View Single Post
Old 01-26-2011, 04:22 PM   #14
Ted Craven
Grade 1
 
Ted Craven's Avatar
 
Join Date: Jun 2005
Location: Nanaimo, British Columbia, Canada
Posts: 8,853
.
I spent a fair amount of time watching videos and reading articles and transcripts by George Soros about his 'reflexivity' concept. In the end, I got as much a summation of its relevance to handicapping and wagering from a few lines in Jerod Dinkin's article as I did from Soros (no poor reflection on him: he's a well spoken philosopher/investor and not a horse-racing specialist), so thanks again, Jerod!

The key observation for me is: in events comprising 'thinking participants' (e.g. people wagering on the outcome of a horse race, people trading stocks, commodities and other financial instruments) - all decisions are made with uncertain and imperfect understanding, thus stock pricing and oddslines are subjective guesses, estimates, and not facts. And these imperfect guesses cause feedback-loops by causing other individuals to respond (by placing bets or avoiding bets) which in turn cause other bettors to reassess their oddslines and their betting actions, etc, etc. It's a chaotic, noisy, messy storm of opinions, actions, counter-actions and modified opinions right up until the bell rings (and after...). And, filled with plenty of odds pricing and guessing mistakes: that's why most horses don't pay between EVEN and 3/2.

You may decide a horse is worth 5-1 and bet it at 6-1, but your own bet (what to speak of the bets of others, either independent of your actions, or in response to your actions) can cause the odds on the horse to go below 5-1, in which case you never would have bet it to start with (and neither would some others in response). Reflexivity; uncertainty; feedback loop.

I've never been able to appreciate the usefulness per-se of the Betting-Line oddsline in any of Doc Sartin's software and I believe he put it in there (the Bottom Line/Betting Line readout) as a 'get off my back' type response to those who nagged him about 'how could you omit such a fundamental aspect of wager decision making - you're not serious'. As it stands, the Betting Line is nowhere close to a 100% or even 120% oddsline because it isn't designed for that purpose (as far as I can understand).

Houndog's original post wondered about the applicability of reflexivity and uncertainty in pre race odds estimation to Sartin's concept of 'wagercapping'. Wagercapping (in my understanding, and sometimes application) involves using your key readouts (e.g. BL/BL), potentially hiding a Top 3 ranked horse having below certain cut-off odds (e.g. 5/2, 2-1) if you have to, then selecting 2 horses from among the remaining Top 3 such that the net odds (and resulting average mutuel) relative to your historical hit rate from doing so, at least breaks even (and presumably better). The keys to making a go of this approach are:

1) reliable readouts which identify horses with excellent chances to win the race (properly selected pacelines, omitted non-win-contenders due to form) - e.g. BL/BL, V/DC,
2) either acceptance of M/L odds as proxies for post-time odds, or waiting until close enough to post-time that you know the odds on your horse(s) won't drop unacceptably,
3) either, insight into when a top ranked low odds horse is the real thing, then to single it or pass that race, OR an emotional equanimity and good records to comfort you when you bet against that top ranked horse in favour of 2 of your next 3 ranked, and it wins anyway.

Wagercapping is not (in my understanding) about taking your 3rd ranked horse at 3-1 on the BL and betting it because it pays at 9/2, or your top ranked horse at 9/5 BL because it pays 5/2 (or both). It is about working with your ratings, composing a bet usually with 2 horses (or perhaps, equally, WP on 1 horse, or W on 1 horse and P on another, etc) such that your net odds (and resulting mutuel) if the lower paying one hits the percentage of the time your records shows it does, combined with the net odds of your higher paying one at its recorded hit rate - will result in a positive ROI over a number of events (e.g. a cycle of 20 bets). And everything is quite imperfect - final odds fluctuate from when you bet, your historical records on hit rate, average mutuel and factor ranking by mutuel (e.g. the traditional 'Wager Decision Form' showing mutuels by BL/BL tier) may fluctuate significantly cycle to cycle, track to track, distance structure to surface, etc - and your knowledge from bet to bet about how low a net mutuel you can accept can be imprecise (and likely a total crap-shoot and mere gambling if you have no records to consult).

