|
Dear Professor,
I had this "little" project on self-competing AI that was based on SC2, and problems you've described, were challenged by me as well. Truth to be told, this is extremely complex task, nonetheless can be atomized with an appropriate level of understanding the game and tailored toward group of any participants. What I had covered was from simple to mid-level tasks, while high-leveled ones were only in theoretical state at that time.
If you'd want to know more from my perspective - I'm more glad to help. Just please PM me, and we'll exchange e-mails to push the case further.
|
On February 18 2019 20:25 SimonDennis wrote: Ok, got it. Does anyone know of a way to truncate replays?
Simon.
Hi Simon,
One very simple way to truncate the replay is to watch it, then use the "take command" function from the UI at 5 min and surrender right away.
Following this method, SC2 will create a new replay file that stops at any time you wish and the original replay remains unaltered.
I think Rodya's suggestion to choose a game like Serral vs Maru at WESG 2017 where both players almost die is a good one.
Another way to simplify the problem would be:
Starting from a situation where player A took a significant advantage during the first 5 minutes and player B is clearly behind.
First you have to ask whether the group of subject agrees that player B will be fighting an uphill battle or not. Does the latter have any advantages like upgrades, tech, units? Then, you could ask what are the options available from there so that he/she can come back and win? Finally, what is the decision that will have the highest chance of winning in this situation?
|
Its interesting. When I was first thinkung about this I was assuming teams would reason primarily from the mzcro, but its becoming clear now that the better teams would use the meta too.
Even the micro could be useful. If there is an initial battle you mihht get a sense of the micro skills of each player. If the weaker player is setting up for a major fight and it is going to be pretty even in terms of units then msybe you favour the player who is strongest on micro. But if the weaker micro player is playing like they are going to avoid a battle then maybe not.
There certainly seems to be plenty of scope for advanced reasoning. I like it as a domain as it has many of the aspects of geoploitical conflict, which is our ultimate target, but in a more measurable and more knowable form.
QUOTE]On February 18 2019 21:51 Oshuy wrote: Given a snapshot of the state of the game at a given time, there may be some reflexion needed to perdict a future state. Knowing how the game developped (replay), depends if there are still upcoming decisions or not. I agree that if the game contains harass/micro battles in the prediction time it is hard to predict how much it will slow things down.
If there is still an upcoming decision in the tree, main problem would be to identify the build. Would mostly be dependent of the meta which changes from season to season. Process would probably be:
- identify the map, given the map a date range is known,
- given the date range a list of builds that were/are used can be found
- match the current status to the list of status corresponding to those builds to get a set of probable decisions
There would be a number of possible choices, so answer would be a list of probabilities on the decision tree. (in a game, if there is no common response to the main choices, player either scouts for additional information or gambles)
If all basic decisions have been made and it is just a matter of adding the worker/units produced down the line, it is probably too easy. Moreover, if the input is available as a replay, anyone should be able to come with a pretty close result by simply playing out the next few minutes using the "resume from replay" function.[/QUOTE]
|
Yes, so the base set up needs to be taken into account when choosing. Probably the positioning also. If someone has a proxy then that is different than if they don't.
Simon.
On February 18 2019 22:27 BisuDagger wrote: Instead of using timestamps for your approach, you could try different economic states instead. By this I mean, give teams replays that you pause based on the number of bases each player has.
For example show: TvZ: * 2 base vs 3 base * 2 base vs 4 base * 3 base vs 3 base
Taking TvZ 2 base vs 4 base so many factors come into play leading up to this moment. Worker damage and harrasment, tech delays, army engagement, worker cutting, creep spread, map control. On paper you'd say an undamagerd 4 base zerg player with all gases is definitively ahead and should win the game. But that's the perfect world. From their you need to start to breaking down in what ways that advantage is mitigated.
|
One thing to keep in mind is that the plan would be to have people predict multiple facets of the game - the numbers of different unit types, the builds, the tech, the base positions etc. In the longer term these interact, but rolling forward just a bit even if a team is off on one measure they might be quite accurate on another. Maybe predicting the number of queens, for instance, might be easier than the number of zerglings. Predicting tech might be easier than predicting unit numbers. My intuition is that the different predictions range from being very difficult to quite easy, which is good for this task, because it means we can tell the difference between the reasoning ability of teams over a wide spectrum. For our purposes, the question isn't will the best teams get it perfectly right, it is will there be a difference between good teams and bad teams.
On February 18 2019 02:17 Rodya wrote: Too unpredictable, I have to say. If there is any major fight which occurs during the period between the replay ending and the prediction then it's unreasonable for there to be muchaccuracy. Especially because in general, if a fight occurs and one person loses, it might as well have not occured as the would-be loser could have pulled back his army and waited for a bit. Or alternatively, if a protoss is in the game, then the fight can depend critically on how good their micro is (which is so hard to judge, that you see the best protoss players in the world walking into big fights and getting smashed because they miscalculated how much mileage their skill would give them).
The only situation which I think could be good for your paper is after some unusual early game shenanigans where both players almost die and then you stop the replay and have them predict the game state a minute or two into the future. For instance, Serral vs Maru Game 1 at WESG 2017 at 7:45.
Problem is that you need high level players to play the games (since everyone else kind of sucks at this game to the point that predicting their behavior accurately is hopeless), and so it will be hard to find real examples.
|
|
|
|