The winning bid that we want to predict at Pricebender is completely random. But we are not totally blind with information. We know at least the range of the winning bids by scanning through the records of winners.
We can use two kinds of strategies.
1. We can use logical prediction. - This means, you look at the value of items out for bidding and the the types of bidding methods. (regular, first time, junior, big dog, double MRP, and No auto bidding).
You can look at the records of each item bid within the range of the winning bids. It is good to choose 1 bid price for same item with such method of bidding method and stick with it until you win.
2. We can use scientific statistical analysis. - I have done this myself, I down loaded 6 months of winning data and put them in excel. I sorted it according to item.
Then each item I also sorted it according to time that it was won (24 hours). from this classification, I took the average.
After that I made a matrix, with different items in rows versus the 24 hours in a day in the column. I put the average winning bid at the intersection cell.
I use it to predict the winning bid of certain item at a certain time. I stick to that number until I win. So, far so good.
note: Instead of average (mean) you can also use the mode- the highest number of winning bid occurrence.
less
The winning bid that we want to predict at Pricebender is completely random. But we are not totally blind with information. We know at least the range of the winning bids by scanning through the records of winners.
We can use two kinds of strategies.
1. We can use logical prediction. - This means, you look at the value of items out for bidding and the the types of bidding methods. (regular, first time, junior, big dog, double MRP, and No auto bidding).
You can look
...
more
