time, win often enough, at a sufficient price, to produce a consistent long term profit..

Yep, right on. Doing the form and selecting the winner.

Well I never thought about selling a method, isn't that like writing a book on how to make a million, hoping to sell enough copies to actually make the million? No, I am far more practical than that, I want to build a system, that will give me true financial independence, whilst at the same time letting me do something I have loved all my life and 'PAID' to participate in.

I accept your definition of making a profit over time, but raise the question of how you can possibly put a price on the years it took you to amass the knowledge to even come close to making a 'real' system that makes a profit over time. What price do you factor in their and have you deducted that from your bottom line, and if so, perhaps you are no longer in profit. If you made a profit, was it a fluke, a convergence of positive elements, or just luck? If it worked in the first 5 years will it fall over in the next five years?

I doubt that many people ever have a static (Finished) system that yields long term profits. I suspect that a system would need constant tweaking and constant appraisal and modifying, to mimic the changing nature of the industry it is based on.

As the rules and trends change, so must the system makers, to me this is simple logic. I will never say my system is built, or finished, because I know that would be the start of the end. I would opt for a dynamic system that had the ability to change as conditions demanded it.

There are some factors that you just can't quantify, some factors that can't be anticipated nor converted to a numeral or string in a program. For example, take the effect of betting exchanges, television wars, government tax changes and so on. These things have a profound and long term effect on the industry, thus the industry actually changes intrinsically, therefore the system must also change by some degree to keep pace with the ever changing landscape of what we call the racing industry. Not to mention the menace of drugs in sport and people simply not trying to win. These are the great unknown intangibles, some you can try to adjust for and some you just have to accept will never go away.

I accept everyone's voice, for there is truly no right or wrong, its not about the math whatsoever, its not about the ability of the participants, its not even about the money, it is about the entire game, lock stock and barrel, or as they say in the classics, "the whole shooting match', or is that 'Box and Dice'?

My point is, a system is never finished...not really.

Tailwag

Hi Tailwag
 

"I accept your definition of making a profit over time, but raise the question of how you can possibly put a price on the years it took you to amass the knowledge to even come close to making a 'real' system that makes a profit over time."

in that learning period, (which never ends) a lot of time and some money was thrown at ideas that didnt work out...nothing ventured, nothing gained. Months and months spent dry running methods..no computers back then...

You are correct..I cant put a price on that..nor my time every Saturday and travel to a metropolitan course every race day, and public holiday, for many years.... When I moved overseas, my methods were instantly unworkable to usable for me, and the biggest thrill for me was being able to claim some of my life back..namely all the Saturdays! What a relief to not have to work Saturdays nor arrange my life and family around public holidays..

I used other people's knowledge rather than just my own. (Anyone who doesnt do their own form but uses other peoples ratings is in same catagory.). I still maintain that one does not have to be a form expert to win at racing. Luckily for me, there are some great form analysists out there, so I dont have to get too involved in that arena. The methods I use today are nothing special..use someone elses ratings and filter them, selecting horses which appear to represent greater than average value for money..

One does have to have nerves of steel and discipline though, and it is these which are usually the undoing..

You are under no illusion that your system will change over time..
While my original plan appears to still work twenty five years on, changing circumstances has forced modification.

However, my tastes have changed also.. I'm seeking less action these days, so I intend to introduce further restrictions to arrive at less bets. I introduced Brisbane into my betting a few months ago, to spread my risk, but now I realise that I dont really want the extra workload.

To me, the primary objectives in developing a new system should include

1. Objectivity... black and white criteria..not open to interpretaion.

2.Simplicity....cant stress that one enough.. a system is only as good as the ability of operator to ensure accuracy in selecting. The simpler it is the less chance of errors. and the more enjoyable it can be. With so many race meetings today...this perhaps more important today than ever before.

3. Choosing materials and resources that appear to be here for the long haul, to minimise the incidents of change..

W924
 

Couldn,t agree more on keeping it simple.

In my experience complicated methods invariably lead to short priced favorites.

Cheers.
darky.

Hi Darkydog2002 and Tailwag
 

"In my experience complicated methods invariably lead to short priced favorites."

I imagine that may well be true.. Picking winners aint nearly as difficult as correctly backing horses which will produce a long term profit.

Tailwag, good luck with your efforts ...a fresh look at racing and a periodical re-evaluation of every "Given" or maxim in racing may give you an edge..

