When adding an additional size to an AI solve tree, I’m proposing that the increase in EV by adding that size is calculated and displayed.
Currently the only way to ‘compare’ this is by looking at the aggregate EV of each player in each of the two sims individually. However, the increase in EV by adding a size to a players strategy at a node is actually calculated by taking the max EV for each combo from the two sims, and then re-calculating the aggregate using these values.
As a simplified example of this, if a player’s range is AA, KK and QQ at a node, and in the single size sim they have EVs of 20, 15 and 10 respectively, that player has an aggregate EV of 15 [(20+15+10)/3]. When an extra size is added to the sim, those EVs become 18, 16 and 11 respectively, but this still results in the same displayed EV using the current calculation method [(18+16+11)/3 = 15]. However when we account for the opponent treating our strategy as if it were a single sized one, the actual EV of adding the size is 15.67[(20+16+11)/3].
The logic behind this is that if we add a size to our strategy (whilst retaining the previous ones), and our opponent does not adjust accordingly, then they are unable to reduce the EV of our hands (AA in the example above) to that of the multi-size strategy. However they will still lose EV to the hands which benefit from a multi-size strategy.
Please feel free to reach out for additional details.
I previously mentioned this in this Discord comment: https://discord.com/channels/756231695522332692/928166782131400754/1523544688173518921