- Total Posts 3386
VDW Conclussions – Part Two.<br>As I said earlier we had formed a syndicate to use the DGA Pro-Punter program. The syndicate was formed because of the lengthy process of manually loading the data required to perform the program operations. It would take an average of two to four hours to input one race. So the basic criteria for being a member was owning an Amstrad computer. Each person would be alloted a couple of races to input and this would give us at least a couple of good bets for the Saturday afternoon.
The nature of the program familiarised me with form reading. We retained the records and within three or four months had built, what we thought was a good source of data. It was against this background that we researched the VDW methodology, amongst other systems, over about a year.
I have made this quite a lengthy post so you can decide for yourself whether the decisions made were sound or not. I’m sure that I have only remembered the salient points. In addition, I have added my own further research for your perusal, as I feel a shorter response may be misunderstood.
Informations presented.<br>System Aim = 80% strike rate<br>Statistical Filters = Betting Forecast / Form Evaluation / Class Evaluation.<br>Suitability Filters = Race Type / Number of Runners.<br>Identification Markers = Days since last run.<br>Betting = Illogical Staking Plan.
Attempt One……<br>Working the system with the information presented.<br>Betting Forecast – Eliminate outsiders, captures 75 – 80% of the market. Eliminates a higher percentage of the field in Non-HCP’s than it does in HCP’s and higher class Non-HCP’s. We decided that this was how we should determine the number of all top selections from the check list. For example, if the betting forecast selects four horses and deletes four horses, only the top four for Form and Class would be considered.<br>Form Evaluation – in it’s own right, throws up winners. Bad in Non-HCP’s better in smaller field HCP’s.<br>Class Evaluation – Bad in lower class races. In HCP’s can push the selections up the weights. Major draw back, does identify horses coming up in class.
So, taking these filters and applying them in the proscribed VDW cross refferal method we arrived at these trends
Betting Forecast x Form Eval. x Class Eval. = Favourites<br>Betting Forecast x Form Eval. = Favourites and second Favourites.<br>Betting Forecast x Class Eval. = limited success, lack of data.<br>Form Eval. x Class Eval. = Short Priced Favourites.
Race Type and Number of Runners were ignored in the first pass. Doing this would allow an assessment of advantage when applied retrospectively.
We quickly realised that Class data was a bad filter. Not only was the lack of data not providing meaningful ratings any filter applied to it eliminated the longer priced horses it did throw up.<br>During this phase we were getting a strike rate of around 20 – 30% and no viable selection method. Days since last run made little or no difference between the longest and shortest number of days. Drawing a line at this we had a change of heart.
Attempt Two………..Modification of Data.<br>We decided to discard the Class Eval. Filter and days since last run, rightly or wrongly, and introduce two new filters.
We applied a Primary Filter – Number of Runners – limiting the size of field being considered to 6 to 15 runners.
Form Rating – not wanting to abandon class altogether and not being overly confident in our VDW form figures, we decided to bring newspaper form ratings into the equation. We figured that so long as they were form ratings and not speed ratings Ãƒâ€šÃ‚Â we’d be okay. A minimum of 50% of the field should have a rating, to ensure they had any relevance.
So we set up another set of filters and repeated to process over again.
We were still not there, we were down to two horses being the optimum or should I say more likely route, but couldn’t get over 60% strike rate and then only in cetain types of races and with limited conditions.
Final System……..<br>Delete any race with less than 6 runners and more than 15.<br>Bet only in proper Non-HCP races.<br>Delete any race with a joint first or joint second favourite in the F-SP.<br>From the remaining races, select the forecast first and second favourite.<br>One or both selections must be in the select list for Form Evaluation.<br>One or both selections must be in the select list for Form Rating.<br>Combined SR = 60 -70%.
The syndicate research ended at this point and we agreed that this was the best we were going to get based on the data gathered. The system was abandoned because it only generated two or three bets a week. I believe that this is the VDW selection system. This invaluable research allowed me to go on further and reinforce my belief that this was it.
To be continued…………………<br>