Thank you very much. Good luck to you w924

Tailwag

I'm all for tweaking systems, that's what makes them work better (hopefully), but if you can get a set of iron clad rules to work by (at least until the next tweak) that's wonderful. I find nothing more annoying than having a system which forces you to make decisions on a race by race basis. Shall I back this one... I'm not to sure about that one... etc.
If you trust your system implicitly it probably means it's doing it's job and being successful. If you're dithering over every bet it recommends the system probably isn't ready yet and you need to work on it some more before you use it.

KV

I think that generally we (well me anyway) were referring to the beginning stages of a system, if as you say you have one that works and works well, I agree, you would be better off leaving it alone.

Tailwag

Tweaking the system
 

I am currently trialing a system that follows the results of different types of selections. The system basically determines what type of selection to use next according to race results as they come to hand. I don't want to provide too much information about the program only to say that it is dynamic in that the system itself is changing constantly throughout the day.

The program works on all types of races (horses, harness and greys) and is automated because I could not possibly keep up with the constant collection of results and subsequent number crunching.

I have been developing the program over the last few months and whilst I have not been collecting results during that period I could see it was looking very promising. Recent results gathered over the last 18 days shows it won on 16 out of the 18 days. But not to get too cocky - time will tell how it fares in the long run.

What is unique about the program and pertinent to this topic is that it is a system that is not constant but is changing throughout the day.

Very interesting, you have happening something like what I have in mind. I believe that essentially there are only a few key areas of difference in any system. Form (meaning events that have happened in the past), betting (meaning what people think in the now) and of course predictions (the selections that are going to happen in the future). This is all bound by the wagering methods.

I doubt that backing a tipsters 2nd selections is really a system based on my analogy above, however it might yield good results depending on the wagering method.

I also believe that a system may indeed have a high incidents of selecting winners and that you can still lose because of unsound wagering principles.

Be that as it may, a dynamic system appeals to me because it has the sense of amalgamating all the factors above (past, present & future) into one 'living' system that by its dynamic nature could be the source of some highly remarkable selections.

I wouldn't want my system to come up with every favourite (as was mentioned earlier by a poster), if that was all the system was yielding me, I might as well dump it and buy a race book with tips in it.

The dynamic system appeals to me because of the sense of taking the latest possible information and crunching it at speeds that humans can't. In this sense I definitely feel that I would have an advantage. I am not frowning on experience and gut feeling as recorded earlier but a computer takes the sentiment out and relies on pure data. There is an old saying, garbage in garbage out, an old programmers saying. It has the reverse sense also, meaning that the quality of your input will directly equate to your output, quality data in, quality selections out.

You still need to figure out how to back them, that's another part of the total program, but I sure wish I could hear more about your idea of a dynamic system, without revealing too much of course :-)

Tailwag

That sound like an interesting concept Debug. I'm not quite sure what you mean though. Could you give us an example of a 'dynamic' system and how that would work in practise, so I can further get my head around it ? Thanks.

Hi KV
 

I find nothing more annoying than having a system which forces you to make decisions on a race by race basis

Yep Im with you on that one..Im not geared up for that type of betting...mental wise..I like to know well in advance as to which selections I have for the day. The chances of making errors can increase dramatically, while, the clock ticks down to the jump...

It gets back to finding something that suits one's own makeup. The above type is well suited to some people, although I would suggest that it would be a very small perecntage of punters...having the calmness and ability to handle "realtime" pressure.

tailwag/crash
 

Sorry about the delay in replying.

I don’t want to give too much away about the core workings but here is some information about the program.

What I meant about dynamic is that the program does not roll along (and it may run for hours) selecting runners on a fixed method of selection - dutching the first 3 favourites for instance. It will check the results of a range of previous selection methods to see how they have fared and switch to the one that is producing the best results. How it measures results is another issue but it will also take into consideration tracks and codes (racing, harness and greys). Sometimes it will not make a selection if the results are not favourable. During the course of a program run it will be constantly assessing results and switching selection methods.

The program when it starts has to run through several races to collect information before it kicks in and begins it selection process.

There is a pop up window control panel that has range of switches (tick in the box type switches) for switching features in and out of the program. I set these switches before the program starts or I may decide to change them during the running process. Specific race tracks, races or race codes can be preset if required.

