Low security in PBEM

[ferokapo]

Do you have a specific algorithm in mind that would work with the current system?

I´m not actually aware of any specific algorithm for this but may be it could be implemented a file key that is checked for reloads.

I don´t see any way how this should work. The only way I can think of, aside from a central server registering every transaction (upload, download, loading, saving), is to watermark a game file with a unique ID. Once your file is loaded by your opponent, his CEAW would have to store that watermark in an encrypted registry. Only then would CEAW be able to tell whether it has ever loaded a specific file before, and write this information into the output game file, from where it could be read by your CEAW, issuing a warning like "Your previous turn has been reloaded 4 times by your opponent." Even then, it would be easy to work around this by having more than one CEAW installation. Never mind that you would still have to modify CEAW code first.

[GaryChildress]

To make things worst combat results system is excessively random so this is like a vicious cycle: people are more tempted to cheat because of the bad combat results they have got in a single turn and so on.

I don't know how easy or hard it is to create a reload warning for the game but IMO two EASY and VERY effective ways to combat cheating would be the following:

1. Make combat results less random. Seeing 10:2 odds turn out 1:3 is annoyingly ridiculous. Less variation from the predicted results would go a LONG way in curbing the temptation to cheat.

2. Make reloading WAY more cumbersome. I would even go so far as to remove the "load game" option from the main menu, except ONLY when you first launch the game. Basically when you first launch the game you could be given the option to load a saved game and then once in the game the only way to load a save game would be to quit to desktop and relaunch the game. Seriously, this would probably be even more effective than #1. Who is going to cheat if you have to go around the world just to get back to square 1 again?

In my experience cheating is more a function of how easy and how desirable it is to cheat. Cheating in PBEM is not the only way the game is spoiled. IMO cheating in single player spoils the game just as much.

[rkr1958]

1. Make combat results less random. Seeing 10:2 odds turn out 1:3 is annoyingly ridiculous. Less variation from the predicted results would go a LONG way in curbing the temptation to cheat.

To me arguing that combat results should be less random to reduce the incentive to cheat is a flawed argument. That really makes no sense to me. Either the system works and produces historically believable results (e.g., Battle of Midway) or isn't. If folks believe the latter then they need to make the case as to why they believe that.

2. Make reloading WAY more cumbersome. I would even go so far as to remove the "load game" option from the main menu, except ONLY when you first launch the game. Basically when you first launch the game you could be given the option to load a saved game and then once in the game the only way to load a save game would be to quit to desktop and relaunch the game. Seriously, this would probably be even more effective than #1. Who is going to cheat if you have to go around the world just to get back to square 1 again?

I've currently got 5 games going and to have to shut down after each would to load another turn would be a major pain in the butt.

[GaryChildress]

@ rkr1958: Then what do you recommend? As the game stands now, cheating is ridiculously easy and pays off big rewards. So long as that is the case it makes more sense to cheat than it does to play by the rules.

As far as randomness, I agree that war has some unpredictable factors that play in victory and defeat. I'm not saying do away with randomness completely, only have less variation. Radically divergent combat results should be the exception rather than the rule. I don't think Midway was as random as many think it was. It was in part testimony to things like poor Japanese carrier design, poor AA escort philosophy, superior Allied damage control practices and other factors. If you could "restart" Midway 10 times, the Allies probably would have won 9 out of the 10.

WWII was largely a war of statistics. Attrition played the biggest aggregate role in the outcome of it. Germany and Japan lost because they failed at the numbers game, not because they were unlucky in battle. At the macro scale that CEAW represents basically whoever has the greater numbers and whoever uses their forces best should win, not whoever happens to be luckier. There shouldn't be 10:2 odds with 1:3 outcomes. Sorry but I don't see the logic in so much randomness. Why even have attack and defense values or have estimated outcomes if everything is so random?

[Plaid]

1. Make combat results less random. Seeing 10:2 odds turn out 1:3 is annoyingly ridiculous. Less variation from the predicted results would go a LONG way in curbing the temptation to cheat.

It happens really rare, almost never when you have proper air support. Also personaly I always ask myself - what if this tank score 1:1 on that yellow corps unit. And if I see that entire offencive will be ruined with it, I most likely will not attack at all. Saves resources and oil, by the way.

Getting 2 this bad rolls in a row considered ~impossible.

[rkr1958]

As far as randomness, I agree that war has some unpredictable factors that play in victory and defeat. I'm not saying do away with randomness completely, only have less variation. Radically divergent combat results should be the exception rather than the rule. I don't think Midway was as random as many think it was. It was in part testimony to things like poor Japanese carrier design, poor AA escort philosophy, superior Allied damage control practices and other factors. If you could "restart" Midway 10 times, the Allies probably would have won 9 out of the 10. ?

