FAQ ---- Poologic Calculator
Use Ctrl-Refresh (IE) or Shift-Reload (Netscape) to get the latest data. Wait for the applet to load...
The ROI calculator is designed to help you make the best picks for champion, taking into account your opponents picks. Bet one sheet per champ, in the order of the ROI values of the possible champs. Create each of your entry sheets by using the Poologic Calculator with the champ constrained to the required team.
If the winner of your pool is often decided before the championship game is played, then the ROI calculator is not applicable to your pool. It applies best to pools with only standard scoring, since we have no estimate of the relative value of an EPM sheet in a pool with incentives for picking upsets, because we have no theoretical model for the betting pattern of a typical player.
My simulations support the claim that this ROI calculator does a pretty good job of predicting the top four teams that will provide the the best ROI. But it tends to overestimate the ROI value for all teams. Your actual ROI is probably not going to be much more than $2 for each $1 your bet in entries fees in a standard scoring pool. In other words, simulations predict that you will perhaps double your money on average over the long run if you are playing a pool where nobody else is using Poologic.
Estimated total entries: Enter your estimate of the total sheets entered in your office pool, including your entries.
Estimated value of a EPM sheet: This is the estimated value of the EPM sheet calculated by the Poologic Calculator relative to the typical sheet bet in your pool. From simulations of a limited number of pools from past years, I estimate the value to be 2.5 for standard scoring pools with round multipliers of 1,2,3,4,6,8. This is the default value, and it it is probably a good choice for all standard scoring pools.
Probability Model: This is the data source for the win probabilities. For information about the probability models, go here.
Est. entries: For each team of interest, enter your estimate of the number of sheets that will be bet for that team in your pool, not counting your entry sheets. You don't need to provide an estimate for all teams, but just the teams that you are considering for champ. Yahoo Tournament Pick'em helps with the estimates. Yahoo is providing frequency distributions on the entries in their nationwide pool. The 2005 link is here and the same link may work for 2006. Take the percentage from the Yahoo distribution for the final round winner and multiply it by your estimate of the total opponent entries in your pool. For instance, if you estimate 150 opponents in your pool and Duke has been picked by 30% of the Pick'em entries to win it all, then estimate that 45 of your opponents will pick Duke. This should provide good estimates for all teams except localized favorites. You can try to calibrate for localized favorites or just avoid betting any localized favorite as champ.
Remember to hit the Return key to cause the ROI to be updated. The ROI (Return on investment) is your expected return when you bet this team as champ. For instance, if ROI=1.5 then, on average, you should have an expected return of $1.50 for each $1 that you bet. (Note that the ROI may be somewhat underestimated, especially in very small pools, because it does not cover the possibility that you can win the pool when no player correctly picks the championship.)
If "ERROR" Appears in the ROI column, then check that you entered valid non-negative numbers in the data entry fields. Use zero, not the letter O.
The applet calculates ROI = P*T*(V/(O+V)), where P is the probability that a team will win the championship, T is the estimated total entries in the pool, V is the estimated value of an EPM sheet, and O is the estimated opposing entries where the team is bet as champion. (If you set T=1 and O=0 then ROI = P, the probabilty that a team will win the championship.)
Note that, unlike the Poologic Calculator, the ROI Calculator is not an implementation of a method that has been published in a peer-reviewed statistical journal. So, user beware.