Bets can be set as win, place or each way. With an each way bet there is the option of 1 unit win to 1 unit place, or, 1 unit win to 2 units place - which is what I prefer.

Preset race selections can also be used - Tab 3 in Randwick race 5 for instance. The program will use this selection for this race instead of doing its auto selection.

It also has the provision for putting information up onto the net. Because the races for the day need to be in chronological order I often put this info onto the net (with the click of a button) for anyone who wants to use it.
See http://www.videocam.net.au/koala/racetimes.html

The program as I see it will always be evolving. Changes being made new features added. What I like about the program is it does not dither, unlike me who could not pick a winner in a 2 horse race.

Looks like your pushing a personal barrow here Debug. Not that I mind, but management might :-)

Crash
 

No, there is nothing to push.

It is a personal program for my own use.

I put stuff up on the net simply simply as an excercise which I pass on to a few people who I swap discussions with.

Hi Debug, thnakss for posting that link..it could be very handy for a lot of punters, regardless of which methods they use ..

Thank you for your reply, and I could not agree more with the final sentence of your posting quoted above. These are words I have spoken often!

I also took the liberty of downloading your programs and looked at them, you have done a lot of work and are very serious about what you do, so from that point of view I wish you every success. You are a doer and not a talker :-)

In as far as most systems are similar in fashion but change in design, they are all dependant of several key areas, I can see the dynamic nature of your system, and it represents a real-time/race day solution. I want to develop my system to have a dynamic capability NOT on race day but the day before or several days before.

This of course would mean that current prices would not be part of that criteria, however predicted prices certainly could play a role in the selection process and then if predicted prices are matched with real prices on the day, you could back the selection then.

My ideal system would involve collecting the fields for a meeting several days before the event, running the program to come up with selections, give me a print out of same to pop into my pocket and take with me to the races or any other place where I have access to placing a wager.

Then armed with my list of selection and probable price ranges, I sit and wait to see if the odds expected match the odds on offer and I place the bet, the rest of course is up to fate and history :-)

So my form of dynamic in this sense is the interplay between the selection criteria and the prices on the actual day. I would of course only have to concentrate on the races where selections have come up according to the selection criteria. Other systems would have you scanning every race for a possible match and that is a full time job.

My second form of dynamic is totally different than described above. In this form of dynamic system, I would have all the above as well as a direct current odds feed. The program would flag me when a certain price point (odds) matched a desired predetermined set of selections according to my selection criteria which I can run on the day the fields come out.

The essential difference between the two systems are in the first, you get a print out and can go and have a life and place your bets over the phone or in a Tab outlet etc. The system would rely on you doing a personal observation of the prices on offer for the selection in question.

The second system would require you to be at your computer with a direct constant feed of prices waiting for the all important matching of odds to required price to make your bet.

Both systems do the same thing, they differ only in releasing you from your computer dungeon, if you want to get away occasionally and live like a real person.

The bottom line is, regardless of how much you automate your system, how much you rely on dynamic, real-time information, you simply can't avoid making a human decision when it comes to betting and nor should you want to. At best you must oversee your creation and be totally aware of what is happening at all times, especially if you have a healthy respect for your money.

At worst, you will need to spend countless hours, copying and pasting and doing manual tasks to format your data and make sense of it and eventually give you a selection and then leave you with a decision to bet or not. This is the hard way and I think would be a very good first step for any programmer/punter before moving up to the next level of semi or full automation.

These are just random thoughts on the subject of dynamic. Dynamic could be measured by the amount of time/work you spend doing tasks to actually make a list of selections. In this sense, whether or not you make a profit hardly seems to be the focal point. It becomes a question of how much time and effort you invest in your system versus any profits or otherwise :-)

Tailwag

Just a word about the race times for anyone using the them. I try and get them up on the net the evening prior to the race day and occasionally there can be an alteration to the starting times at one of the tracks the next day for some reason or other. Just be aware to check the list if relying on it. Also, I may not always be available to upload the times so there could be some missing days.

tailwag,
Good luck with your project, it looks promising. It all takes time as you are probably aware. One of the traps that is easy to fall into - and I am guilty of this - is to lose site of the objective to make a quid out of the TAB and get sidetracked with bells and whistles for the program.

Thank you for that advice and I can relate to it, I am prone to spending months on how a gui looks and as you say, drift off the core issue.

Regards
Tailwag