Up until recently I would have agreed with that but I don't know. US dive bombers found their mark with 3 of the 4 Japanese big carriers with their decks full of armed planes and with the below decks stacked full of bombs just off loaded from those planes to rearm them with torpedoes to turn the tide of that battle. Before that the three torpedo squadrons from the three US carriers had been obliterate because they had attacked without fighter cover. I think only 3 or 4 crews out of 21 or 22 torpedo crews survived. The Japanese fighter cover had been pulled down by the torpedo bomber attacks and was off chasing after survivors. The dive bombers were at the absolutely maximum of their range. If fact, they only found the Japanese carrier force by detecting and following a Japanese destroyer. The Japanese had made a fateful decision to rearm their planes to attack the US carrier they had just discovered, which delayed their launch by an hour. Growing up I thought the Battle of Midway was a slam dunk for the US but after researching it in much more detailed it could have easily gone the Japanese way. The lost was a result of tactical combat decisions and good luck (for the US).

By the way when we're talking bad rolls were not talking about an entire battle. We're talking about one attack by a corps, air or naval unit. While I can't think of specific examples I'm sure there are numerous examples in WW-II when a corps performed much better, or much worse, than it was expected too.

I guess if you get a bad roll that devastates you then it part that due either to a dire situation or poor planning. This is not unlike in Avalon Hill's 3rd Reich having 2:1 odds and rolling a 4 followed by your opponent rolling a 6. Ouch ... while this is definitely poor luck it's possibility, even with odds of 1 out 34, was known before. Now with poor planning rolling a 5 on a 1:1 attack could also be just as devastating but not as unexpected. Also, these bad rolls (i.e., 2:1 odds with 4 followed by a 1 or 1:1 with a 5) resulted in the elimination of all (non-naval) attacking forces; whereas, in GS it's only a single corps, air or naval units that's impacted.

[GaryChildress]

OK. Just performed a test:

Predicted outcome = 10:2

actual outcomes:

7:0
7:2
6:0
5:1
6:1
4:2
9:1
6:1
7:1
7:0
6:1
7:2
3:1
3:2
7:2

1st thing to notice: the predicted results NEVER appeared out of 15 tries.

From there we have best (9:1) all the way to worst (3:2). Within the fist 6 tries I can get from 7:0 to 4:2.

Where is the 10:2? Why even have an outcome prediction at all?

[GaryChildress]

But why were the Japanese planes on the deck and not on the US carriers? If the Japanese were able to warm up their plane engines in the hangar deck as US carriers could, then they might not have been there. Had the US planes all arrived on time, then they would have had fighter escort and bombers would probably have gotten through the IJN fighters just the same. The torpedo bombers probably would have been the primary targets of the fighters either way and the Devastator was a horrible aircraft, regardless of when it arrived on the scene. Midway was probably stacked in the Allies favor from the get go. What happened was probably little more than what should have happened.

[rkr1958]

OK. Just performed a test:

Predicted outcome = 10:2

1st thing to notice: the predicted results NEVER appeared out of 15 tries.

From there we have best (9:1) all the way to worst (3:2). Within the fist 6 tries I can get from 7:0 to 4:2.

Where is the 10:2? Why even have an outcome prediction at all?

Nice test. The average outcome was 6:1.1 and the best was 9:1 and the worst was 3:2. Are you arguing that the results were too varied or that the average didn't match the odds given? These are two different issues. By the way, the combat and prediction systems are fundamental parts of the vanilla game and something we haven't dare touch because they are so fundamental.

[GaryChildress]

Another quick test: how long does it take to reload a saved game, including the time it takes to exit to the main menu to get the "load game" option.

Answer: as little as 20 seconds. In one minute I can just about give it 3 tries to get a better combat result.

Just think if it took even 1 minute to reload. How many gamers have the patience to sit there a minute between tries. Would 1 minute be a pain for someone who had 5 games going? I wouldn't think so.

[GaryChildress]

OK. Just performed a test:

Predicted outcome = 10:2

1st thing to notice: the predicted results NEVER appeared out of 15 tries.

From there we have best (9:1) all the way to worst (3:2). Within the fist 6 tries I can get from 7:0 to 4:2.

Where is the 10:2? Why even have an outcome prediction at all?

Nice test. The average outcome was 6:1.1 and the best was 9:1 and the worst was 3:2. Are you arguing that the results were too varied or that the average didn't match the odds given? These are two different issues. By the way, the combat and prediction systems are fundamental parts of the vanilla game and something we haven't dare touch because they are so fundamental.

