Testing spot plays in access

[Carl]

Can't get it together to put a stamp/envelope/check parlay pick three together for my share of the board. Humm.....made me feel guilty, so I don't feel like I "should post here" till I pay my share.

Well, I overcame that anyway. (BTW Rick, I do have a credit card on file with Ken. If you could just take my share out of that it would be EXCELLENT! Would save me a lot of "thinking about". Let me know.)

On to access.

I am presently using 16 spot plays. Someone posted a month or so back that they wanted to do a double bet when a horse was picked by two spot plays, a triple play when picked by three, etc.

I liked this idea and wanted to check it out. Turned out to be an interesting access project for me. To find out if this approach was "better" for me, I went through the following manuvers:
Merged my data into one BIG data base. Impact, HTR, XF1. Ended up with 390,411 race lines for 47,872 races.
Pulled each of the race lines out on a query for each of my spot plays. Coded them with a column of their own, then replaced them in the data base. Each of the plays could now be run with a single number rather than a multiplicity of factors. I could easily test how plays did together.
This large data base runs surprisingly fast on my little pentium two 233 HRZ antique. Don't know why. Thought it would be much slower.
What was interesting to me was that

I had 5729 spot plays bets in total. They showed a 14% win percentage/45% ROI in the large database.
813 of these plays were picked out by two or more spot plays. These showed a 16% win percentage/101% ROI.
220 plays/horses were picked out by three or more spot plays. These showed a 20% win percentage/159% ROI.
76 plays were picked out by four or more spot plays. These showed a 26% win percentage with a 202% ROI.
Now IF ONLY these darn little spot plays could be trained to perform as well in the future as they did in the past, retirement here I come. Opps, forgot, I may be retired already and I have been doing a little better horseplaying of late, back up to $7 bets from the $3 bets I mentioned last post. Few winners will do that for you.....

later,
Carl
4)

[fred4now]

Carl,
I started out doing that, if horse came up 3 times betting triple etc. I made an table and appended to it so each horse would show as many times as it got picked for that day. My theory behind this was that each play should be able to stand on its own and each play should be bet regardless of the others. I still think it is a valid idea but the ups and downs come very quickly, think it would be better if using a bigger bankroll than I was at the time.

BTW, I really didn't think a new board was necessary, but am starting to like this one. Would like it to remeber me so I don't have to sign in each time though(I already have cookies enabled)

[Carl]

Thanks Fred.

I got into trouble with my ESB account betting triple, etc also. It just accentuated the fact that my picks sucked, got me to where I was going faster. Feeling better now that I have integrated Impact factors into my spot plays, but......we'll see.

Easiest solution to the multiple bet thing as far as I can see is to use less percentage of bankroll on each unit. And, since I have now "proved" (to myself at least, the only one as far as this betting stuff goes that matters to me) that multiple plays should be bet multiple , I'll have to do that.

BTW, the browser here remembers me fine, even calls me "Carl" every time I have hit the site the last two days. I am using Internet explorer, I think I used the default Rick had set up, which I think was "disable cookies", but I am not sure. So don't despair, it will remember you soon (I think).

later,
Carl

[Glen]

Always been curious how plays that come up multiple times by different queries do. I had taken the stance that if 2 or more spot plays picked them then the public was more likely to find them as well. I have since moved away from that stance but still have no answers as to the correct answer.

[Carl]

Glen,

Looks to me like the public does not look for the same things I look for in a spot play, i.e. some good factors to have a shot to win it, some bad factors to drive the price up. Few thoughts.

I noticed my average pay did not correlate with the number of times the horse was picked. Using the above example that started this post, the average pay for all winners was $20.06.
1) If a horse was picked by two spot plays or more the average pay was $24.31
2) If a horse was picked by three spot plays or more the average pay was $26.57
3) If a horse was picked by four or more spot plays the average pay was $22.96

These numbers are probably a little crazy because
1) There are some low pay/high percentage plays in the full sample that do not get doubled up on.
2) Obviously some backfitting here, in that I have been using past data to optimize my current roster of spot plays. And it could be that some of them are fatting on the same longshots.

Still, I feel safe in saying that what we see as "good" with HTR is not necessarily what the public sees as good. The Keenland thread you guys have going about $70 horses etc. is real for me also (although I did not have the horses you guys mentioned). And different spot plays could easily pick up one of these strange ones from multiple angles. I know I have seen it more than once, and I can assure you that not all do get bet down (and some do).

later,
Carl

[tomcat]

Hi, I love your discussion about angles and access. I am doing about the same. We should
share our plays sometime.
I wanted to comment abut ROI. A little of people worship at the alter of ROI. All things are
good if approved by ROI and bad if the the great GOD ROI disapproves.
But ROI is a history of the past. In determining angles, I think % is more important. Then
an examination and judgement of the current odds. It has never made sense to me to
refuse to bet a 5-1 horse because the ROI was so low.

[Carl]

Hi Tomcat,

Starting to come around to your point of view about win percentage also, albeit slowly. And then again, that would mean I would have to also embrace win percentage's big cousin, IV.

The thing I like out win percentage and/or IV is that it provides a stable counter to the high ROI plays I so enjoy. I now have a nice little spot play mostly based on IMPACT that I have been using that hovers around the 40% win rate and comes up with three or four plays a day. I like it a lot.

Still looking forward to Vegas, and of course as always "external circumstances" are fighting me. Same old same old.....

later,
Carl

[tomcat]

Hi Carl, I never say I'm right and your wrong or visa versa. I hope I have more class than that.
I just have an opinion based on my experiences. I think I have a little more experience dealing with horses than most, and have been luckier than many with my wagering.
The biggest thing I like about horse racing is the really nice people I have met. If I never cash another ticket, I will always be grateful for friends I have met through Ken and HTR.

Good luck

[hurrikane]

Tomcat...I'm leaning toward your barn on ROI. Although the nice ROI looks good it is just a reflection of the past. I have had plays that have a low win % but high ROI that if I played them automatically would put me in ruin. On the other side of the spectrum is the high win % but low ROI..ie the ****K play George was running earlier. High precent wins but the low mutuals could put you in the poor house fast if it went south real quick. I suppose somewhere out there is a comfortable mix of the two. If you or anyone finds it please let me know.

[Glen]

Carl,

Good to hear the that the double and triple plays still perform well. Although I have not tested it, I am kinda tracking it on netcapper player picks under poster GR. Did great last week with above 2.00 ROI but think I'm 0 for 5 or so this week.

Regarding ROI, I will stay in the ROI is important barn. This is a difficult question to answer, perhaps a question with no correct answer. When ever I come across a query with a positive ROI, I search out and see if there was a $100 or $150 type winner to throw it way off, and also check for consistency in the winning prices. IMHO, the prob with win% is the lower the "off odds" the higher the win%, the higher the "off odds" the lower the win%. This is an area of interest to me though that I would like to learn more about. Also been fiddling with creating an "odds line" based on those factors. However, I have not put a lot of effort into this area because it is rare that I will watch the tote and wager since all my wagers are placed during "lunch break".

[MikeDee]

Just a couple of comments on this thread. I don't have any data to back it up but I equate the multiple spot play thing to multiple no. 1factors. For example, if the no. 1K horse wins at 30%, it can also be the no. 1 Power, PPscan and whatever else but it still wins around 30% of the time. Multiple no 1's do not seem to change the win% all that much. (multiple picks don't make the horse any better!!)

Re: ROI I think it has value as a indicator, nothing really can project how a spot play will do going forward and we all have many examples of one that can look good backward, but don't work well going forward.

I like to look at the wager series. How many times did the series go under water? how long were the losing streaks? What was the return on the actual out of pocket wagers and not the ones when you would have been playing on winnings. Sort of a cash flow analysis.

[MtKen]

Hi All:

I just ret'd from a little spring fling in So. Utah (it's snowing here today so the trip south was much appreciated). The new board looks great & I'm sending $ to Rick today.

I just wanted to mention that I always break my plays or queries up into sections that are grouped by the final odds. >=2/1 to <4/1, >=4/1 to <6/1, >=6/1 to <9/1, etc. That way the ROI isn't misled by the 1 big winner in the sample & the better win % of the 2/1 horses' doesn't influence my thinking about the 35/1 shots.

After re-reading many of the old posts, particularly George's, & testing my db ( which is actually in about 15 spreadsheets) I'm a convert to track selectivity. Does everybody test their plays on a track-by-track basis?

MtKen

[Carl]

Tomcat,

I was attempting to agree with your point of view about win percentage being important. If it didn't come across that way, sorry for the misunderstanding. I very rarely communicate with anyone over the age of 11 these days, my skills may be rusty.

later,
Carl

[Carl]

Mike,

About the multiple "1"s being a better or worst play. Don't think it is necessarily analogous to how multiple spot plays will do.

