In The Eudaemonic Pie book "The Newtonian Casino" Chapter 9 entitled "Lady Luck" is the following: 'Yet another croupier offered the Eudaemons an elegant demonstration of what gamblers call the dealer's "signature."'
An explanation is then given of how dealer signature can be achieved and exploited! This is pure fiction. There is not the slightest proof that even one dealer has ever been able to throw a signature. No research has been done, no data exists, nor is there a method to gather data. Although there is not one scientist who has any evidence of dealer signature none condemned or disputed the claims in this book. On the contrary, it is embraced as an authoritative work. Dr Ed Thorp, despite his disbelief in dealer signature, checked the manuscript and gave the book a big wrap.
The account now moves from theory to practice. The player, Norman Packard, gave the dealer chip tokes to bet at his discretion. The dealer always placed the tokes on number '17'. When asked why, the dealer said, 'Because if I do everything just right, I can actually hit '17'. He added, 'I flip the ball in this nice regular way when the zero is lined up in front of me, and I swear I can hit '17' with better-than-average odds.'
Norman's team-mate Doyne Farmer, with no data to justify his action, and only the dealer's word that he could, and would, do as he said, placed additional bets on '17' and won God knows how much. One would think they would have to have won many thousands of dollars if it was to convince him of the dealer's signature and so be worthy of mentioning in the book. Doyne gave not a shred of proof that he applied any statistical tests to check out any dealers for signature. Let us assume he did apply 'stats' tests. How many spins would he need to get a statistically significant result to conclude a dealer had a signature. At least 1000. Presumably, the dealer only lined up zero when a Eudaemon placed a toke on '17'. But how often did they place a toke? Surely not every spin or the casino pit-bosses would have woken up and objected. Let's say every 5th spin. That's 5000 total spins needed from this dealer on this wheel. The total time to acquire those spins would take nearly a month of continuous play, and that's assuming he always operated this wheel. How many other dealers did they check out? Just how much did these side issues interfere with and degrade and distort the data for the prediction theory they were primarily researching?
The book tells of "That night we all went away happy." It doesn't say how much they all won that night, but it might be interesting to speculate. At about 40 spins per hour for a shift that's about 300 spins. Assume the dealer attempted his signature about every 5th spin that means 60 bets. Only two hits of '17' were needed to show a profit but how many hits were actually thrown to convince them that the dealer was throwing a signature and was not just being lucky.
Why did this supposedly clever dealer only ever use the one throw, always lining up '0' to hit '17'. There are 38 such pairs using the identical throw that this dealer could choose from. Surely he would have learned to throw any pair at will. Alternative throws seems a sensible idea as they would help conceal his skill from his casino bosses who would object to him spinning winners for players and some for himself. It might be thought that he used '0' only because it was easy to identify, but why did he never line up the more obvious '00' and shoot for '18'. This is on the opposite side of the wheel but all it needs is the identical throw. (Check it yourself). Surely he would have discovered that. Also, the 17/18 split would make a nice alternative bet for either throw.
But even if '17' did come up above average there are reasons other than dealer signature that could be the cause. Let's say the wheel itself was biased enough to cause '17' to come up more frequently than by random chance. For all we know it might have dropped into '17' regardless of what the dealer lined up for his throw merely because '17' trapped the ball because of some fault. Did Doyne Farmer do a 'stats' test to check the wheel for bias? No! Did he or any of the team question anything about their methodology? No! They have assumed that the dealer is the only possible cause for the excess '17s', which is appallingly unscientific.
It makes one really wonder how much of their win the Eudaemonic Pie players should have attributed to (a) their prediction theory, (b) how much of it to the dubious dealers' signatures and (c) how much to the wheels that were very likely biased but were never checked. Despite a plethora of words in the book, there is no breakdown of the actual causes for their win. It might be that the prediction theory component actually lost and their net win, assuming they did win, was attributable to biased wheels only. They offer nothing to the contrary. The book contains no data, no spins, no formulae, no statistics, no results, no dollars won, no proofs, nor anything to justify their conclusions. Has anyone questioned it since? No! It has been swallowed hook, line and sinker!