a customer retention science tool.

  
   
DHA, inc.    

When your GOOD INTENTIONS are foiled it may be because you can't make everyone happy this year, but you can pass three tests .


 
 


SOME ESP MADE POORER

     1. When shadowprice.com autopilot pitted aggressive challengers against the incumbent’s standard offer, its solver could not find Pareto efficient solutions for all competitors in price war or price peace. This is the same as saying the solver’s brain discovered (in repeated trials) that in this market setting, some competitor(s) had to finish the year less profitable than their starting values anticipated. For an alternative to aggressive challengers pitted against the incumbent’s standard offer, the beta-test limited the challenger posture to only requiring TV spot advertising for:
  1. a competitor who offers renewables in its supply mix

  2. a competitor who offers a performance-based flat rate

  3. a competitor who offers a cash incentive to switch providers.

This does not mean that only challengers who meet these requirements will advertise. shadowprice.com always requires that at least 4 challengers (including the player if he/she so chooses) advertise in every series game. It does mean that in every series game it is possible for one or more challengers, not including the no-frills provider slot, to not advertise.

  
             
  THREE PASSING GRADES     2. shadowprice.com's Game Assignment Seed algorithm passes three tests of pseudo-random selection:

 
             
  - for random,
uniform 		     
  1. A Chi-Square test says we cannot reject the null hypothesis that the seeds are random and uniformly distributed throughout the game assignment space. This test was passed in 60 of 62 90-game series comprising the beta test reported here. The two series failures (Autopilot Series 12-1 and 12-4) were just across the line denoting pass.
  
             
  - for challenger
slot landing 		     
  1. On average, half of a 90-game Autopilot Series’ player assignments should be to the four lowest-priced wholesale electricity slots (regardless of base load power cost regime) and half to the four highest-priced wholesale electricity slots, under the bus stop option “Player is just as likely to be the LCOS as any other challenger” selected for the beta test. For the 4140 beta-test games played in which shadowprice.com autopilot could find Pareto efficient solutions in price wars possible under market power manifestations 1, 5, 6, 7, and 8, the average number of four lowest-priced wholesale electricity slots was 45.8 per 90 game series.
  
             
  - for price wars
expected 		     
  1. The expected number of price wars per 90 game series should equal the number played under market power manifestation 4 plus half the number played under game strategies 1, 5, 6, 7, and 8. For the 4140 beta-test games played in which shadowprice.com autopilot could find Pareto efficient solutions in price wars possible under market power manifestations 1, 5, 6, 7, and 8, the average difference between the expected number encountered and actual number encountered per 90-game series was 0.74.
  
             
  MUTUAL DISADVANTAGE BIAS     3. When shadowprice.com autopilot pitted non-aggressive challengers who were only required to advertise if they also did one of the three things cited in number 1 above against the incumbent’s standard offer, it could not find mutually advantageous (Pareto efficient) solutions for all competitors in price wars unless the incumbent was prohibited from making price offers to customers below its wholesale cost for power. If shadowprice.com fails to find Pareto efficient solutions for all competitors in repeated trials, it draws new game assignment seeds until Pareto efficiency is achieved under price peace. Moreover, as it encounters difficulty finding Pareto efficiency under price peace it continues drawing game assignment seeds until it finds a distribution of player assignments to wholesale electricity slots that is mutually advantageous to all competitors. This skews the Game Assignment Seed probability space away from the expected number of lowest priced wholesale electricity slots as well as the expected number of price wars per 90-game series. Specifically, the actual average number for the four lowest-priced wholesale electricity slots, for the 1,440 beta test games in which skew occurred, was 54.5 per 90-game series. For all 5580 beta-test games played (in which shadowprice.com autopilot could sometimes find Pareto efficient solutions in price wars possible under market power manifestations 1, 5, 6, 7, and 8 and sometimes not), the average difference between the expected number of price wars encountered and actual number encountered per 90-game series was 8.13.  
             
  ARTISTIC LICENSE TAKEN     4. Between autopilot Series 4 and 5, in an attempt to improve solution probabilities from reseeds of the shadowprice.com profit optimizer, we altered the topology solution space by increasing the change in random seed by about 18% for each reseeded game trial. We verified that doing this did not exceed the long integer boundaries set by Delphi Pascal for random number seeds for the largest possible drawing. Increasing the change in random seed does reflect artistic license taken under beta-test conditions.  
             
 

SHORT SERIES TEST DESIGN INTENT

     5.

Series 6 – 1 through 7 – 2 were run under the high base load power cost regime simply to test the proposition that the shadowprice.com profit optimizer solver’s inability to wage price wars under market power manifestations 1 and 5 through 8, with the incumbent pricing unrestricted, was owing to the beta-test design intent to run the low base load power cost regime first. This proposition was confirmed.

  
   
  What are some of the awesome details from which beta test pics were derived?
   
  You can purchase Autopilot 3.5 on a 4 gigabyte memory stick and test your own favorite portfolio design and most likely market power manifestations where you reside.
 

 

 

      

 

© 2001-2008 Dan Hamblin & Associates, Inc. All rights reserved.