Looks to me like the the multiple "1" play is almost always a higher percentage lower average pay situation, making it a wash. I think George has some spot plays with lots of one's that win 65% of the time (and I think forget to show a profit in so doing).

For whatever reason, this did not happen with my spot plays. This surprised me a little. Frankly, I was not sure when I embarked upon this project if horses picked out by multiple spot plays would even be as productive as the single plays (for the reasons Glen mentioned, I too thought "everyone might be on it").

And I wouldn't mind a second opinion. Anyone out there (HTR user or no) tested their spot plays for either greater or less profit with multiplicity?

later,
Carl

[Carl]

Think I'll try my first HTR attachment.

Here's a "one'd up" Query that tests at 44% winners (all races Jan-March 2001). Actually not too bad a "one'd up" horsey, play only loses 3 cents on a dollar.

later,
Carl

[fred4now]

What seems most important to me in the ROI vs. Win% deal is the proper ratio between them, like if Win% is 8 and ROI is 1.75, I'm probably in trouble.

I did the multiple bet thing and when things were good they were very good and when they were bad uh oh!

Missed most of this thread today, am over in Sedona AZ on anniversery, so better get off computer and do my duty.

P.S. When i am testing new query I put <50 in nWIN to make sure it isn't skewed by a couple of big ones.

[MikeDee]

I see your point about multiple no1s Carl

MtKen - I am also a track by track player. I'll try my hand at a attachment since I was the major complainer about not being able to attach on the old board:)

The attached is for all fast dirt at 6 different tracks for races in 99, 00 and 01 thru last week. the filter is Espeed no 1 with a e dif of >=10.

It is interesting to note the large swings in the number of qualifing plays, win% and ROI.

Carl have you ever looked at your spot plays on a trk by trk basis?

[George]

Interesting thread. MtKen, we are not allowed to use dirty words on here.
To a southerner "snow" is a four letter word. No more use of it please.
Think track selectivity is extremely important. Have found very little that
will hold up across tracks.

Carl......my K1 play over time has become a 56% winner with a 6.7% roi.
Roi would be better if late simulcast money did not seduce me into wagers
at 2-1 that end up 4-5 at the wire. Agree with Tomcat that win% in my
experience is more important when developing a play. I also look at the
series of plays similar to Mike Dee looking for bombs distorting the roi and
for how far negative a series goes. Like a happy medium.

The best overall play I am currently running is a "multiple" spot from impact
and velocity section of HTR. When one comes up I "double" wager. Have
not ran stats in access to justify the double-bet but can tell from bankroll
that it is working out all right.

[Rick]

Originally posted by MikeDee
I'll try my hand at a attachment since I was the major complainer about not being able to attach on the old board:)

I have just removed xls and mdb from the attachment list. The whole spreadsheet is copied over to the online database. The files are too large.

When you want to upload your spreadsheet, copy and paste to Word or Wordpad and save as a rtf. That takes a lot less space and will allow for many more attachments in the online database.

If you have a choice, use rtf files for up loads instead of txt for any formatted text like something in columns. The rtf files will display correctly online where the txt files will not.

Thanks

[Carl]

Rick,

Mike's excel file opened up beautifully for me. Only funny was that it asked for my HTR username and password first.

Mike,

Interesting to me that your worst track (RD) and best track (MTR) had similar win percentages, I think 19% and 20% respectively. The big difference was the average pay.

Have you checked the average off odds of the plays at both tracks? IF the average off odds were similar also, I would say it is at least possible your play just got "lucky" at MTR, and "unlucky" at RD, and would be willing to bet that the phenomena reverses itself next season. If the average off odds were also 50% different, I would agree you are much better off playing at MTR.

Leaves me back at Tomcat's point about average win percentage also saying something.

later,
Carl

[Rick]

Carl,
Thanks. I really put my foot in my mouth that time.

After your post I decided I better check my facts instead of just ASSUMEing that what I think is right.

It turns out that the xls file takes up less disk space than the rtf file. Maybe I should remove the rtf??:eek:

Can't do that because they display properly, as you so clearly noted that so do the xls.

[MikeDee]

Carl, I'll take a look at RD and MNR on the ave odds to see if that's it.

Rick - Glad to see we can keep the spreadsheet. But I also did my first fixed format using the code control. That works real good as well and you can put your results right in the message.