BOTH! So where is the predicted 10:2 outcome (even 5:1 only happened once)? And yes, 3:2 is a WHOLE different ballgame from 9:1. It may not seem like it but it is easily the difference between conquering Poland in 2 turns versus 3. Give me a 9 and I'll destroy the enemy unit in 2 hits. Give me a 3 and it'll take 3 or 4 hits.

[gerones]

The only way I can think of, aside from a central server registering every transaction (upload, download, loading, saving), is to watermark a game file with a unique ID. Once your file is loaded by your opponent, his CEAW would have to store that watermark in an encrypted registry. Only then would CEAW be able to tell whether it has ever loaded a specific file before, and write this information into the output game file, from where it could be read by your CEAW, issuing a warning like "Your previous turn has been reloaded 4 times by your opponent." Even then, it would be easy to work around this by having more than one CEAW installation. Never mind that you would still have to modify CEAW code first.

Something like this would be a start on this keeping in mind that one can have 2 or more installation but pbem files go always to the same GS/data/pbem subfolder placed in My documents folder.

2. Make reloading WAY more cumbersome. I would even go so far as to remove the "load game" option from the main menu, except ONLY when you first launch the game. Basically when you first launch the game you could be given the option to load a saved game and then once in the game the only way to load a save game would be to quit to desktop and relaunch the game. Seriously, this would probably be even more effective than #1. Who is going to cheat if you have to go around the world just to get back to square 1 again?

We have to keep in mind that other strategical wargames as Advanced Tactics has so effective PBEM cheating prevention that even re-installing the game won´t let you cheat at all. So it´s not a question of making the things a little bit hard for the cheaters but a question of implementing a really effective cheating prevention.

[GaryChildress]

We have to keep in mind that other strategical wargames as Advanced Tactics has so effective PBEM cheating prevention that even re-installing the game won´t let you cheat at all. So it´s not a question of making the things a little bit hard for the cheaters but a question of implementing a really effective cheating prevention.

How does Advanced Tactics prevent cheating? I'm not familiar with the game.

Basically making cheating more difficult would not only curb PBEM cheating but also make the game more enjoyable in single player IMO. From there you can also add reload alerts for PBEM games.

[GaryChildress]

Let me also say that CEAW is one of the FUNNEST games I own. It's one of my favorites. So I'm not arguing that the game is no good, only that it can be made better. Even with its few flaws I still recommend it to anyone who likes to wargame.

[schwerpunkt]

OK. Just performed a test:

Predicted outcome = 10:2

1st thing to notice: the predicted results NEVER appeared out of 15 tries.

From there we have best (9:1) all the way to worst (3:2). Within the fist 6 tries I can get from 7:0 to 4:2.

Where is the 10:2? Why even have an outcome prediction at all?

Nice test. The average outcome was 6:1.1 and the best was 9:1 and the worst was 3:2. Are you arguing that the results were too varied or that the average didn't match the odds given? These are two different issues. By the way, the combat and prediction systems are fundamental parts of the vanilla game and something we haven't dare touch because they are so fundamental.

BOTH! So where is the predicted 10:2 outcome (even 5:1 only happened once)? And yes, 3:2 is a WHOLE different ballgame from 9:1. It may not seem like it but it is easily the difference between conquering Poland in 2 turns versus 3. Give me a 9 and I'll destroy the enemy unit in 2 hits. Give me a 3 and it'll take 3 or 4 hits.

Something to note: All forecasts are inflated by 50% so a 9:3 forecast is really a 6:2 - thats not to say that 7,8,9 damage results cant happen as you have noted, just that they are at the low end of probability. The other thing to note is that the 3 return damage result will only apply if you dont score much damage (eg 0 or 1 steps) to the target (thats my observation from hotseat combat result tests).


Also, like others, I'd like to see a little less variation in combat results as I'd like skill to be a bit more dominant than luck - this is one of the reasons that I gave up playing the board game World in Flames as I was a notoriously poor dice roller. Not that I'm alone in that - one of my mates playing Germany currently in a game has lost the whole of his airforce during the first few turns of Barbarossa as the allied player continually rolled 11 or 12 on two dice whilst he couldnt roll above 5 on the two dice.

[ncali]

My understanding of the combat odds caluclator is that it isn't a predictor of results at all. Instead, it compares the relative strengths of units and does not even try to calculate results based on the sequence of combat. Thus, I think it is more accurate for air to air combat, which is simulataneous, than it is for ground combat, which follows the order (1) attacker suppression of defense unit, (2) defense unit fires and damages attacker, and (3) attacker fires. I think this means that when a defense unit has high efficiency, the calculator is wildly inaccurate (and suggests it is a more favorable attack than it is) since the defending unit will still be fairly effective after suppression and will then get to fire first.

But I could be wrong! In any case, it would be preferable if the combat odds calculator were just that. For example, if it ran say 10-50 attacks and took the averages for attacker and defender losses.