I have tried to apply wagercapping (again, as I understand it) for years, and in general often had rather dissatisfying (read: emotionally trying) results. As often as I'd pat myself on the back for 'hiding' the Top BL/BL horse at 2-1 or 6/5 and cashing in on either a 9/2 or 8-1 of my 2 other bets, I would lose my entire bet when that Top ranked 'hidden' horse won. I tried to develop measurements of when to accept the Top ranked horse at face value: 2 points gap on BL, lone VDC #1, lone Early, best Late in a field of challenged Earlies, top 2 betting odds, etc. Sometimes it worked, sometimes it didn't. Maybe I missed something.

The past year or so, I have arrived at a betting practice which respects some things I have observed about myself, and which fits with my longer term goals:

1) I hate to lose money,
2) I feel worse about losing money than I feel great about winning money (see point #1),
3) I want to bet a lot of races and significant money (always relative to my bankroll size, hit rate and advantage - i.e. some flavour of Kelly) and earn good rebates for every bet.

To that end, and perhaps accepting that times have changed since Doc penned his immortal studies and theories, I now never 'hide' any horse from my screen, and consider a 2 horse bet if I can get 3/2 on the low side (i.e. minimum $5 payout for $4 bet, or 25% ROI on that event) and at least 3-1 on the high side (i.e. minimum $8 payout for $4 bet, or 100%). In this, I find myself focusing on the horse and its ranks first, and on odds second. My betting records tell me my choice of 2 horses to win-bet result in a winning wager about 50% of the time on average (though including some wild swings), which is the end result of paceline selection, recency and condition requirements, class of competition, performance in the paceline selected, a few factor readouts, odds ranks from the tote board, and to a certain extent a factor decision model. Most of the time, my odds focus is knowing whether I'll get my 3/2 minimum (and that, for a buffer against sometimes dropping below EVEN after I bet) or more specifically what my low to high range of payouts might be, then passing when I don't like the components of the range. I am now more happy betting a 3/2 and a 3-1, or a 2-1 and a 5-1, and never say no when offered a pair of 3's. Around 8-1 or so, I like to split my wager WP on one of the horses (again, I hate to lose money, and I want to bet frequently and collect rebates). Occasionally I will single my top horse at EVEN odds. Quite simply put, there are many more horses below 3-1 than there used to be, so either we have to get accurate enough to play with this distribution of odds, or play less, waiting for higher odds events. The difference between these two stances is a psychological/emotional one and depends on one's objectives in participating in racing (i.e. casual enjoyment/stimulation versus significant return on capital deployed, requiring betting either more events or higher units - not saying the latter objective cannot also be both enjoyable and stimulating).

But I never spend any time worrying whether I got my oddsline right on my top 3 or 4 horses. I worry about whether my paceline selection is consistent, whether I have reasonably assessed a horse's current form, whether there is too much unknown or unknowable about a given race. I appreciate that when anyone waits for 10-1 (on presumably their top few horses) they buy a lot of room for error but presumably bet less (in the straight pools). And you don't have to worry overly about an accurate oddsline, whatever that is, given how much is subjective and unknowable in race analysis: most horses win at less than 10-1, so it's an easy way of betting 'overlays' - though I appreciate that the skill required to both find them and wait for them is considerable and to be applauded.

Here's a relevant and interesting recent thought by Dick Schmidt (http://www.paceadvantage.com/forum/s...1&postcount=11)

Quote:
Originally Posted by Dick Schmidt
I have a solution to this problem that is extremely effective and extremely unpopular: don't look. Make your bets without reference to the odds or morning line, then go on to work the next race. Don't watch the odds move, don't watch the race, don't check the results or payoffs until you are done for the day. Spend your time handicapping and then entering the outcome into a good record keeping program when the day is done. So long as your bankroll is larger at the end of the day, no single race, or any group of races matters. Put your time and mental effort where it actually pays off, not watching dumb animals run around or listening to the opinions of other dumb animals betting on them.

I told you that you wouldn't like it, but it is how most professionals play.
So reflexivity = uncertainty = feedback influence when thinking (including crowd/herd responsive) participants are involved. Markets are not efficient, pricing errors (odds mis-estimations) happen EVERY day. That's as much as I understand about reflexivity and wagercapping so far. But always ready to learn more! If anyone has alternate interpretations of what Doc Sartin meant by 'wagercapping', I'm all eyes and ears.

Any and all feedback appreciated.

cheers,

Ted
__________________

R
DSS -
Racing Decision Support System™

Last edited by Ted Craven; 01-27-2011 at 02:12 PM. Reason: spelling, clarity
Ted Craven is offline   Reply With Quote