Add training workflow, datasets, and runbook

training_data/curated/text/b38517fb5508e4a7ddc8b72a255f4da7f4bc0e3819a71f5fd5d587ca4b4863cc.txt

@@ -0,0 +1,50 @@
+286
+A Complete Guide to the Futures mArket
+series means that past prices in a continuous series will not match the actual historical prices that 
+prevailed at the time. However, the essential point is that the continuous series is the only linked 
+futures series that will exactly reflect price swings and hence equity fluctuations in an actual trading 
+account. Consequently, it is the only linked series that can be used to generate accurate simulations in 
+computer testing of trading systems.
+The preceding point is absolutely critical! Mathematics is not a matter of opinion. There is one 
+right answer and there are many wrong answers. The simple fact is that if a continuous futures price 
+series is defined so that rollovers occur on days consistent with rollovers in actual trading, results 
+implied by using this series will precisely match results in actual trading (assuming, of course, accu­
+rate commission and slippage cost estimates). In other words, the continuous series will exactly paral­
+lel the fluctuations of a constantly held (i.e., rolled over) long position. All other types of linked series 
+will not match actual market price movements.
+T o illustrate this statement, we compare the implications of various price series using the sideways 
+gold market example cited earlier in this chapter (i.e., gold hovering near $1,200 and a forward/
+nearby contract premium equal to 1 percent per two
+­month spread). A trader buying a one­year for­
+ward futures contract would therefore pay approximately $1,273.82 (1.016 × $1,200 = $1,273.82). 
+The spot price would reflect a sideways pattern near $1,200. As previously seen, a 60­day constant­
+forward price would reflect a sideways pattern near $1,212 (1.01 × $1,200). A nearest futures 
+price series would exhibit a general sideways pattern, characterized by extended minor downtrends 
+(reflecting the gradual evaporation of the carrying charge time premium as each nearby contract 
+approached expiration), interspersed with upward gaps at rollovers between expiring and subsequent 
+futures contracts.
+Thus the spot, constant
+­forward, and nearest futures price series would all suggest that a long 
+position would have resulted in a break­even trade for the year. In reality, however, the buyer of the 
+futures contract pays $1,273.82 for a contract that eventually expires at $1,200. Thus, from a trading 
+or real
+­world viewpoint, the market actually witnesses a downtrend. The continuous futures price is 
+the only price series that reflects the market decline—and real dollar loss—a trader would actually 
+have experienced.
+I have often seen comments or articles by industry “experts” arguing for the use of constant
+­
+forward (perpetual) series instead of continuous series in order to avoid distortions. This argument 
+has it exactly backwards. Whether these proponents of constant
+­forward series adopt their stance 
+because of naïveté or self ­interest (i.e., they are vendors of constant ­forward­type data), they are 
+simply wrong. This is not a matter of opinion. If you have any doubts, try matching up fluctuations 
+in an actual trading account with those that would be implied by constant
+­forward­type price series. 
+you will soon be a believer.
+Are there any drawbacks to the continuous futures time series? of course. It may be the best 
+solution to the linked series problem, but it is not a perfect answer. A perfect alternative simply 
+does not exist. 
+one potential drawback, which is a consequence of the fact that continuous futures 
+accurately reflect only price swings, not price levels, is that continuous futures cannot be used for any 
+type of percentage calculations. This situation, however, can be easily remedied. If a system requires 
+the calculation of a percentage change figure, use continuous futures to calculate the nominal price