The main resons are:
1. The cancel time and the signal delay tolerance/allowance given--up to 30 seconds total depending on rules in jurisdiction of signal exporter AND ALSO the jurisdiction of the bettaker: in case of delay somewhere along the transmission line(possibly caused by satellites, power glitch, internet relays, maybe old telephone lines, etc.)when the racetrack sends the stop signal they account for transmission delays glitches in that signal arriving to/from all simulcast locations the world over, in addition to the void time given so that the bettaker is not stuck with a # mistake, customer not having the funds, etc.
Few in the industry will willingly divulge all the nuts & bolts to the general public, but: off-site simulcasting & its entailing huge amounts of money being wagered, would not be so proliferate if the all parties were not covered from the get go re any potential exposure(They have to account for all worst-case scenarios & eventualities down to the very last Podunk BetTaker in some remote island with customers on crappy dialup-- the signal has to travel to/from customer at home through crappy wiring, satellites, undersea cables, beacoup internet relays all over the world, to Podunk, to one, maybe two middleman hubs, to the racetrack).
This is borne out by reports i've gotten of past-posting online at one of the major bet takers, 'after the race was over'-- further, keep in mind that supposed live racing on both net & airwaves is often delayed by at least 15-30 seconds!
2. Betting at BOTH ends does not always end when the first horse loads, but rather when the field is off....
In spite of the recent spotlight on issues like these, to this day there are still teams at racetracks that bet the races after they're off-- with both live tellers & at the SAM's.
Why not? Does there exist a whale anywhere that wouldn't want to at least cancel his bet if it encounters early trouble-- or even further: whether betting with live tellers or having direct computer access to the feed & a batch betting utility in place-- why not instead put all the money that would go to one horse, the same amount on each of the likely winners & some longshots in all pools, then after the quarter mile press cancel on the ones you want?
Anyhow, even with no cheating going on-- if you add the refresh cycle, delete times & transmission delays-- it is almost impossible to have a final odds BEFORE a sprint is over.
And finally, keep in mind two final thoughts:
1. historically favorites tend to drift down
2. more and more each day: tote money before 2-3 min to post is mostly uninformed money, & to boot only a small proportion of the final total. With the ever-increasing proliferation of simulcasting, internet betting, & particularly rebate shops & betting exchanges in far-flung locations-- it's mostly wise money coming in extremely late: in the form of non-parimutuels laying off excess exposure & their sharpest customers' bets, plus hedging/scalping of exchange bets-- most horses that have a decent shot of winning relative to their odds, not just faves, will get late money.
The ONLY solution to the falling odds after the race starts, in addition to the very latest technology applied EVERYWHERE, is for betting at BOTH ends EVERYWHERE to end "the instant the first horse touches the starting gate as it's being loaded". And in order to appease all places over concerns of lost revenue & to train bettors, a new system can be started thusly:
Normally these announcements are made: "the field is approaching the starting gate", then a bit later "the field has reached the starting gate"-- whereupon this can be added just before loading the horses:
a final 30-second ditty from the track bugler and a Countdown-- with a new year's type ball falling descending from the 30 second mark on a corner of all screens.
If that is adopted everywhere AND a consensus mandated at ALL jurisdictions WORLDWIDE re no bet aftwards, maybe, just maybe............
ointer:
