function NewQas=UpdateQFinger(Play,Payouts,Qas)



%   This program updates the value table (as it depends on possible plays)
%
%   PlayNim also includes definitions of JStartEnd and Qas
%
% Input:
%
%    Play  is a vector. The jth entry is the choice made by the jth player.
%
%
%    Payouts is a vector. The jth entry contains the payout
%         to the jth player of the current game.
%
%     Qas:   This is a cell. The jth entry corresponds to the jth player
%           and is comprised of two columns. The kth row contains the current estimate of
%           the value of playing k and the number of times that k has already been chosen
%           by the jth player.
%

% Output:
%
%    NewQas is the updated version of the values, after compensating for
%    the last
%                   round of rewards given by Payouts
%
%%     Algorithm   is of the TD type
%
%     Q_k+1=Q_k+alpha(r-Q_k)
%
%    r is the reward that is eventually delivered
%      at present: there is no discounting
%%
%
%    alpha is  the learning rate;

%% %   Add a weigthed version of Payouts (average it in) to the old state values
%     to get the new state values
%
%  the more the magnitude of the TDError is non-zero, the more that learning is
%      taking place
%
NumPlayers= length(Play);
NewQas= Qas;                    %  The new  values

for jSubj=1:NumPlayers
 
    MoveInd= Play(jSubj);
    TDError = Payouts(jSubj)-Qas{jSubj}(MoveInd,1);                             % CHANGE ME

    NumAlreadyp1 = Qas{jSubj}(MoveInd,2)+1;
   
    NewQas{jSubj}(MoveInd,1) = NewQas{jSubj}(MoveInd,1) + (1/NumAlreadyp1)*TDError;   % CHANGE ME
    
    NewQas{jSubj}(MoveInd,2)=  NumAlreadyp1;    % update the number of times this move was made
%
end


% NewQas= Qas + alpha*(ToAdd-Qas);
%%