[gerones]

How does Advanced Tactics prevent cheating? I'm not familiar with the game.

The system seems to work with a key file that is created for every single PBEM turn file and only for this. So even you reinstall the game that PBEM file won´t work anymore for any single game you play.

So I'm not arguing that the game is no good, only that it can be made better. Even with its few flaws I still recommend it to anyone who likes to wargame.

Me too

[KingHunter3059]

There is a solution for this, I gave it some thought over the weekend. For this to work, you would do it with three people. Two would be contestants, and the third a moderator. You could send the moderator what your moves are and what combat you would want (You could do this in a saved Hotseat for example, or JPG Screen shot, then the moderator (Or Game Master if you prefer) would in-turn, move the contestants and do the battle moves via hotseat, then return the results back to each player. Of course this eliminates any would be 'cheating' as the moderator should be neutral in all cases. I have no issue at being the moderator, between two combatants, as my skills at Generalship may be a bit under proven.


Jay

[jjdenver]

@ rkr1958: Then what do you recommend? As the game stands now, cheating is ridiculously easy and pays off big rewards. So long as that is the case it makes more sense to cheat than it does to play by the rules.

As far as randomness, I agree that war has some unpredictable factors that play in victory and defeat. I'm not saying do away with randomness completely, only have less variation. Radically divergent combat results should be the exception rather than the rule. I don't think Midway was as random as many think it was. It was in part testimony to things like poor Japanese carrier design, poor AA escort philosophy, superior Allied damage control practices and other factors. If you could "restart" Midway 10 times, the Allies probably would have won 9 out of the 10.

WWII was largely a war of statistics. Attrition played the biggest aggregate role in the outcome of it. Germany and Japan lost because they failed at the numbers game, not because they were unlucky in battle. At the macro scale that CEAW represents basically whoever has the greater numbers and whoever uses their forces best should win, not whoever happens to be luckier. There shouldn't be 10:2 odds with 1:3 outcomes. Sorry but I don't see the logic in so much randomness. Why even have attack and defense values or have estimated outcomes if everything is so random?

This is a great post. I quit starting new PBEM games of CEAW about 6-9 months ago specifically because of people cheating in PBEM (or what seemed like it) removing all fun from the game for me. It's a great game but I'd only do PBEM again if the game can implement some way to prevent/mitigate cheating. All of the great work done by the GS mod team is wasted if PBEM is not fixed. That should be their number 1 priority w/ a bullet instead of all the bell and whistle enhancements imo.

Some possibilities:
1) System similar to Advanced Tactics to prevent PBEM cheating
2) System similar to SSG games to prevent PBEM cheating
3) Use the server-based PBEM that a lot of Slitherine games seem to be using to prevent PBEM cheating
4) (least preferred) DRASTICALLY reduce randomization of results to give less benefit to PBEM cheating. The predicted outcome should be the outcome 75% of the time, with very minor variation by 1 point 25% of the time.

If the game ever gets this problem fixed it'll rocket back to the top of my favorite games list.

Cheers

[rkr1958]

This is a great post. I quit starting new PBEM games of CEAW about 6-9 months ago specifically because of people cheating in PBEM (or what seemed like it) removing all fun from the game for me. It's a great game but I'd only do PBEM again if the game can implement some way to prevent/mitigate cheating. All of the great work done by the GS mod team is wasted if PBEM is not fixed.

I'm sorry you feel that way and have had a bad experience playing PBEM because you suspect your opponents of cheating. However; I couldn't disagree more. The GS expansion, which came out of the BJR-mod, was developed by a group of fans that what to increase the realism of the game. Originally we were content to just play it among ourselves but then decided to share it with the community, in general and as is for free. This group of fans do this as a hobby and for free. For us, improving the realism of the game within the constraints of keeping it fun and balanced (as best as possible) has always been our main goal. Where we could make improvements in security with modest effort we have, such as: (1) hashing the PBEM passwords so players can't get them by using a text editor to open the saved game, finding their password (which they know) and getting their opponents password and (2) adding a checksum to alert players if they have a different, or changed, general.txt file from the one their opponent used last turn.

That should be their number 1 priority w/ a bullet instead of all the bell and whistle enhancements imo.

Again, I couldn't disagree more and I'm sorry that suspected cheating has spoiled the fun of PBEM play for you. I have two comments and then I'll end my rant. (1) I believe PBEM security can only be solved by playing on a server based system, which is a business decision for Slitherine, and which is not for me as a hobbyist, to make. (2) In terms of enjoyable PBEM play within the currebnt system I would recommend finding a group of people that you can trust and play with them. Ingames where you suspect your opponent of cheating then I would recommend that you stop the game and don't' play them anymore.